The Ledger

“I think that one of our favorites was Charles River.”

Episode: Citrini: Monkeys Are The Next Bottleneck | Citrini on AI Drug Discovery Experiment & Reflexive Financing Risk to AI Buildout & Hardware Trade

at 36:04

After using Claude to design scorpion anti-venom mini-binders (17 of the designs bound in the lab), Citrini concluded that AI compresses the dry-lab design work from months to days while synthesis and testing still take months, so the trade is to be 'long experiments'. Charles River is a favourite: about 60% of its revenue is from discovery, safety assessment and early-stage experiments, it is emerging from a cyclical trough with spare capacity, so each new contract is higher margin and earnings growth could inflect. Separately, monkeys for preclinical testing are the single most supply-constrained step before clinical trials (China cut exports and breeding takes about four years), which he sees as a bottleneck harder to expand than memory supply. He calls it the highest-conviction long from the piece, and says monkey pricing data would be the key to the trade.

How sure

4 out of 5

Copy citationCopy link

“I'm bullish on Meta because I think that they're really killing it with Muse.”

Episode: Citrini: Monkeys Are The Next Bottleneck | Citrini on AI Drug Discovery Experiment & Reflexive Financing Risk to AI Buildout & Hardware Trade

at 1:07:31

Citrini turned bullish on Meta this summer, before its Muse model was released, and says Muse confirms the thesis. Meta holds vast data on its users across Instagram, WhatsApp and Facebook with ARPU around $200-something; if Muse gets any meaningful adoption he draws a straight line from that data to ARPU rising about 4x by 2030, with more room to run on its existing AI infrastructure. He holds it as the long leg of a long-Meta, short-Google pair.

Time frame

by 2030

What they own

Says on air that long Meta, short Google is a trade he has on.

How sure

3 out of 5

Copy citationCopy link

“I'm still pretty bullish on NVIDIA from here.”

Episode: Citrini: Monkeys Are The Next Bottleneck | Citrini on AI Drug Discovery Experiment & Reflexive Financing Risk to AI Buildout & Hardware Trade

at 1:18

Citrini says his research shop moved away from a broad long-semiconductors position from around March, shifting toward the users and enablers of AI ('agentic utilities'), because the infrastructure trade now requires underwriting significant expectations. He is explicit that this does not mean he is done with semiconductors: he remains pretty bullish on NVIDIA from here, while saying the more asymmetric opportunities now sit elsewhere in the AI trade.

How sure

3 out of 5

Copy citationCopy link

“And so I think that's an example of a company that is going to benefit from this commoditization of phase one assets.”

Episode: Citrini: Monkeys Are The Next Bottleneck | Citrini on AI Drug Discovery Experiment & Reflexive Financing Risk to AI Buildout & Hardware Trade

at 55:26

Citrini expects large pharma to show significant dispersion as AI improves, playing out over a long period. Eli Lilly has a strong record of acquiring molecules and bringing them to FDA approval, and is a trailblazer in using AI itself: its drug-design model trained on its own proprietary data requires outside users to contribute designs, so it keeps gathering data. He sees it as an example of a company that will benefit from the commoditisation of phase one assets, while weaker companies reliant on a single asset in a well-understood mechanism lose out as the time between first and follow-on drugs keeps shrinking.

Time frame

this will play out over a longer period of time

How sure

3 out of 5

Copy citationCopy link

“And then we're pretty bullish on life sciences in general because we think that funding for experiments will go up even though rates have come up.”

Episode: Citrini: Monkeys Are The Next Bottleneck | Citrini on AI Drug Discovery Experiment & Reflexive Financing Risk to AI Buildout & Hardware Trade

at 1:00:04

Citrini's thesis is that AI is commoditising drug design and will create an explosion of new molecules and experiments, mostly in the preclinical space, while the experiments themselves cannot be sped up much. Beneficiaries are the companies that run experiments or make them quicker: contract research organisations, consumables and lab automation (which he calls a really clean way to play it, since speed matters more than cost to AI labs). AI labs and AI-biology startups become new customers because they need experimental data to improve their models. The sector also looks like a cyclical recovery from a post-COVID overbuild at trough pricing, with a secular story on top, like semis in early 2023. He is bullish on life sciences in general because he expects funding for experiments to rise even with higher rates.

How sure

3 out of 5

Copy citationCopy link

“Actually, we kind of got a little bullish on the CoreWeave bonds specific.”

Episode: Citrini: Monkeys Are The Next Bottleneck | Citrini on AI Drug Discovery Experiment & Reflexive Financing Risk to AI Buildout & Hardware Trade

at 1:11:01

Citrini thinks the recent CoreWeave scare was way overblown and got a little bullish on CoreWeave's bonds specifically (the bonds, not the shares). He does not expect CoreWeave to default, noting it has a sizeable equity cushion it can tap, though that can shrink fast as the stock falls. If the AI cycle does turn, he expects it to end with a whimper, through downward revisions to combined lab ARR, rather than with a CoreWeave default breaking the cycle.

Not sure

Yes, said with doubt

How sure

2 out of 5

Copy citationCopy link

“I think if you said to me, George, forget about the next three months, next three years, I think energy is going to absolutely run rings around the S&P, not even close.”

Episode: Coreweave's a Zero" | George Noble on Why He's Short Neoclouds, Why AI is a Shale 2.0 Bubble Headed for Debt/Depreciation Brick Wall, and Why He's "Coming For The Momentum Bros

at 1:06:28

Noble says his energy call predates the Iran war: energy stocks are cheap, demand is robust, production depletes about 5% a year after 15 years of underinvestment, and the Permian can no longer supply the marginal growth. On a gold-to-oil ratio crude looks cheap. The war has distracted from that case, and supply security now has lasting value. He concedes energy could fall in the short run if peace breaks out, but over three years expects it to far outperform the S&P.

Time frame

next three years

What they own

Owns energy stocks, including Schlumberger; said energy was his call for the year in December.

How sure

4 out of 5

Copy citationCopy link

“So I love Schlumberger. It's not the only stock I own, but I think energy as a group is going to do incredibly well.”

Episode: Coreweave's a Zero" | George Noble on Why He's Short Neoclouds, Why AI is a Shale 2.0 Bubble Headed for Debt/Depreciation Brick Wall, and Why He's "Coming For The Momentum Bros

at 1:09:27

Noble says he loves Schlumberger. Within his energy bull case, he cites Robert Mullen's argument that SLB is an overlooked AI play: its proprietary subsurface data and its data-centre construction business could on their own be worth its market cap, before counting the oilfield-services business.

What they own

Says it is 'not the only stock I own', i.e. he owns SLB.

How sure

4 out of 5

Copy citationCopy link

“I can talk about the balance sheet of land research, which I'm bullish on.”

Episode: Coreweave's a Zero" | George Noble on Why He's Short Neoclouds, Why AI is a Shale 2.0 Bubble Headed for Debt/Depreciation Brick Wall, and Why He's "Coming For The Momentum Bros

at 1:24:09

Farley says he is bullish on Lam Research and knows its balance sheet. Ahead of its report that evening he says 'I think it's going to rip', as part of his wider expectation of a semiconductor rally and of chip spending above sell-side estimates.

Time frame

tonight's earnings report

How sure

3 out of 5

Copy citationCopy link

“So I think the price of gold is going to go up.”

Episode: Coreweave's a Zero" | George Noble on Why He's Short Neoclouds, Why AI is a Shale 2.0 Bubble Headed for Debt/Depreciation Brick Wall, and Why He's "Coming For The Momentum Bros

at 55:00

Noble expects the gold price to rise because irresponsible monetary and fiscal policy is destroying the value of money: no one will raise taxes or cut spending, foreign buyers of US debt are retreating, and he expects authorities to have to print again. He sees gold as an insurance policy against central-bank error and a way to preserve wealth in real terms, adding that he lost a little money in gold this year.

What they own

Says he lost a little money in gold this year; some of his gold stocks fell and then rebounded.

How sure

4 out of 5

Copy citationCopy link

“And I say SSRM has been my favorite name.”

Episode: Coreweave's a Zero" | George Noble on Why He's Short Neoclouds, Why AI is a Shale 2.0 Bubble Headed for Debt/Depreciation Brick Wall, and Why He's "Coming For The Momentum Bros

at 55:00

Within his bullish gold case, a debasing dollar and fiscal excess forcing more money printing, Noble names SSR Mining (SSRM) as his favourite gold stock, after his gold picks fell and then rebounded this year.

What they own

Says SSRM 'has been my favorite name' among his gold picks.

How sure

3 out of 5

Copy citationCopy link

“I do think that we are in for a pretty foul semiconductor rip, and we are also shorted SpaceX, but against semis, so we have way more semi-longs.”

Episode: Coreweave's a Zero" | George Noble on Why He's Short Neoclouds, Why AI is a Shale 2.0 Bubble Headed for Debt/Depreciation Brick Wall, and Why He's "Coming For The Momentum Bros

at 1:32:55

Farley, the host, expects a sharp semiconductor rally after nine straight down days. He thinks sell-side 2027 earnings are too low because hyperscalers and xAI will spend heavily on chips, so NVIDIA may be cheaper than its 17-18 times forward earnings suggests. He says semis are a better way to own the AI trade than neoclouds. His model portfolio holds far more semi longs than shorts, with SpaceX, AXT and SanDisk shorts as hedges.

What they own

His Substack model portfolio is net long semis (about 50-60% beta to semiconductors), short SpaceX against them, and also short AXTI and SanDisk; he covered a CoreWeave short the night before.

How sure

3 out of 5

Copy citationCopy link

“I think the opposite is true with crude. I think energy prices are sustainably higher.”

Episode: Coreweave's a Zero" | George Noble on Why He's Short Neoclouds, Why AI is a Shale 2.0 Bubble Headed for Debt/Depreciation Brick Wall, and Why He's "Coming For The Momentum Bros

at 1:19:28

Noble expects oil prices to stay sustainably higher, a 'higher for longer' environment, because of years of underinvestment, depletion and the end of easy Permian growth, and because rising deferred futures prices show the market expects higher prices. He is not calling for $150 oil and does not want a spike that would trigger recession talk.

How sure

3 out of 5

Copy citationCopy link

“My point is, we're not going back to goods deflation, and if we're not going back to goods deflation, we cannot disinflate back under, say, 3% on overall inflation.”

Episode: Jim Bianco: Why I’m Finally Buying Bonds After Six Years

at 10:07

Bianco's inflation argument is an arithmetic one about composition. From China's WTO accession in December 2001 to 2020, goods PCE inflation averaged minus half a per cent while services ran two and a half, and the goods deflation is what held the total near 1.6 to 1.7 per cent -- central bankers took credit for an outcome cheap Chinese goods delivered. Post-COVID de-globalisation, tariffs, wars and chronic supply-chain problems have taken goods from minus half a per cent to plus 2, which puts the total in the low threes. He accepts the San Francisco Fed's finding that tariffs drive most of the goods inflation, and says it changes nothing: the tariffs are a deliberate choice both parties now make, because the price of that goods deflation was gutted manufacturing towns, which he ties to the Case and Deaton deaths-of-despair work. His test case is BYD: a Seal that matches a Tesla Model 3 at 15,000 dollars and a smaller electric car at 10,000, banned by Biden rather than Trump, because letting them in would bankrupt Detroit -- and Europe, which did let them in, is now watching its manufacturing base and its politics turn. So there is no route back to goods deflation and no route back under 3 per cent. He is equally firm on the other side: he is not forecasting 4, 5 or 6 per cent, and he is not in the recession camp. The diesel channel is his evidence that it filters through -- Ukrainian strikes on Russian refineries have pushed European gas oil up 130 to 140 per cent, the US is the marginal supplier, and because every commodity is, as he quotes Jeff Currie, diesel and dirt, the cost sits in every price on every loading dock. PPI, which he would rather call business inflation, is running hot on chips and diesel and eventually feeds consumer inflation, with the current tariff refunds only masking it for a while.

How sure

4 out of 5

Copy citationCopy link

“I'm a gigantic AI bull. I think it is as revolutionary as the railroads, anything we've seen in the last 100 years.”

Episode: Jim Bianco: Why I’m Finally Buying Bonds After Six Years

at 50:39

Bianco calls himself a gigantic AI bull and rates the technology as revolutionary as the railroads or anything in the last century. He grants the bubble: every technology ends in one and this will be no different. His argument is about timing. A bubble peaks when the skepticism is gone and only wild-eyed optimism is left, and right now there is too much skepticism and not enough compute -- he reads Ed Zitron's popularity as evidence of the narrative people want to believe, and says the moment that narrative is rejected is the moment to worry. His analogy is Greenspan's irrational exuberance at 1,100 on the Nasdaq in December 1996: the index went up roughly 400 per cent more before peaking at 5,100 in March 2000, and after the bust it bottomed at 1,200, never returning to the level at which the bubble was first called. So he puts this nearer December 1997 than spring 2000 and expects the market to go up a lot more than people think before the peak. On adoption he argues the financial industry systematically underrates AI because FINRA and internal IT have throttled it, so practitioners conclude there is nothing to it; he thinks it will redefine how work is done rather than destroy jobs, and that firms will willingly pay AI fees that already dwarf hardware costs because the productivity gain is orders of magnitude larger. His stated worry is not failure but success: if the productivity gains arrive, nominal growth and interest rates go higher, and the parts of the economy that cannot be made more productive -- the corner grocer, the new restaurant, landscaping, construction -- face 8 or 9 per cent borrowing costs, which is the strain he thinks is underpriced. He draws the parallel to George Pullman, whose casket sits in fifteen feet of cement in Chicago because the railroad boom worked and the anger it caused was real.

Time frame

three years or so to the bubble peak; he likens now to December 1997

Not sure

Yes, said with doubt

What they own

Says 40 per cent of the 5 per cent convertible-bond weighting in his index is AI-related and that he expects that space to go up; says he would probably not buy CoreWeave.

How sure

4 out of 5

Copy citationCopy link

“So I see that as being localized, idiosyncratic problems in the credit market. I don't see it as being a broader issue that, first the hyperscalers can't borrow, and then the credit market goes bust, and then the next thing you know, Kroger can't borrow, and they got to start closing stores. We are a long, long way from that happening, if it ever happens at all, which I don't think it will.”

Episode: Jim Bianco: Why I’m Finally Buying Bonds After Six Years

at 50:39

Bianco rejects the reading that widening AI-linked credit is the start of a credit cycle. His starting point is the balance sheet: corporate credit peaked at 290 per cent of GDP in March 2008, the month Bear Stearns failed, against 100 per cent for government-linked debt, and the two have since crossed -- corporates down to about 140 per cent after an 18-year deleveraging, government up to about 160. The corporate market is therefore not over-levered and not dependent on rolling risky loans. Then he decomposes the move: high-yield spreads at their widest since April, about 360 over and an 8 per cent all-in yield, with essentially all of the widening coming from CCC and single-B and double-B barely moving. CCC, in his analogy, is detention -- a small set of problem children that issue little new paper and mostly refinance, among them CoreWeave, two over-levered cable companies, Aston Martin, Hertz, video-game operators and the French lottery operator. They are in trouble because they are over-levered and refinancing has got harder, not because they are signalling anything about the school. Hyperscalers paying wider spreads is simply the higher expected return they are supposed to offer; if they cannot afford it they get sent to detention too. He sees the stress as localised and idiosyncratic and is a long way from believing it reaches the ordinary borrower.

What they own

His index holds a 5 per cent weighting in convertible bonds, about 40 per cent of which he says is tied to the AI story, and he says he expects that space to go up.

How sure

4 out of 5

Copy citationCopy link

“Bonds are no longer expensive, meaning the yields are too low relative to the economy. And I'll step up and own longer durations, meaning longer maturity bonds with higher yields because I want to capture those yields.”

Episode: Jim Bianco: Why I’m Finally Buying Bonds After Six Years

at 1:48

Bianco turned constructive on the bond market in early October for the first time in six years, and he is careful about what the call is: bonds are not cheap, they are no longer expensive. His anchor is nominal growth -- roughly 3 per cent sticky inflation plus about 2 per cent real growth, so a 5 to 6 per cent economy -- and by the start of the month every yield from the three-year to the thirty-year carried a five handle, which he reads as the appropriate level reached at last. On that basis he has stepped up and bought longer duration to capture the yields, and he argues the arithmetic now protects him: because the yield is the biggest determinant of bond returns, it would take a 14 to 15 per cent ten-year within five to seven years to produce a negative ten-year return, and in that world bonds would be the best thing to own anyway because the loss would be about 1 per cent while everything else collapsed. The painful part, getting rates off zero to 5 per cent, is behind us. He is explicit that this is not a recession call and that rates going higher from here is always possible, and he names the escalation scenario -- a further widening of the war and 200 dollar crude -- as what would break it. His own risk case is the opposite of the bear case: if AI delivers, nominal growth rises and the five handle becomes a six.

Time frame

the next 10 years

Not sure

Yes, said with doubt

What they own

Says he stepped up and went long longer-maturity bonds; runs the index behind the WTBN ETF, which holds a 5 per cent weighting in convertible bonds.

How sure

4 out of 5

Copy citationCopy link

“So we have definitely liked energy stocks and let's say more broadly resource stocks for quite a while because of their diversifying characteristic.”

Episode: The AI Bubble Will Pop from Equity Supply Says This $80 Billion Asset Manager | Ben Inker | GMO

at 46:42

Inker's case for energy is correlation rather than a price forecast. He says most stocks are hardly correlated with each other at any frequency, and resource companies are the one group whose correlation with everything else keeps falling as the holding period lengthens, so there is a bad case for the rest of the portfolio that is not a bad case for them. He does not claim they hedge every inflationary shock, but says a chunk of them are very well positioned for the kinds that matter. Asked what he would own knowing nothing else about the world, he names two groups on a secular basis: resources and high-quality stocks, the latter because they carry less economic risk without costing you long-term return. The discipline he attaches is valuation and crowding, that you do not want to own energy when everyone is saying to own as much energy as possible, and in practice today the firm expresses this alongside a preference for non-US over US and value over growth.

Not sure

Yes, said with doubt

What they own

Says GMO has liked energy and resource stocks for quite a while for their diversifying characteristic, and names resources and quality as the two groups he would be over-allocated to knowing nothing else.

How sure

3 out of 5

Copy citationCopy link

“I guess the short answer is no, because we overweighed emerging markets.”

Episode: The AI Bubble Will Pop from Equity Supply Says This $80 Billion Asset Manager | Ben Inker | GMO

at 50:30

Asked whether he is underweighting the AI winners in Korea and Taiwan, Inker says no, because GMO is overweight emerging markets and those are the two largest markets in the index. His reasoning is that the Korean and Taiwanese beneficiaries have not re-rated the way the US AI names have: valuations have not gone up that much, what exploded was profits, so Taiwan is not full of companies priced as if they will be the most amazing businesses in history. In his numbers profits went up roughly tenfold while prices went up sevenfold, which is why the firm is still a holder. He has trimmed, because he does not believe those profit levels are fully sustainable, and he is less excited than when the prices were much lower. The allocation sits inside a broader preference he states for non-US over US equities on price.

Not sure

Yes, said with doubt

What they own

Says GMO is overweight emerging markets and still holds the Korean and Taiwanese AI beneficiaries, a little less than before; also says the firm would rather own non-US than US stocks and value than growth.

How sure

3 out of 5

Copy citationCopy link

“But one of our favorite holdings is Orsted, which owns the windmill farms and a secondary one in the space is Cattler.”

Episode: Top Performing Value Investor Explains Why the Market Isn’t Overvalued and Hasn’t Hit the Blowoff Top Phase | Charles Lemonides | ValueWorks LLC

at 55:32

Lemonides counts Orsted, owner of offshore wind farms, among his favorite holdings, as his main way to own offshore wind after the sector was kicked to the curb, including by the US administration.

What they own

Among the fund's favorite holdings.

How sure

3 out of 5

Copy citationCopy link

“And energy prices are high. And I think there's a good story to be had in those energy names.”

Episode: Top Performing Value Investor Explains Why the Market Isn’t Overvalued and Hasn’t Hit the Blowoff Top Phase | Charles Lemonides | ValueWorks LLC

at 18:41

Lemonides built a large energy position in the pandemic, scaled it back after the stocks exploded, and says the remaining holdings have since retreated a lot even though energy prices are high, which leaves a good story in the names. He says he missed refiners and tankers and thinks it is too late for them; his preferred places are domestic independent E&Ps and offshore service companies. He also runs short energy positions against the longs.

What they own

Still holds part of an energy position built during the pandemic.

How sure

2 out of 5

Copy citationCopy link

“And I don't think we've seen that lift off quite yet in this cycle. So I think you have to, on the one hand, be prepared for the market turn whenever it happens. And on the other hand, be prepared to participate because it could last another two, four years, five years, who knows?”

Episode: Top Performing Value Investor Explains Why the Market Isn’t Overvalued and Hasn’t Hit the Blowoff Top Phase | Charles Lemonides | ValueWorks LLC

at 1:58

Lemonides says the market is fully valued but not overvalued, with only a handful of large companies at excessive prices. He argues that after 18 years without a correction of more than 25% it is silly to time the end, that the best returns come in the final years of a bull market, and that bull markets have a blow-off to the upside before the downside, which he does not think has happened yet. Rates are rising because growth is strong, higher rates take a while to slow growth, and he sees nothing indicating a slowdown six months out. He hedges by saying investors should also be prepared for the turn whenever it comes.

Time frame

it could last another two, four years, five years

Not sure

Yes, said with doubt

How sure

3 out of 5

Copy citationCopy link

“So we're getting at that stock for nothing, and we're getting the parent company at very, very fair prices.”

Episode: Top Performing Value Investor Explains Why the Market Isn’t Overvalued and Hasn’t Hit the Blowoff Top Phase | Charles Lemonides | ValueWorks LLC

at 53:34

Lemonides gets quantum-computing exposure through Honeywell, which owns about half of the newly listed Quantinuum. He owns Honeywell and argues he is getting the Quantinuum stake for nothing and the parent at very fair prices.

What they own

Owns Honeywell.

How sure

3 out of 5

Copy citationCopy link

“We were more enthusiastic to build on that exposure in the middle of 20s than Intel in the high 80s, low 90s.”

Episode: Top Performing Value Investor Explains Why the Market Isn’t Overvalued and Hasn’t Hit the Blowoff Top Phase | Charles Lemonides | ValueWorks LLC

at 14:09

In the AI-chip sell-off Lemonides preferred to add to 'Wolf', which he calls a somewhat different chip company, after it fell from a high of about 80 to about 22, rather than add to Intel in the high 80s and low 90s. The rationale given is price: the bigger fall made it the better place to build chip exposure. The transcript renders the name as 'Wolf'; read here as Wolfspeed.

What they own

Has been a buyer and built the position in the mid-20s.

How sure

3 out of 5

Copy citationCopy link

“The other one in that space is one of our favorite names which is Hornbeck.”

Episode: Top Performing Value Investor Explains Why the Market Isn’t Overvalued and Hasn’t Hit the Blowoff Top Phase | Charles Lemonides | ValueWorks LLC

at 31:48

Lemonides calls Hornbeck, an offshore supply-vessel operator, one of his favorite names. He owned it through its bankruptcy and private period; about a month ago it merged into an existing public company of about $1.5 billion that was renamed Hornbeck, and he thinks its first combined quarterly results will surprise people. The thesis is the same as Tidewater's: a tight vessel market and a price well below replacement cost.

What they own

Owned through the bankruptcy and still held.

How sure

4 out of 5

Copy citationCopy link

“On the other hand, look, the growth is happening in the data center stuff and we don't think it's going to slow down anytime soon.”

Episode: Top Performing Value Investor Explains Why the Market Isn’t Overvalued and Hasn’t Hit the Blowoff Top Phase | Charles Lemonides | ValueWorks LLC

at 6:18

Lemonides describes the current advance as investment-led rather than consumer-led, with data centers and chip purchases driving growth while rate-sensitive sectors such as housing and autos are weak. He does not expect the data-center growth to slow any time soon, and expects market excitement to stay in the exciting growth companies through the final stage of the bull market. He notes that if it did slow, policymakers could cut rates and revive housing and construction.

How sure

3 out of 5

Copy citationCopy link

“And yeah, now we like the offshore wind space.”

Episode: Top Performing Value Investor Explains Why the Market Isn’t Overvalued and Hasn’t Hit the Blowoff Top Phase | Charles Lemonides | ValueWorks LLC

at 55:32

In contrast to the power sector, which he finds far too expensive, Lemonides likes offshore wind because it has been kicked to the curb, including by the administration. He expresses it through Orsted, which owns the wind farms, and Cadeler, which owns the vessels that install the turbines.

How sure

3 out of 5

Copy citationCopy link

“And we see the same potential in names like Rivian, to Instacart, to Lyft, and a whole bunch of other ones.”

Episode: Top Performing Value Investor Explains Why the Market Isn’t Overvalued and Hasn’t Hit the Blowoff Top Phase | Charles Lemonides | ValueWorks LLC

at 8:20

Lemonides' approach is to buy companies with growth elements that trade at value prices and wait for the market to re-rate them as growth stocks, as happened with Micron. He sees the same potential in Rivian. Later he adds that Rivians will be close to fully self-driving within three years and that Rivian, like Lyft and Instacart, carries investor fatigue in its share price; he expects the payoff within two to five years rather than by year-end.

Time frame

in the next two to five years

How sure

3 out of 5

Copy citationCopy link

“But you're getting that business at a song. And if new management can make it sing, you're gonna get a triple out of the stock.”

Episode: Top Performing Value Investor Explains Why the Market Isn’t Overvalued and Hasn’t Hit the Blowoff Top Phase | Charles Lemonides | ValueWorks LLC

at 53:34

Lemonides calls Boeing super cheap at about a $175 billion valuation for one of only two big Western aircraft makers for a long time, with the company seeming to get its house in order. It also gives space exposure through a partnership with Blue Origin. If new management can execute, he expects the stock to triple.

Only if

if new management can make it sing

How sure

3 out of 5

Copy citationCopy link

“But one of our favorite holdings is Orsted, which owns the windmill farms and a secondary one in the space is Cattler.”

Episode: Top Performing Value Investor Explains Why the Market Isn’t Overvalued and Hasn’t Hit the Blowoff Top Phase | Charles Lemonides | ValueWorks LLC

at 55:32

Lemonides holds Cadeler as a secondary position in offshore wind: it owns the vessels that go out and install offshore wind turbines. The transcript renders the name as 'Cattler'.

What they own

A secondary holding in the offshore wind theme.

How sure

2 out of 5

Copy citationCopy link

“But because these are cyclical businesses and because they're one of the only folks in the game and because their yields are getting better and their technology is absolutely state-of-the-art, we think the torque to the upside is pretty dramatic.”

Episode: Top Performing Value Investor Explains Why the Market Isn’t Overvalued and Hasn’t Hit the Blowoff Top Phase | Charles Lemonides | ValueWorks LLC

at 12:11

Lemonides names Intel as a candidate for the next Micron: it has gone up about four times but is still cheap on his metrics. The case is both the CPU trade and, more importantly, the foundry: Intel is one of the few companies in the world with real manufacturing capacity at a time when no one can get enough chips, it has brought new capacity on stream over the past year and a half with yields improving each quarter, and a long-unprofitable cyclical business could swing hard. He bought around $55, added at $45-47, trimmed into $100-120, and bought a healthy stake again on the recent pullback.

What they own

Bought earlier in the year around $55, added near $45-47, trimmed at $100-120, rebuilt on the pullback; heavily weighted in the portfolio.

How sure

4 out of 5

Copy citationCopy link

“I specifically like Cord, which is an $8 billion valuation, and I think it's at three, four times what we think will be cash flow going forward, which is hard to argue with.”

Episode: Top Performing Value Investor Explains Why the Market Isn’t Overvalued and Hasn’t Hit the Blowoff Top Phase | Charles Lemonides | ValueWorks LLC

at 30:04

Lemonides likes Chord, an $8 billion domestic independent producer, at what he estimates is three to four times forward cash flow. It drills a few more wells each year, sells its oil and gas at close to full prices, and trades at a very low multiple of cash flow. The transcript renders the name as 'Cord'.

How sure

4 out of 5

Copy citationCopy link

“But that fleet of ships is going to generate massive cash flows because supply and demand has gotten completely tight.”

Episode: Top Performing Value Investor Explains Why the Market Isn’t Overvalued and Hasn’t Hit the Blowoff Top Phase | Charles Lemonides | ValueWorks LLC

at 31:48

Lemonides likes the offshore supply vessels that service drill ships, and Tidewater has been the biggest public name. Rebuilding its fleet would cost about twice the price, while the market values it near today's price, and the fleet will generate large cash flows because the vessel market has gone completely tight: the drill-ship count rises every year, each needs four or five vessels, and there are no spare ones. He sees offshore production growing for another five to ten years and buys these businesses on price relative to replacement cost.

How sure

4 out of 5

Copy citationCopy link

“And by no multiple is Micron expensive today. So it's still in the portfolio and we're going to let it run for a while still.”

Episode: Top Performing Value Investor Explains Why the Market Isn’t Overvalued and Hasn’t Hit the Blowoff Top Phase | Charles Lemonides | ValueWorks LLC

at 10:21

Lemonides has owned Micron since at least 2000 and says it went from about 70 to about 1000 in a year and a half as a value stock got priced as a growth stock. He accepts that memory stocks are most dangerous when they look cheapest on peak earnings, expects the earnings to hold for another couple of years, and is less sure about the stock at today's price than two years ago, so he has scaled the position back. He keeps it because it is not expensive on any multiple and momentum can beget momentum.

Time frame

another couple of years

Not sure

Yes, said with doubt

What they own

A core holding since at least 2000, scaled back.

How sure

2 out of 5

Copy citationCopy link

“So the names that are exactly like that, to my mind, are we're long Lyft, which is second to Uber in terms of valuation, but it's roughly a quarter of Uber's valuation.”

Episode: Top Performing Value Investor Explains Why the Market Isn’t Overvalued and Hasn’t Hit the Blowoff Top Phase | Charles Lemonides | ValueWorks LLC

at 18:41

Lemonides is long Lyft as a secondary player in a growing industry that trades at about a quarter of the leader's valuation while growing consistently and staying cash-flow positive. He rejects the winner-take-all view because rideshare is not a zero-marginal-cost business (drivers are the marginal cost) and Lyft already has the scale it needs. On autonomy, he argues driverless vehicles will be bought from automakers and become ubiquitous, so the bottleneck is the rider network, which Lyft and Uber already own. He concedes investor fatigue and possible price wars are part of why it is cheap, and hopes for a 2x to 4x within three to five years.

Time frame

in some random year in the next three or five

What they own

Long Lyft.

How sure

4 out of 5

Copy citationCopy link

“Where I think there's good value are the domestic independent E&Ps.”

Episode: Top Performing Value Investor Explains Why the Market Isn’t Overvalued and Hasn’t Hit the Blowoff Top Phase | Charles Lemonides | ValueWorks LLC

at 30:04

Asked where to be in energy, Lemonides says refiners and tankers have already exploded and it is probably too late, but domestic independent producers are good value: they pump oil and gas from acreage they already own at nearly full prices and trade at very low multiples of cash flow. He names Chord and the much smaller Unit as examples.

How sure

3 out of 5

Copy citationCopy link

“Listen, one of my favorite investments is Mesabi Trust.”

Episode: Top Performing Value Investor Explains Why the Market Isn’t Overvalued and Hasn’t Hit the Blowoff Top Phase | Charles Lemonides | ValueWorks LLC

at 42:05

Lemonides calls Mesabi Trust, a couple-of-hundred-million-dollar royalty on one Minnesota iron ore range mined by Cleveland-Cliffs, one of his favorite investments and a fair price for a royalty. Cash flow has swung between zero and $8 a share; at about $20 the yield is roughly 10-12% depending on iron ore prices. The risk is that Cliffs has been hostile and once stopped mining the range after losing an arbitration, but it is now operating the mine and paying, and he values the trust on normalised iron ore prices.

What they own

Holds it ("one of my favorite investments").

How sure

4 out of 5

Copy citationCopy link

“But I think they're going to generate $75 million worth of cash flow over the next year, and they've got $100 million of net cash. So it's small, but a bit of a gem.”

Episode: Top Performing Value Investor Explains Why the Market Isn’t Overvalued and Hasn’t Hit the Blowoff Top Phase | Charles Lemonides | ValueWorks LLC

at 30:04

Lemonides names Unit, a domestic independent producer with about a $300 million equity cap, as a small gem: he expects about $75 million of cash flow over the next year and it holds about $100 million of net cash. He concedes it is so small it may not count.

Time frame

over the next year

How sure

3 out of 5

Copy citationCopy link

“And we're long Instacart, which is perceived as second to DoorDash. And again, it's trading at one quarter of the valuation.”

Episode: Top Performing Value Investor Explains Why the Market Isn’t Overvalued and Hasn’t Hit the Blowoff Top Phase | Charles Lemonides | ValueWorks LLC

at 18:41

Lemonides is long Instacart as the perceived number two to DoorDash at about a quarter of its valuation, even though it has grown consistently with much better margins than DoorDash. Online grocery delivery is a secular growth business, and it is not winner-take-all because the shopper and the groceries are real marginal costs, so scale does not let the leader squeeze it out. He attributes the discount partly to investor fatigue and expects a 2x to 4x over the next three to five years.

Time frame

in some random year in the next three or five

What they own

Long Instacart.

How sure

4 out of 5

Copy citationCopy link

“Like if I had to lean a little toward, maybe it's AMAT a little bit, just because they have more DRAM exposure.”

Episode: Stacy Rasgon: “Demand Is Off The Charts” in Semiconductors… And Set To Double Again Soon

at 29:26

Of the three semicap names he covers (Applied Materials, Lam Research and KLA), Rasgon likes all of them but would lean slightly toward Applied Materials, because it has more DRAM exposure (the industry needs a lot of DRAM wafer capacity to make HBM, which takes three to four times the wafers per gigabyte) and because it is a little cheaper. He stresses that correlations are high and you could own any or all of them. This rests on his broader bullish WFE outlook.

How sure

3 out of 5

Copy citationCopy link

“But at some point, like I said, they're sort of creating the AI market, like ultimately doesn't work, I think, if they don't. So we've liked them.”

Episode: Stacy Rasgon: “Demand Is Off The Charts” in Semiconductors… And Set To Double Again Soon

at 21:47

Rasgon has liked Broadcom alongside Nvidia, arguing the AI market does not work without the companies creating it and that the gap with this year's 'bottleneck' trades has to normalize. Broadcom has said it can grow revenue 100% next year and again the year after and do over $30 of earnings by 2028, which against a ~$363 stock is about 11-12x if delivered. He acknowledges Broadcom has not rebounded as much as Nvidia, citing market-share concerns, and that the market is worried these are peak earnings.

How sure

3 out of 5

Copy citationCopy link

“You know, but I cover AMAT, LAM and KLIC. We like them all.”

Episode: Stacy Rasgon: “Demand Is Off The Charts” in Semiconductors… And Set To Double Again Soon

at 29:26

Rasgon says he likes all three semicap names he covers, including KLA (rendered 'KLIC' in the transcript), on his outlook for wafer fab equipment spending rising from about $150bn in 2026 to over $250bn in 2028 and potentially $300bn+ by the end of the decade. He gives no KLA-specific argument; his slight preference within the group is Applied Materials.

How sure

3 out of 5

Copy citationCopy link

“But ultimately, every AI bear case that you can think of, ultimately I think boils down to that. Answering that question, is there a return or is there not? I think there is.”

Episode: Stacy Rasgon: “Demand Is Off The Charts” in Semiconductors… And Set To Double Again Soon

at 8:22

Rasgon argues the AI build-out is generating a return and the demand behind it is real. Neoclouds rent capacity at around $30bn a gigawatt, implying a one-to-one-and-a-half-year payback; some model companies' revenues have gone vertical; and new consumer agents such as Meta's Muse point to mainstream adoption that is still early. He sees no stockpiling of GPUs: what gets built is used immediately, unlike the dark fiber of 2000, and the build is happening with demand rather than ahead of it. Supply constraints in power, land and clean rooms meter the pace, which he thinks may make the cycle last longer. Caveats: it is still early, and if nobody ends up willing to pay, that is the bear case; data-center backlash is a political risk in the midterms.

How sure

3 out of 5

Copy citationCopy link

“But my guess is we've still got some room, and especially after the pullback and we're going to earnings, and it feels to me like things still have more room to run off of this base.”

Episode: Stacy Rasgon: “Demand Is Off The Charts” in Semiconductors… And Set To Double Again Soon

at 1:02

Rasgon argues semiconductor stocks have room to run despite falling multiples. The SOX is up about 80% year to date while earnings are up more than 100%, and forward estimates have kept rising since the June peak even as the stocks fell, so multiples have compressed further. Investors are pricing a cycle peak, but he does not see one in the fundamentals: Nvidia and Broadcom talk of 70-100%+ growth next year, memory makers expect supply to be even tighter in 2027-28, and hyperscalers say nobody has enough compute. He is not worried about a peak this year or next, and does not know about 2028-29. Caveats: forecasting is an unsolved problem in the industry, companies only see the orders in front of them, and if data-center spending rolled over it would cause big problems in and out of semis. He sums up as 'fairly constructive overall'.

Time frame

not this year, not next year

How sure

3 out of 5

Copy citationCopy link

“You know, but I cover AMAT, LAM and KLIC. We like them all.”

Episode: Stacy Rasgon: “Demand Is Off The Charts” in Semiconductors… And Set To Double Again Soon

at 29:26

Rasgon says he likes all three semicap names he covers, including Lam Research, on his outlook for wafer fab equipment spending rising from about $150bn in 2026 to over $250bn in 2028. He describes Lam as overweight NAND but credits it with diversifying over the years into foundry, logic and some DRAM, citing how it stayed profitable through a COVID-era collapse in NAND equipment spending. His slight preference within the group is Applied Materials, not Lam.

How sure

3 out of 5

Copy citationCopy link

“We've liked AMD.”

Episode: Stacy Rasgon: “Demand Is Off The Charts” in Semiconductors… And Set To Double Again Soon

at 49:52

Summing up his semiconductor view, Rasgon names AMD among the stocks he likes alongside Nvidia, Broadcom and semicap. Earlier he says that if AMD can ramp its rack-scale products it could do $40-50bn in AI revenue, and that a company of AMD's size only needs to take a small share of a near trillion-dollar market for it to be meaningful, so he does not worry about market share.

How sure

3 out of 5

Copy citationCopy link

“But because of that, I'm actually quite bullish on the semi-cap names I like semi-cap in terms of which one do we like?”

Episode: Stacy Rasgon: “Demand Is Off The Charts” in Semiconductors… And Set To Double Again Soon

at 29:26

Rasgon is quite bullish on semiconductor capital equipment stocks. Working back from data-center gigawatt forecasts to racks, silicon content and fab capacity, he expects wafer fab equipment spending of about $150bn+ in 2026, over $200bn in 2027 and over $250bn in 2028, and has no problem with $300bn+ by the end of the decade. He thinks demand would already support $300bn next year if clean rooms and the toolmakers' own capacity were unconstrained; the clean-room shortage meters the build-out, which he argues may make the cycle stronger for longer. Because correlations within the group are high, he says investors could simply own the group. Caveat: the forecast assumes the data centers actually get built.

Time frame

2026-2028, end of the decade

How sure

4 out of 5

Copy citationCopy link

“But at some point, like I said, they're sort of creating the AI market, like ultimately doesn't work, I think, if they don't. So we've liked them.”

Episode: Stacy Rasgon: “Demand Is Off The Charts” in Semiconductors… And Set To Double Again Soon

at 21:47

Rasgon has liked Nvidia even as it lagged the 'bottleneck' trades this year. He argues the bottleneck stocks and the accelerator makers cannot both be right and the gap has to normalize, because Nvidia and Broadcom are creating the AI market. On Nvidia's own guide of at least 70% revenue growth next year he puts earnings at $15-16, rising to about $20 if revenue doubles, against a ~$240 stock, roughly 11-12x; if this is not a peak the stocks are 'absurdly cheap'. He attributes the lag partly to technicals: investors use it as a funding source and many are capped on position size at ~8-9% of the S&P. He views competition from AMD and custom ASICs as a 'perceived threat' while the market keeps growing.

How sure

3 out of 5

Copy citationCopy link

“then dry bulk, in our opinion, is the most attractive segment out there.”

Episode: The Most Extreme Shipping Market in History, Explained | J Mintzmyer on Why Oil Tanker Rates Are Up 25x and Why It Can’t Last (and Why Dry Bulk, not Tankers, is the Next Shipping Boom)

at 47:47

Mintzmyer calls dry bulk the most attractive segment in shipping, with several firms at very attractive valuations. Capesize rates have moved from about $20,000 to $40-50,000 a day, which he places around the fourth to sixth inning, comparable to tankers in late 2025 before they took off. The order book is only slightly above average while the fleet is as old as tankers', and ton-miles should rise as China shifts iron ore imports from Australia to Guinea's Simandou, more than three times the distance, while power plants in Asia switch from LNG back to coal. He thinks rates could stay very strong for two or three years. The big caveat: if there is a global recession or a manufacturing crisis in China, dry bulk and its stocks will go down, and he says he is personally concerned about broad stock market valuations.

Time frame

two or three years

Only if

As long as there is no global recession or China manufacturing collapse

How sure

4 out of 5

Copy citationCopy link

“Fantastic opportunity, I think, a catch-up trade for product tanker rates far below where they should be in parity.”

Episode: The Most Extreme Shipping Market in History, Explained | J Mintzmyer on Why Oil Tanker Rates Are Up 25x and Why It Can’t Last (and Why Dry Bulk, not Tankers, is the Next Shipping Boom)

at 23:07

Mintzmyer had expected the unprecedented gap between crude and product (clean) tanker rates to close, with LR and MR product tanker rates doubling or tripling; instead they pulled back for a couple of months because the product fleet was more evenly dispersed and China cut refined product exports. As of the 30 September recording he says the thesis is finally working, with product rates hockey-sticking, and calls it a fantastic catch-up opportunity because product rates sit far below parity with crude. He adds that product rates could keep rising even if crude tanker rates pull back. Risk he names: a longer-lasting US diesel export ban would be very bearish for product tankers.

How sure

3 out of 5

Copy citationCopy link

“We are very bullish that company even at current prices around $19 a share.”

Episode: The Most Extreme Shipping Market in History, Explained | J Mintzmyer on Why Oil Tanker Rates Are Up 25x and Why It Can’t Last (and Why Dry Bulk, not Tankers, is the Next Shipping Boom)

at 47:47

CMB.TECH (CMBT) was his firm's public top idea for 2026, and Mintzmyer remains very bullish at around $19 a share, up from $8 last fall, seeing potential for $25-30 by next year if dry bulk rates stay modestly strong. He likes that it is mainly a dry bulk company with modern, dual-fuel Newcastlemax vessels earning above the benchmark, and that it is harvesting most of its tankers at peak prices, using about half the proceeds to pay down debt and paying out much of the rest as dividends, which should rise with dry bulk rates while cutting its exposure to the volatile tanker market.

Time frame

by next year

What they own

Says he still has a long position in CMBT and that it is in Value Investors Edge's model portfolios.

How sure

4 out of 5

Copy citationCopy link

“We like TRMD a lot.”

Episode: The Most Extreme Shipping Market in History, Explained | J Mintzmyer on Why Oil Tanker Rates Are Up 25x and Why It Can’t Last (and Why Dry Bulk, not Tankers, is the Next Shipping Boom)

at 34:14

On the product tanker side, where he thinks rates have catching up to do, Mintzmyer names TORM (TRMD) as his firm's favourite company, citing its distribution policy, corporate governance, savvy management and fleet profile, and calls it a premium company that he prefers slightly to Hafnia and Scorpio Tankers. He sees Hafnia's buying of the TORM shares Oaktree has been selling as smart rather than as a negative signal.

What they own

TRMD is in Value Investors Edge's income model portfolio; he says he has no personal position in it.

How sure

3 out of 5

Copy citationCopy link

“Obviously, I've also been a long-term Apple investor. That's been the same thesis that I've had. Like, everyone, why would you get an iPhone? You get an Android, you get a Samsung. This, no, it's the software layer that matters, that improves over time.”

Episode: Token Bill Dwarfs CPU Bill | Ben Pouladian on Meta's Muse, Anthropic S1, and Why All Roads Still Lead to NVIDIA

at 27:33

Pouladian brings Apple up unprompted as the template for the principle he is arguing about semiconductors: he wants to own businesses with a software layer that holds margin and improves over time, rather than commodity hardware that looks good in a squeeze cycle and falls when supply catches up. He says he has been a long-term Apple investor on exactly that thesis - the standing objection was why buy an iPhone rather than an Android or a Samsung, and his answer has always been the software layer. He gives no price, no horizon, no sizing, and makes no claim about Apple's AI products in this episode.

What they own

Says he has been a long-term Apple investor.

How sure

3 out of 5

Copy citationCopy link

“Ben, just to be clear, I am overweight Nvidia in my personal portfolio as well as my model portfolio.”

Episode: Token Bill Dwarfs CPU Bill | Ben Pouladian on Meta's Muse, Anthropic S1, and Why All Roads Still Lead to NVIDIA

at 18:04

Farley states his own position before pressing the other side of the argument: he is overweight Nvidia both in his personal portfolio and in his model portfolio. He makes the disclosure precisely so that the bear questions he then puts to his guest are not mistaken for his book - he goes on to argue the case for the hyperscalers' custom ASICs, that Google and Amazon are not regretting their own silicon, and later that the tightest parts of the supply chain (DRAM, NAND, HDD, substrates, capacitors) rather than GPUs are where pricing power sits. At this point in the conversation he gives no rationale for the overweight itself, no price target, no horizon and no sizing beyond the word overweight.

What they own

Says he is overweight Nvidia in his personal portfolio and in his model portfolio.

How sure

3 out of 5

Copy citationCopy link

“So I think it's, you know, 18 percent revenue growth is like Lam's average revenue growth. And we're clearly in one of the biggest semi-cap equipment cycles ever, perhaps probably the biggest one. So it just seems ridiculously low. So, you know, on my sub-stack, I've been saying that some of these analysts who cover Lam should be in jail.”

Episode: Token Bill Dwarfs CPU Bill | Ben Pouladian on Meta's Muse, Anthropic S1, and Why All Roads Still Lead to NVIDIA

at 1:06:11

Farley pitches Lam Research on the narrow test he says the big hedge funds and pod shops actually run: are earnings expectations too high or too low, and with what confidence. He accepts the guided quarter is already in the numbers - Lam guided the quarter now underway to very large sequential growth, so analysts carry it. His objection is the year after: estimates imply only about 18 percent growth, which he says is roughly Lam's own average revenue growth, from the biggest semi-cap equipment supplier to memory, where the memory makers are earning hundreds of billions of dollars a year in aggregate, in what he calls probably the biggest semi-cap equipment cycle ever. On that mismatch he calls the out-year number ridiculously low and says on his Substack that the analysts covering Lam should be in jail. He gives no price target and no position disclosure on the name.

Time frame

the fiscal year after the quarter now underway

How sure

4 out of 5

Copy citationCopy link

“I just opened up a position in Meta on yesterday's pullback. I think they've proven to obviously innovate on physical AI products with their glasses and the necklace.”

Episode: Token Bill Dwarfs CPU Bill | Ben Pouladian on Meta's Muse, Anthropic S1, and Why All Roads Still Lead to NVIDIA

at 1:05:12

Pouladian says he opened a Meta position on the previous day's pullback, when the host's question asked for an AI beneficiary other than Nvidia. His case is that Meta has built the first AI assistant an ordinary consumer actually uses - he gave Muse to his wife, who uses it for Instagram, appointments and school admin - and that while it is probably not a frontier model it does ninety percent of what people want. He compares the frontier labs to Michelin-starred restaurants and Muse to a McDonald's version for the average consumer. On monetisation he expects the revenue to come from advertising rather than subscriptions, because people will not reach into their wallets for Facebook or Instagram, and because the rest of the world - he names India and Brazil - already runs on WhatsApp and cannot pay 200 dollars a month for Pro tiers; Meta can give AI away against its existing ad network and purchase-intent queries. He reads the hire of MongoDB's CJ Desai as a signal Zuckerberg is serious about a business product. He also credits Meta with innovating on physical AI hardware (the glasses and the necklace) even though he would not wear them himself, and notes that Meta's own coding assistant may displace Anthropic, which discloses Meta as a customer in the teens of revenue. He expects Google to follow with its own agentic offering. He gives no price target and no horizon.

What they own

Says he opened a position in Meta on the previous day's pullback.

How sure

3 out of 5

Copy citationCopy link

“Before, you needed some guy to go through all this stuff. Now, you just feed it to these AI models or you create your own model and it runs and runs and runs, and you can find some targets to hit. I think that is the new paradigm shift with AI and biotechnology and science.”

Episode: Token Bill Dwarfs CPU Bill | Ben Pouladian on Meta's Muse, Anthropic S1, and Why All Roads Still Lead to NVIDIA

at 56:42

Pouladian argues compute has only now become powerful enough to model cells, organs and tissue, which are complex and multivariable, and to pan for drug candidates that would otherwise take years of animal work and trial and error. Robots and automation running tests continuously generate the data, and the models search it, which he calls the paradigm shift for biotechnology and science. His implementation is deliberately not drug picking, which he says is too hard to predict: his published basket is the tools and measurement layer - he names Danaher, Thermo Fisher, Schrodinger, Natera and Illumina, with Guardant as a competitor, and Twist as a riskier name that re-rated on its Anthropic partnership. He points to Anthropic opening a wet lab in San Francisco as evidence AI is entering the physical sciences, and to Demis Hassabis and Jeff Dean leaving Google for AI biotech as a talent signal. He discloses he is an investor in Ture, a private applied-AI molecule screening company he says has partnered with Bristol Myers Squibb. The motivation he gives is personal - both parents diagnosed with cancer - and his hope is earlier detection plus novel targets. He gives no price, no horizon and no sizing.

What they own

Says he runs a published basket of AI-in-biology tools stocks at BEP Research and discloses he is an investor in Ture, a private applied-AI company.

How sure

4 out of 5

Copy citationCopy link

“I mean, the way things are looking, we live in this crazy exponential times. And if you're thinking in exponentials, and people are gonna probably give me a lot of prep for saying this, it could probably double again, if there's enough compute.”

Episode: Token Bill Dwarfs CPU Bill | Ben Pouladian on Meta's Muse, Anthropic S1, and Why All Roads Still Lead to NVIDIA

at 53:56

Asked where the combined annualised revenue of OpenAI and Anthropic lands by the end of next year, Pouladian clarifies the question as where they exit 2027 and says that in exponential times it could probably double again, conditional on there being enough compute. He flags that he expects pushback for saying it. The base he is doubling is the figure established earlier in the conversation: he says OpenAI leaked an ARR hitting 70 billion dollars that morning and puts the two leading labs together at perhaps north of 150 billion, against roughly 30 billion at the start of the year. He adds that the labs are not spending blind - they can extrapolate from their own usage metrics - but notes the matching compute commitments are enormous, citing an Anthropic S1 with over half a trillion dollars of compute contracts through 2027 and beyond, which leaves whether they can afford the bills an open question.

Time frame

exit 2027

Only if

if there is enough compute available

Not sure

Yes, said with doubt

How sure

2 out of 5

Copy citationCopy link

“So if you have a way or you have a system that optimizes from the full stack, again, I've written about this a few times, like all roads and rockets lead to Nvidia.”

Episode: Token Bill Dwarfs CPU Bill | Ben Pouladian on Meta's Muse, Anthropic S1, and Why All Roads Still Lead to NVIDIA

at 5:57

Pouladian frames compute as a commodity business where the unit that matters is cost per token per megawatt, and argues that only a fully optimised stack - power, GPU, networking, software, security - gets that cost low enough to hold margin. On that logic he says all roads lead to Nvidia, citing Elon Musk abandoning Dojo for the NVL72 Vera Rubin platform across Tesla and SpaceX as the clearest evidence that the best in-house alternative loses to the integrated system. The second leg of the argument is financeability: a GPU is fungible, so a cluster one tenant walks away from can be re-let to another, which is why private credit will lend against GPUs; custom silicon like TPUs and Trainium is internal, not democratised and not open to outside improvement, so he asks what street those assets are on. He also points to Nvidia extending the moat into agentic security - a sandbox that runs on x86, Arm and Nvidia chips but not AMD - and to DPU-level isolation of rogue agents. His expectation is that as models converge on good enough, Wall Street will price token production cost, discounting the higher-cost producers. He gives no price target and no horizon.

How sure

5 out of 5

Copy citationCopy link

“And I keep coming to that because I want to make the point that I've always thought that asset managers are great businesses. And if you have the right cost structure and you have the right leader at the helm, they are really great businesses. And so, no, I don't think that alts are over. I don't think that alt valuations are never going to come back.”

Episode: The Private Credit Boom is Over: Redemption Requests Exceed Liquidity | James Elbaor | Marlton LLC

at 45:59

Asked whether the alternative managers' business models are in trouble after a year in which Blue Owl fell about 35% and Apollo, Blackstone, Ares and KKR fell 12-22%, Elbaor says the selling is discriminating rather than structural. The market is penalising the managers that pushed hardest into private credit and built large teams to sell the wealth channel, because those are expensive commitments that are hard to unwind when the flows stop. What it is rewarding is permanent capital: he points to Pershing Square roughly doubling from its April IPO and being awarded the peak multiple Apollo and Blackstone carried a year ago, on an 80% payout ratio, industry-leading margins and under a hundred staff. His conclusion is that the management fee on capital that cannot be redeemed is an annuity worth a high multiple, that these firms are all moving that way, and that the alt managers will find their next growth engine -- so the asset class's valuations come back even if the private-credit-concentrated names take longer to turn the ship.

Not sure

Yes, said with doubt

How sure

3 out of 5

Copy citationCopy link

“For us, we're saying if you have a view on private credit and you are most interested in making money on that, we think the BDC market is very particularly the best place to look at that right now.”

Episode: The Private Credit Boom is Over: Redemption Requests Exceed Liquidity | James Elbaor | Marlton LLC

at 27:13

Elbaor separates the loans from the wrapper and says the money is in the wrapper. Public BDCs are marked daily while the private interval funds holding similar paper are not, and the gap between them is telling you one of two things: either NAV is real and the cost of liquidity is 60 to 65 cents on the dollar, or NAV has to come down to 60 cents. His firm's view is that it is the cost of liquidity, which makes the discount an opportunity rather than a warning. The play he describes is to redeem out of the privately listed interval funds and buy publicly listed yield-producing BDCs instead. The edge he is actually underwriting is event-driven rather than mean reversion: he expects the funds trading at narrower discounts to be the acquirers of the wider-discount funds on a NAV-for-NAV basis over the next fiscal and calendar year, so a holder moves from a 30% discount to a 20% discount without needing the gap to close on its own. He says explicitly that he is not buying something just because it trades at a 38% discount and expecting it to revert.

Time frame

the next fiscal year and the next calendar year

What they own

Says his firm is doing this itself -- redeeming private interval fund exposure and buying publicly listed yield-producing BDCs -- and declines to name the positions for proprietary reasons. Says of BPRE specifically that he is not saying people should buy it, only that it should be watched.

How sure

4 out of 5

Copy citationCopy link

“But the space that performs best when risk comes off is generally the three to five year area of the curve. Generally. So that's sort of where I would be focused from a broad brush perspective.”

Episode: Why the Fed May Have to Hike Far Higher Than Expected | Henry Peabody on the Fed’s Triple Mandate and Uneven Transmission of Monetary Policy

at 43:53

Peabody does not like duration in general, but where he does want it, it is the belly. He expects further curve flattening -- real rates are staying high, ten-year one-year forward is up while breakevens are not, and improving ISM and regional survey data argue for more flattening -- and at the margin that pushes an investor out the curve. He frames the trade as a hedge with an acceptable cost of carry: if rates keep moving higher it will hurt risk assets, so he would rather hold the belly and have it not work while being able to buy risk assets more cheaply. He also makes a counterintuitive point about timing: risk typically performs well right through the inversion and only gets hurt when the curve re-steepens, so the re-steepening is the thing to watch, not the inversion. He leaves the long end to people with more fortitude than him, and separately argues long rates should be anchored to nominal GDP, where he does not think current levels are wildly out of bounds.

How sure

3 out of 5

Copy citationCopy link

“in order to subdue that demand for credit, rates may need to move higher than they would otherwise. And that, pure and simple, adds risk, but it also creates opportunity when you see volatility seeping out of that.”

Episode: Why the Fed May Have to Hike Far Higher Than Expected | Henry Peabody on the Fed’s Triple Mandate and Uneven Transmission of Monetary Policy

at 53:42

Peabody argues the Fed now has a de facto third mandate -- the transmission of credit -- and that it is failing. Because the system has become less banked, the borrowers who matter most are insulated from policy: large corporations that termed out their debt at low fixed coupons, the AI complex borrowing against aspirational returns, and private-credit lenders whose cost of capital he describes as effectively zero or negative, since they earn the yield and a fee on deployment. In a banked system a few hundred basis points forces a risk manager to widen spreads or stop lending; in this one it does not. Meanwhile credit to consumers and to CCC-rated marginal borrowers is genuinely tight -- mortgages in the sixes, used-car loans north of 7%, CCC clearing coupons well north of 10%. He calls it a K-shape in credit on top of a K-shaped economy. His conclusion is that to actually subdue credit demand among those with unconstrained access, the Fed has to move rates higher than it otherwise would, and that in every previous cycle built on a growth story -- fiber optics, mortgage brokers in 2007 -- the Fed ended up having to go higher than it expected. He explicitly accepts the corollary: the risk of overshooting and making a mistake is, in his words, absolutely real, and the higher they go the more they will have to ease afterwards. He also notes the constraint on the other side, that fiscal sustainability eventually requires rates below nominal growth, which is why he frames the end state as a second act in which the Fed has to cut aggressively once inflation falls.

How sure

4 out of 5

Copy citationCopy link

“So that sort of space, which is less correlated with credit risk, that still gives you a fairly decent return, is an awfully decent place to hide out. And in our multi-asset strategies, we're overweight there and underweight vanilla credit risk, whether that's high-yield investment grade or EM.”

Episode: Why the Fed May Have to Hike Far Higher Than Expected | Henry Peabody on the Fed’s Triple Mandate and Uneven Transmission of Monetary Policy

at 37:57

This is where Peabody wants the risk budget instead of corporate credit. He describes structured product as a large, heterogeneous market running from one-to-two-year auto ABS out to long-dated student loans and thirty-year 3% coupon agency mortgages, and says the part GMO wants is deliberately the dull end: up the capital structure at AAA, seasoned, short spread duration of roughly two years, in CLOs, student loans and non-agency mortgages. The appeal is predictability and low correlation with credit risk while still paying a decent return -- he calls it an awfully decent place to hide out, and says they are overweight it in multi-asset strategies and underweight vanilla credit risk in high yield, investment grade and EM. He frames it within a broader argument that duration can no longer be relied on as the hedge for credit and equity risk, because the post-2020 inflation regime has destabilised that relationship, which forces investors toward genuinely uncorrelated sources of return. He flags the constraint honestly: there are only so many of these seasoned assets, and as they run off, new issues come with longer tenors and are bought by investors taking more risk -- he will look there when valuations become compelling but is in the short-duration bucket for now. He also concedes the cost: risk keeps trading at the moon and the position tests your fortitude.

What they own

Says GMO is overweight this space and underweight vanilla credit risk in its multi-asset strategies.

How sure

4 out of 5

Copy citationCopy link

“I've loved European defense. It bottomed this year at a really weird time and I was like, this is a great spot to take advantage of fiscal stimulus.”

Episode: Energy Shock and Rate Hikes Could Cause a 2022 Style Bear Market | Eric Wallerstein | Clocktower Group

at 36:10

Although he dislikes European growth and would bet on European rates relative to the US, Wallerstein likes certain pockets of European equities, above all defense, which he says bottomed this year at an odd moment and offered an entry to benefit from European fiscal stimulus. Later he adds that in a downturn European defense could still do well, alongside healthcare as a defensive, while financials would struggle.

How sure

3 out of 5

Copy citationCopy link

“And I think Japanese stocks are going to be great, going to be better than everywhere else, no matter what happens. And in this bear market scenario, I'm long Japan.”

Episode: Energy Shock and Rate Hikes Could Cause a 2022 Style Bear Market | Eric Wallerstein | Clocktower Group

at 26:56

Wallerstein expects Japanese equities to outperform other markets whatever happens, including in the global bear-market scenario he is leaning towards. He contrasts Japan with South Korea: Korea's market is essentially beta to the AI boom, whereas Japan is a real economy with several drivers, manufacturing among them, that do not depend on AI alone. He separates this from his rates view: he loves Japan but would still sell its bonds.

How sure

4 out of 5

Copy citationCopy link

“And I way more favor like the Korean won than the yen, because the new Bank of Korea governor who came from the BIS, I think totally has the Korean government on sides with the fact that rate hikes are needed.”

Episode: Energy Shock and Rate Hikes Could Cause a 2022 Style Bear Market | Eric Wallerstein | Clocktower Group

at 26:56

Wallerstein strongly prefers the Korean won to the yen. The new Bank of Korea governor, formerly of the BIS, has the government on side that rate hikes are needed, because the income boom from Samsung and SK Hynix threatens to reinflate the whole economy. He sees the Korean government accepting a stronger won, quick repatriation of capital and incentives to invest domestically, and a policy-rate gap over Japan (Korea at about 3.5%, possibly 4%, if Japan reaches 1.75%) that he finds compelling for neighbouring, similar economies. He calls the won 'a much more sure thing' than a yen turn.

How sure

4 out of 5

Copy citationCopy link

“I think what's interesting, it's a story that is playing out right now, you know, like we need more copper right now, not just in five years. So it's maybe less like uranium, it's less speculation. And that's why you already see a big rise in copper price, because like it's right now, you know, even my copper forecast right now, the price is like 6.5 USD. Our price target is 7 I could have put 8, but it's for Christmas target.”

Episode: Outlook on 5 Key Commodities: Metals Bull Market is Just Getting Started (Gold, Copper, & Uranium) | Jérémie Boyer | Aurelion

at 55:12

Boyer prefers copper to most other metals because the deficit is a present fact rather than a ten-year projection — he says he is not a waiting investor and dislikes supply-demand work whose shortfall only arrives a decade out. AI and the energy transition are the demand drivers, and he notes copper has stopped trading with gold and the other miners, which he attributes to the AI bid; even if AI momentum cools he expects copper demand to persist. New mine supply takes ten to fifteen years to bring on, so the shortfall cannot be answered quickly. He sets a copper price target of $7 against a spot around $6.50, calling it a deliberately conservative Christmas target when he could have written $8. He prefers mid-cap producers coming out of their capex phase into what he calls the earning phase, where free cash flow rises at low cost, over crowded large caps. The risk he names is that data-centre construction talk has faded relative to memory and GPUs, which could soften near-term copper demand.

Time frame

Christmas target

What they own

Says Aurelion's model portfolio holds one copper company — a high-conviction, roughly $10bn producer with copper, gold and zinc exposure across Peru, the US and Canada, trading near 17x earnings — added recently, and that commodities are about 25-30% of the portfolio.

How sure

4 out of 5

Copy citationCopy link

“This is kind of easy, but maybe the safest one would be gold. I think like I don't see how it won't go back to like 5K USD an ounce.”

Episode: Outlook on 5 Key Commodities: Metals Bull Market is Just Getting Started (Gold, Copper, & Uranium) | Jérémie Boyer | Aurelion

at 2:16

Boyer names gold as Aurelion's strongest and safest commodity call and says he does not see how it fails to return to roughly $5,000 an ounce. The trigger for turning bullish again was gold decoupling from oil: after the Hormuz disruption gold had been trading inversely to oil, so every oil rally knocked gold down, and the break of that relationship in about July is the confirmation Aurelion had been waiting for. He argues rates and the Fed now explain only about 30% of gold, with geopolitics, Japanese intervention and central-bank buying (he cites a Goldman Sachs figure of roughly 66 of 100 tonnes of June global central-bank purchases coming from China) carrying more of the story, and he points to fund inflows from managers afraid of missing a move that has taken miners up about 40% in a month. He frames the horizon as six to twelve months rather than multi-year, calls himself not a prediction guy beyond that, and says Aurelion would go neutral if Hormuz reopened and the geopolitical fear premium disappeared.

What they own

Says Aurelion's published model portfolio holds no large gold position yet — "we are going to add one or two gold" — but already owns a producer that is roughly 60% copper and 40% gold, and would size gold miners at about 5-7% of the portfolio rather than 20%.

How sure

4 out of 5

Copy citationCopy link

“For uranium, it's something it's been going on for like 20 years. You know, uranium, it's up and down, but we think maybe this time could be the right one. There's never been as much governments and countries that said, and you know, sometimes a country said, I will build 10 nuclear central in the next 10 years. Sometimes they are good to not do it, but I think this time, it's probably, it's clear that they need green energy, and it should be good for both the ring price and the producers.”

Episode: Outlook on 5 Key Commodities: Metals Bull Market is Just Getting Started (Gold, Copper, & Uranium) | Jérémie Boyer | Aurelion

at 1:06:53

Boyer accepts uranium has been a false dawn for twenty years but thinks the current cycle is more likely to be real, because the volume of announced government reactor programmes is unprecedented — he cites China at about 70 units over ten years, France 5, the UK 7 and a large Indian programme — and because European governments in particular are locked into nuclear by their green-energy commitments and cannot substitute coal the way he thinks a country like South Korea might. He argues the driver is supply rather than price sensitivity: fuel is a small share of a reactor's operating cost, so operators will pay up, and each contract signed (he uses a Cameco five-year US deal as the example) resets the spot reference for the next buyer. He adds that a reactor cannot be switched on and off, which makes securing supply non-optional, and that the AI power demand now accelerating build schedules is on top of a case that already existed. The risk he names explicitly is that the announced reactors simply do not get built. He expects the move to be good for both the uranium price and the producers, and notes the long flat period before the recent move meant there was no hype to unwind.

Not sure

Yes, said with doubt

What they own

Says Aurelion's model portfolio holds two uranium positions (transcribed as "two Iranian positions"): one of the largest producers and one physical holding company, deliberately split so one leg tracks physical price and the other the speculative price. Says he personally doubled his money in a uranium name some years ago, sold out, and has just re-added it.

How sure

3 out of 5

Copy citationCopy link

“So I think a lot of those things are, the sell-off and a lot of those sectors are overdone because I think ultimately, people will want to have the expert review, their tax plan, their financial plan, their state plan, and so on, instead of doing it on their own.”

Episode: Why The 30-Year Treasury Lost Its Biggest Buyers | David Busch on Why High Yields Are Attractive Right Now

at 44:34

Busch says the software companies that sold off when AI began to be widely adopted are the sector he finds most fascinating and is watching most closely, and that the sell-off in a lot of them is overdone. His argument is about demand rather than technology: he grants that software is the industry most exposed to genuine AI disruption, which is why it sold off, but does not believe the majority of people will build their own tax, financial or estate planning software with a general-purpose model. The value sits in expertise, in complying with federal and differing state rules, and in someone willing to stand behind the result -- a rough draft from a chatbot is not the same as an estate plan that actually complies. He uses Intuit as the worked example, noting he has access to the major AI tools himself and still would not ask one to do his taxes, while conceding the sell-off there may be somewhat warranted. What he expects over the next several quarters is these companies adopting AI themselves to improve margins and efficiency, probably with fewer people and more automation, and he says he would not be dismissive of the sector at this point. It fits the first-, second-, third-order framing he applies generally: after the early entrants and the infrastructure builders, the question is who adopts fastest and benefits from adoption. When the host puts a stronger version of the case to him -- that software trades at its cheapest valuations in ten years on a priced-in disruption threat -- he says that is exactly his premise and that the sector is likely oversold.

Time frame

over the next several quarters

How sure

3 out of 5

Copy citationCopy link

“But right now, the real benefit here is that with a positively sloping curve, there's this dynamic that hasn't been around for a good while, meaning that you can buy a higher coupon five year and let it roll down to that three year and then sell it and reinvest it if the curve stays relatively stable and positively sloping.”

Episode: Why The 30-Year Treasury Lost Its Biggest Buyers | David Busch on Why High Yields Are Attractive Right Now

at 27:48

Busch wants Treasury duration in the three-to-five year part of the curve -- he would step out to three-to-seven, but certainly not out to twenty or thirty years -- and gives two reasons. The first is carry and roll-down: with a positively sloping curve, a dynamic he says has not been around for a good while, you can buy a higher-coupon five year, let it roll down to a three year, then sell and reinvest, provided the curve stays relatively stable and positively sloping. The second is the role of bonds in a portfolio. He argues that because rates are so much higher, bonds work as a ballast again after a period when stocks and bonds fell together, and that in a catastrophic black-swan event where investors push into the safety of Treasuries, Treasuries would likely outperform anything with a credit component, because investment-grade corporates, high-yield and private credit would all take a drawdown as spreads widen. Higher yields also mean lower total-return volatility, since you are paid to be in the position. He ties the maturity choice to clientele as well: a typical retail investor sits near the roughly six-year duration of the US Aggregate index rather than buying a twenty or thirty year. This is the belly of the curve specifically, not the long end, whose natural buyers -- insurers, public pensions and sovereign wealth funds -- he says have rotated into private credit and investment-grade corporates and may need a catalyst to come back.

What they own

Says on air that the three-to-five year range is where he looks for the portfolios he runs as CIO, and certainly not out longer than 20 to 30 years.

How sure

3 out of 5

Copy citationCopy link

“As long as earnings support these valuations, I think that we'll continue to trend in a similar direction.”

Episode: Why The 30-Year Treasury Lost Its Biggest Buyers | David Busch on Why High Yields Are Attractive Right Now

at 42:35

Asked for his view on the overall market, Busch says it should continue to trend in a similar direction as long as earnings support these valuations. His reasoning is that what is driving equities today is earnings rather than passive flow: the rise of AI technology and the hyperscalers has kept corporate earnings growing, the companies participating in the AI infrastructure buildout have had stable earnings, and that is why equities have been able to rise at the same time as yields, which he says a pure discounted-cash-flow reading would not have predicted. The condition does real work in his answer. He says the telling thing will be whether those hyperscalers and infrastructure builders start to show margin compression or miss earnings expectations, in which case he expects volatility, and he expects higher borrowing costs to compress margins across the board as existing debt is refinanced and new debt taken on. He thinks the market's focus will shift from capex itself to the monetization of that capex and whether it supports continued earnings growth. He also flags that rates are here to stay, higher for longer, that equity risk premiums could compress and PE multiples get harder to justify as the hurdle rate rises, and that leveraged sectors -- financial companies, insurers, banks, potentially REITs -- are the most susceptible to the current rate environment.

Only if

Holds only as long as corporate earnings continue to support current valuations; he expects volatility if the hyperscalers and the companies building out AI infrastructure show margin compression or come in below forecast earnings.

Not sure

Yes, said with doubt

How sure

3 out of 5

Copy citationCopy link

“So to two, three reasons why I like long bond. So one is that I think that we are in a tremendous, tremendous equity bubble and that historically speaking when that declines, Fed cuts rates, there's a flight to safety. And long bond is where you get the most bang for your buck.”

Episode: Hawks Take Flight | Joseph Wang on Fed’s Hawkish Hike and Warsh’s Gameplan

at 16:45

Wang says he likes the long bond, which he puts at about 5.35 percent on the day of the hawkish September meeting. His main reason is that he believes US equities are in a tremendous bubble, and that historically when such a bubble declines the Fed cuts rates and there is a flight to safety, in which the long end gives the most upside per dollar; bought for cash or financed, it is held with positive carry. He gives a second, low-weight scenario in which a genuine AI revolution pushes unemployment up and produces technology-driven deflation, which would also be good for long bonds. He adds that he feels much better about the bet after this meeting, because Chair Warsh showed he is serious about getting inflation under control rather than moving the goalposts on the inflation index, and that he thinks current inflation of roughly three and a half percent is largely energy, so excluding that the economy would be on its way to 2 percent. He says he can be totally wrong.

Not sure

Yes, said with doubt

How sure

4 out of 5

Copy citationCopy link

“So I have a low conviction view that MPC estimates are too high, but I have a high conviction view that Lam Research estimates are too low. Like I would say it's almost criminal how low these estimates are on Lam.”

Episode: Hawks Take Flight | Joseph Wang on Fed’s Hawkish Hike and Warsh’s Gameplan

at 54:24

Farley argues consensus estimates for Lam Research are far too low. The company most recently guided to roughly 8.2 billion dollars for the quarter now under way, but the forecast for the following year is about 9.6 billion, which is roughly 17 percent growth and about the company's average rate. He expects a much larger step up in demand for its machines, because the buyers are not only logic fabs such as TSMC and Intel but memory makers -- SK Hynix, Samsung and Micron -- which he puts at 30 to 40 percent of the mix and rising, and which are doing everything they can to add capacity. He would be the highest estimate on the board at 10.5 billion and might go to 11 billion, while cautioning that he does not expect Lam to behave like Micron, which he regards as a genuine commodity. He labels this a high-conviction view and contrasts it with his low-conviction view that Marathon Petroleum estimates are too high.

Time frame

September 2026 into September 27

How sure

4 out of 5

Copy citationCopy link

“Especially if you are afraid of inflation, you can just think about 30-year tips. It's 3.1% real yield. That's historically quite high. You haven't seen real yields that high for a couple of decades.”

Episode: Hawks Take Flight | Joseph Wang on Fed’s Hawkish Hike and Warsh’s Gameplan

at 16:45

For an investor who is afraid of inflation, Wang points to 30-year TIPS as the version of his long-bond view that carries no inflation risk. He puts the real yield at about 3.1 percent and calls that historically quite high, saying real yields have not been at that level for a couple of decades. He offers it as an alternative way to own long duration rather than as a separate forecast, and does not state a position in it.

How sure

3 out of 5

Copy citationCopy link

“But I think you can just have us, if we're sitting here in three months, six months having the same debate, I think the refiners are gonna be higher.”

Episode: Hawks Take Flight | Joseph Wang on Fed’s Hawkish Hike and Warsh’s Gameplan

at 49:54

Wiethe argues refiners are not at peak earnings. Refining capacity is shut in behind the Strait of Hormuz, China has been hoarding its refined product rather than exporting it as it normally does and is now back in the market buying crude, and stockpiles of refined products are smaller than crude stockpiles and have been drawn down -- so the crack spread can stay wide rather than spike and immediately mean revert. Against Farley's argument that refining is cyclical and will revert, he says that case assumes the conflict is done by the end of the year, which he doubts. He is explicit that this is not a call for 10 dollar diesel or for another record EBITDA spike: he expects a plateau at current levels rather than a rapid decline, in which the refiners keep printing, and says that if the same debate is running in three to six months the refiners will be higher. He states on air that he owns Marathon Petroleum and Valero.

Time frame

three months, six months

Only if

Assumes the Middle East conflict and the Strait of Hormuz disruption persist rather than resolving by the end of the year, keeping refined-product capacity shut in and crack spreads wide.

What they own

Says on air that he owns Marathon Petroleum and Valero.

How sure

4 out of 5

Copy citationCopy link

“I think the projects and the opportunities at the moment are amazing.”

Episode: Hunting for Value in Mining Stocks Amid Soaring Metals Prices | Freddy Brick | Muddy Waters Capital

at 34:01

Brick explicitly declines to take a view on where gold, silver or copper prices go next, but is clearly constructive on the junior mining opportunity set itself: more than a decade of capital flight from the sector following the post-2011 super cycle left many economically sound projects undervalued and under-covered, starved of both talent and financing. He now expects an M&A wave as cash-rich majors and mid-tiers, who have struggled to replace reserves through new discovery, look to acquire the projects that quietly advanced toward permitting or construction while the sector was out of favor.

Not sure

Yes, said with doubt

How sure

4 out of 5

Copy citationCopy link

“I have long rice. It will be for the next couple of years.”

Episode: Why Farmland With Centuries of Water Is the Ultimate Scarce Asset | John Farris of LandFund Partners on Super El Niño, Food Security, and the Sources of Alpha in Farmland

at 52:34

Farris, an agricultural economist and farmland fund manager, argues rice is the single crop most exposed to upside price risk right now: it is the most-consumed calorie in the world, global food stocks have thinned to roughly 70 days of supply (down from 110 a decade ago), and a forecast Super El Nino is expected to bring drought to major monsoon-dependent rice-growing regions, particularly India, alongside South America. Combined with California's structural exit from rice growing due to water constraints -- shifting production toward less-irrigated regions -- he expects the supply squeeze to keep rice prices elevated for years, not just this season.

Time frame

next couple of years

Only if

Conditional on the forecast Super El Nino materializing as expected and disrupting monsoon-dependent rice production in India and other major growing regions, against an already-thin global food-stock backdrop (roughly 70 days of supply).

How sure

4 out of 5

Copy citationCopy link

“Intel looks like it's bottoming. Oracle looks like it's bottoming”

Episode: I’m Insanely Bullish on Bonds” | Jared Dillian on Copper, Bonds, Semis, and The Awesome Portfolio

at 8:00

A narrow, low-conviction technical observation rather than a thesis. Going through the top 50 S&P charts, Dillian found Oracle — along with Intel — among the very few that look like they are bottoming. He immediately qualifies it: he is seeing a lot more charts rolling over than basing, so this is an exception inside a broadly deteriorating tape rather than a call on Oracle's business. No fundamental argument is offered and no position is disclosed. Note that elsewhere in the same conversation he is bearish the AI capex cycle that Oracle is levered to, so this row should be read as a chart read, not an endorsement of the AI trade.

Not sure

Yes, said with doubt

How sure

2 out of 5

Copy citationCopy link

“So I'm cautiously bullish here. There really isn't support until 3,900 or 4,000, but that doesn't mean it's going to get there.”

Episode: I’m Insanely Bullish on Bonds” | Jared Dillian on Copper, Bonds, Semis, and The Awesome Portfolio

at 18:03

Gold rallied 15% in a month, then broke back down through its 200-day moving average at 4,500 after Warsh's Jackson Hole speech, which Dillian says has the amateur technicians calling for a retest of 4,000. He is dismissive of that signal — his view is that once price has crossed the 200-day several times the level loses its significance — and he doubts the retest happens, while acknowledging there is no real support until 3,900-4,000. The bigger point is the upcoming catalyst: with payroll prints deteriorating and 55,000 expected, a genuinely ugly number (he floats -50,000 or -100,000) would reverse the whole trade and in his phrase make Warsh 'the best gold salesman of all time', because the Fed would be forced dovish.

Not sure

Yes, said with doubt

How sure

2 out of 5

Copy citationCopy link

“I'm insanely bullish. I think 5.2, 5.3 on 30 years is an incredible deal. I think 4.7 on 10s is an incredible deal.”

Episode: I’m Insanely Bullish on Bonds” | Jared Dillian on Copper, Bonds, Semis, and The Awesome Portfolio

at 0:34

Dillian argues the bond bear market is driven by sentiment, not fundamentals — he calls it 'a mind virus' in which the market has convinced itself deficits and inflation are out of control, echoing the late-1970s talk of bonds as 'certificates of confiscation'. His counter is that inflation is not high and is still falling, while the rest of the data is deteriorating: a 55,000 payroll estimate, a Chicago PMI ten points below expectations, weak JOLTS and a soft ISM. On the supply worry he concedes the $2tn deficit is a scary absolute number but points out it is only 6% of GDP, versus 12% in 2010 when auctions still cleared with bid-to-cover of three or better. His core point is that everyone measures bond supply and nobody measures bond demand: in a genuine risk-off event with equities down 20%, buyers would show up and yields would fall. He notes a secondary pressure he does not dispute — AI-related corporate issuance, such as a $40bn Google bond deal, crowding the market in the way Volcker described. He is long in size and treats it as a multi-year position, not a trade.

Time frame

three to five years

What they own

Says he personally moved a huge portion of his money into bonds in the last month and will hold for three to five years.

How sure

5 out of 5

Copy citationCopy link

“I own some of them. The stocks, I think, are reasonably cheap.”

Episode: The Certainty Premium: What’s Driving the Tech Sector Divergence? | Jeff Keller

at 28:07

Keller agrees the easy, outsized-return era of simply owning mega-cap tech (he notes Amazon is up roughly 10x since 2016-17) is largely behind us, given the hyperscaler cloud market has fragmented from three dominant vendors to nine or ten scaled compute providers, adding competition and uncertainty. But he still owns some of the names, considers the stocks reasonably priced given growth, and thinks investors can still make a respectable 15-20 percent a year from here, just with a very different risk/reward profile than the prior golden period.

Not sure

Yes, said with doubt

What they own

States he personally owns some of the mega-cap tech / former hyperscaler-trade names under discussion ('I own some of them').

How sure

3 out of 5

Copy citationCopy link

“I'd argue that the real AI winners are CrowdStrike and some like that rather than maybe some of the semis.”

Episode: The Certainty Premium: What’s Driving the Tech Sector Divergence? | Jeff Keller

at 4:09

Keller splits software into application software, which he's cautious on, and infrastructure/cyber, where he's more constructive because demand is usage-based and benefiting from enterprises racing to defend against new AI-enabled threats. He specifically names CrowdStrike as one of the 'real AI winners,' arguing it has more visible AI-driven demand than some semiconductor names, though he flags that cyber multiples generally (25-30x revenue for large names) have already run a long way and aren't themselves immune to disruption risk.

Not sure

Yes, said with doubt

How sure

4 out of 5

Copy citationCopy link

“I think we're still moving upwards on an S curve with regards to AI adoption.”

Episode: The Certainty Premium: What’s Driving the Tech Sector Divergence? | Jeff Keller

at 1:51

Keller argues the mid-2026 volatility in AI-related stocks, including the Korea-linked deleveraging event, was a normal shakeout within an ongoing structural adoption trend rather than the end of the mega trend. He points to continued double-digit-to-20-percent-plus growth at Apple Services, Google Search and Meta as evidence the underlying AI-driven S-curve of adoption is intact, and separately argues political headwinds around data-center buildout -- labor, permitting, state-level moratorium threats around the midterms -- are overstated and that data-center construction starts will continue.

How sure

4 out of 5

Copy citationCopy link

“I think it's likely that the low multiple names across the AI trade in general are going to do better, even if they're at different cycles of earnings growth.”

Episode: The Certainty Premium: What’s Driving the Tech Sector Divergence? | Jeff Keller

at 10:22

Keller draws a contrast, a point he credits to Gavin Baker, between AI-trade names that all trade off the same underlying factor -- hyperscaler capex and lab ARR -- but at hugely different multiples. He singles out semiconductor bellwethers like NVIDIA and Micron as names that are not extrapolating an optimistic future, since management is explicitly signaling 'we're at peak,' which he reads as a low-multiple, contrarian-favorable setup relative to higher-multiple, more speculative corners of the AI trade.

How sure

3 out of 5

Copy citationCopy link

“I want to be a buyer when people are getting margin called and liquidated.”

Episode: The Certainty Premium: What’s Driving the Tech Sector Divergence? | Jeff Keller

at 52:32

Keller reads the mid-2026 deleveraging event concentrated in Korea -- large numbers of margin calls and heavy hedge-fund and retail forced selling -- as a technically-driven washout rather than a fundamental deterioration, since he thinks the underlying secular trend (AI-linked demand) is 'still going pretty well' for the two large Korean companies at the center of it. He treats forced-liquidation-driven selloffs like this as a buying opportunity.

Only if

Opportunistic: looks to add exposure specifically during margin-call/deleveraging-driven selloffs in Korea, which he reads as technical rather than a change in the underlying secular (AI-linked) demand trend.

How sure

3 out of 5

Copy citationCopy link

“we could get to these $150, $200 a barrel forecasts.”

Episode: The Commodity Bull Market Is Broadening | Jim Wiederhold on Copper, Grains, and Bloomberg Commodity Index

at 49:20

Wiederhold says the 'jawboning' that has kept oil prices from spiking further -- talk of an imminent US-brokered deal, plus increased North American production and China/SPR inventory drawdowns -- has worked for about three months, but he sees signs both sides in the Iran conflict are digging in and that no real deal is coming. If the Strait of Hormuz disruption persists or escalates further (more Red Sea/tanker attacks, no negotiated resolution), he thinks oil could reach the $150-200 per barrel levels that some analyst models point to if the Strait stays effectively closed for six months -- which is roughly where the conflict now stands.

Only if

Conditional on the Strait of Hormuz disruption/Iran conflict persisting or escalating further without a negotiated resolution; he notes the jawboning that has suppressed prices so far has worked for about three months but sees signs that is breaking down.

Not sure

Yes, said with doubt

How sure

3 out of 5

Copy citationCopy link

“as soon as you see big pickups in central bank buying, that's usually pretty good indicator that the price is going to go higher.”

Episode: The Commodity Bull Market Is Broadening | Jim Wiederhold on Copper, Grains, and Bloomberg Commodity Index

at 18:00

Wiederhold acknowledges the standard bear case on gold (it's a non-yielding asset) but says his own read, from watching the data over multiple decades, is that a pickup in central bank purchases is historically a good leading indicator for higher prices. He points to the latest World Gold Council survey, where a record share of respondents (over 40 percent) said they plan to add to holdings over the next 12 months after the recent pullback in price, and reads that as central banks treating the pullback as a buying opportunity.

Only if

Reads a pickup in central bank gold buying (per the latest World Gold Council survey, where a record share of respondents said they'd add to holdings over the next 12 months) as a signal for higher prices going forward.

Not sure

Yes, said with doubt

How sure

2 out of 5

Copy citationCopy link

“I'm more bullish personally on silver than I am on gold.”

Episode: The Commodity Bull Market Is Broadening | Jim Wiederhold on Copper, Grains, and Bloomberg Commodity Index

at 16:22

Farley argues silver's investment case is more analyzable than gold's because its demand drivers (solar panels, AI data-center wiring, and broader industrial/electrification use) are concrete and persistent, unlike gold, which he says lacks a reliable one-to-one relationship to debasement/debt narratives -- citing gold's 19-year bear market from 1981-2000 even as US debt and the fiat system exploded in growth. He's careful to note this isn't a bearish call on gold, just a preference for silver's more analyzable, industrial-demand-driven setup.

Not sure

Yes, said with doubt

How sure

4 out of 5

Copy citationCopy link

“I like the fundamental story behind the industrial metals and the ones that are industrial adjacent like silver.”

Episode: The Commodity Bull Market Is Broadening | Jim Wiederhold on Copper, Grains, and Bloomberg Commodity Index

at 18:00

Wiederhold ties silver's case to the same structural demand growth he sees across industrial metals: electrification, the renewables buildout (roughly 60 percent of silver demand is industrial, concentrated in solar), and AI data-center wiring, alongside supply that is largely a byproduct of gold, copper and zinc mining and therefore slow to respond even to sharply higher prices. He frames this alongside his broader commodity-supercycle view, naming silver as one of the areas where he currently has the most constructive fundamental view.

How sure

3 out of 5

Copy citationCopy link

“So that could lead to potential issues and corn price appreciation in the future.”

Episode: The Commodity Bull Market Is Broadening | Jim Wiederhold on Copper, Grains, and Bloomberg Commodity Index

at 26:43

Wiederhold sees continued upside for grains, particularly corn and wheat, driven by supply-side tightness rather than demand. High fertilizer costs -- which spiked after the Strait of Hormuz disruption, since a large share of fertilizer inputs move through that region -- are pushing Midwest farmers to apply less fertilizer, which he expects to show up as reduced yields and potential future price appreciation in corn. He separately flags wheat: with a large share of Ukraine/Black Sea wheat shipments exposed to renewed attacks on grain-carrying vessels, similar to the Red Sea/Hormuz disruptions, he expects that to be a further source of upward pressure on wheat prices specifically.

Only if

Conditional on farmers continuing to economize on fertilizer application (reducing future corn yields) and on continued disruption to Black Sea wheat shipments.

How sure

3 out of 5

Copy citationCopy link

“It just makes sense to run just a constant overweight within your portfolio”

Episode: Time to Reduce Equity Risk: Why Underappreciated Macro Risks Could Derail the Bull Market | Warren Pies

at 47:52

Pies has held this position since 2022 and says he does not know what would change it. The argument is portfolio-structural rather than a call on the oil price: in an inflationary regime bonds and stocks move together, so trouble in the equity book tends to originate in the bond book, and conventional diversification stops working. Energy-based assets are the one thing that reliably buffers an equity portfolio when geopolitics hits — he cites the Iran war as a correlation-of-one event in which energy equities nonetheless spiked, some names printing their best days ever while everything else printed its worst. He adds that the case strengthens if the Fed is now reacting to oil prices, because the same shock that forces a hawkish Fed is the one energy hedges. He notes energy is only about 3.5–4% of the S&P, which is what makes a standing overweight cheap to carry.

What they own

HELD — structural overweight in the 3Fourteen RAA fund: the S&P energy weight plus a separate energy sleeve held within alternatives, described as a perpetual overweight.

How sure

5 out of 5

Copy citationCopy link

“I think that the data center ROI question was basically solved this earning season”

Episode: Time to Reduce Equity Risk: Why Underappreciated Macro Risks Could Derail the Bull Market | Warren Pies

at 38:30

Pies argues the return-on-investment question that hung over hyperscaler capex was answered by the most recent quarter, citing Microsoft, Amazon and Google's results as conclusive. His mechanism is that a data centre's total cost can be modelled, and GPU rental rates have risen this year — which he stresses is not supposed to happen in a buildout — so as data centres re-contract at those rates it implies a stair-step higher in cloud and data-centre revenue. He supports this with proprietary GPU availability tracking: a spike in his 'fast availability index' was almost entirely H200, whose rental pricing had drifted implausibly close to Blackwell, and Blackwell B200 availability remains at zero, which he reads as demand for the GPU everyone actually wants being undiminished. He is candid that it is 'never settled' and that the technology could run out of gas, but treats the compute spend as ratified by the technology so long as one believes the tech itself. CONTESTED: Kedrosky, on Meb Faber two days later, argues the opposite from the financing side.

How sure

4 out of 5

Copy citationCopy link

“you probably need a higher oil price once we get to that phase of the conflict if there's no resolution”

Episode: Time to Reduce Equity Risk: Why Underappreciated Macro Risks Could Derail the Bull Market | Warren Pies

at 50:23

Pies notes crude has been sitting above $90 for longer than anyone expected — the market had assumed the war would be over and prices normalised by now — and that the SPR release which has been absorbing the shortfall is nearly finished. If the Strait of Hormuz settles into a genuinely different flow pattern with no resolution to the conflict, the marginal barrel has to come from somewhere else at a higher price. He treats this as the next phase of the story rather than a present call, and it is the same oil strength he argues will keep political pressure on the Fed via gasoline and diesel prices.

Only if

Requires no resolution to the conflict and a sustained change in Strait of Hormuz flows, once the SPR release is exhausted.

Not sure

Yes, said with doubt

How sure

3 out of 5

Copy citationCopy link

“healthcare was going to be that next place you go”

Episode: Time to Reduce Equity Risk: Why Underappreciated Macro Risks Could Derail the Bull Market | Warren Pies

at 41:51

Pies says healthcare has been a house sector preference for some months. The reasoning runs through his view that the AI labs have a public-opinion problem — that their spokesmen are losing hearts and minds and that the industry has been sold to the public primarily on job losses — and that the labs urgently need tangible, visible wins to shift sentiment ahead of political risk. Healthcare is where he expects those wins to be found and promoted, citing AI-assisted drug development as the kind of result the industry will push front and centre. The call is therefore as much about where capital and narrative attention rotate as about healthcare fundamentals.

How sure

3 out of 5

Copy citationCopy link

“semis are kind of the core of this market right now”

Episode: Time to Reduce Equity Risk: Why Underappreciated Macro Risks Could Derail the Bull Market | Warren Pies

at 35:41

Pies treats semiconductors as the proper leadership group of this bull market, now roughly 18% of S&P 500 market capitalisation, and as the single best real-time read on the market's belief in AI. He describes the Q3 divergence — software and services up about 20% while semiconductors and semi equipment fell about 7.5% — as technically rather than fundamentally driven: since the April model announcement, pods have expressed the AI trade as long semis paired with 'funding shorts' in software, staples, financials and rate-sensitive names, and the recent move is that pairing unwinding rather than a fundamental rotation. His diagnostic claim is strong and testable: tell him what SMH did on a given day and he can tell you what the rest of the heat map looks like. For the bull market to resume he wants semis and hyperscalers rising together rather than negatively correlated. CONTESTED: Kedrosky is bearish semiconductors on 2028 supply over the same week.

How sure

4 out of 5

Copy citationCopy link

“And the other thing that we find very exciting nowadays is better for you healthy snacking foods. So as I mentioned, we transacted on a meat snack business. So where the trends seem to be going now, and we think are permanent and moving for a long period of time, is high protein, high fiber, healthier in that manner.”

Episode: Ex-Goya COO on the $1.4 Trillion Family Business Opportunity in Three Consumer Sectors | Andy Unanue

at 17:01

Unanue argues that demand for "better for you" snacking -- high protein, high fiber, cleaner labels -- is a durable structural shift rather than a passing fad, saying the trend is "permanent and moving for a long period of time." The mechanism he gives is GLP-1, GLP-2 and GLP-3 drugs: people eat less caloric volume but still need protein and fiber, which pushes demand toward smaller pack sizes, less sugar and cleaner ingredients. He expects the large CPG companies to keep participating in the category by acquisition and by reformulating their mega brands, because in his reading of their quarterly calls and annual reports their growth is no longer coming from the sugary snacks and drinks that drove it historically. His own firm plays the theme mainly through co-manufacturing and through established family businesses rather than by betting on individual upstart brands, which he calls a binary, asymmetric risk he will not take -- for every visible winner he says there are a hundred that fail. The view is about the demand trend in the sector, not a call on any listed company.

Time frame

permanent and moving for a long period of time

Not sure

Yes, said with doubt

What they own

Says on air that his firm transacted on a meat snack business and plays in the better-for-you ecosystem through co-manufacturing and private label, and that it will not take the risk of backing an individual fast-growing upstart brand.

How sure

3 out of 5

Copy citationCopy link

“Big crash, most likely the crash is going to lead to a three, four, five-month strong upside move, most likely.”

at 5:51

The most extreme decline Berg is buying: the KOSPI fell 43.93% in 27 days into the 29 July low, including 22.68% over three days. He treats that as a genuine crash, and his historical read is that crashes are followed by multi-month rallies — he uses the 1987 analogue, where the S&P fell about 46%, bottomed, then came back within a percentage point of the low two months later before recovering. Position is small at 5% precisely because he expects the test. He gives an unusually falsifiable line: a close below the July low would be normal, but a trade below the 29 July INTRADAY low — roughly 19% below the price at recording — would tell him this was only the first leg of a bear market and he would be out.

Only if

a trade below the 29 July intraday low (~19% below the level at recording) invalidates the thesis and he exits

What they own

5% long Korea. Explicitly sized small because he expects a retest.

How sure

3 out of 5

Copy citationCopy link

“10 percent long the Nasdaq 100, 20 percent long the Sox”

at 1:27

The weakest leg of the long book and the one he flags hardest. The Nasdaq 100 made new lows into July while the S&P did not, is around 3% below its 2 June high, and gapped up off the low in a way he reads as possibly exhaustive rather than breakaway — his test being that a genuine breakaway gap should be followed by a straight-up market, and this was followed by a trading range. He remains 10% long while explicitly watching for a short-term top.

Only if

watching for a short-term top; a failure to follow through would confirm the gap as exhaustive

Not sure

Yes, said with doubt

What they own

10% long the Nasdaq 100.

How sure

2 out of 5

Copy citationCopy link

“If somebody asks you where to be positioned, the answer is to be positioned long now, not short, not out of the market”

at 51:57

Berg covered a short at the 29 July low and went long, and is now 100% long unleveraged — roughly 50% S&P via SPY, 20% SOX, 10% Nasdaq 100, 10% Russell 2000, 5% S&P MidCaps, 5% Korea. The case rests on signals from March-April rather than on July: over 30 buy signals fired between 31 March and 27 April, whose median projections run to 8,900-10,600 on the S&P, and he argues the market is still tracking in line with the historical paths of those signals. July adds a positive divergence — the S&P held above its June low while the Nasdaq 100 and the Philadelphia Semiconductor Index made new lows — plus retail net selling of equities at its greatest since 2022, and a breakaway gap through the prior all-time high, which follows through to the upside about 80% of the time. He is explicit this is a trading posture, not a multi-year call: 'for me to say, it's going to last for a couple of years into a major bull move, I can't say that at this point.'

Not sure

Yes, said with doubt

What they own

100% long, unleveraged. Covered a short on 29-30 July, the day of and day after the low.

How sure

4 out of 5

Copy citationCopy link

“We're positioned nearly 10 percent long the Russell 2000”

at 1:27

Part of the deliberately diversified 100%-long book. The Russell carries specific signal weight in Berg's April work: off the March low it posted its greatest 10-day gain in five years — larger than off the April 2025 low and larger than off the COVID low — and separately traded above its 5-day moving average for 16 consecutive days starting from the day of the low itself. Both fired as buy signals with historical median projections well above current levels, and he says the index remains in line with those historical paths.

What they own

Nearly 10% long the Russell 2000; 5% long S&P MidCaps.

How sure

3 out of 5

Copy citationCopy link

“It's a recovery rally, but you see the gap, you had a gap right here, and you have a possible cyclic top, so we're on the lookout for the possibility of a short-term top, but we're still long.”

at 5:51

His largest single-index position after the S&P, at 20% long — and deliberately the most damaged asset. The Philadelphia Semiconductor Index had run over 100% from its March lows to a June peak, then fell 28.73% into the 29 July low. He bought the wreckage on the view it was oversold on panic selling and should at least bounce. The caveat is doing real work: crash lows are normally tested rather than V-bottomed, he reads the gap off the low as possibly exhaustive rather than breakaway because it has not followed through, and a Montgomery cycle date falls in the current week. So he expects a better-than-random chance semis trade back toward the lows within a month or two while he stays long through it.

Only if

expects a better-than-random chance of a retest of the July lows within one to two months, while staying long

What they own

20% long the Philadelphia Semiconductor Index.

How sure

3 out of 5

Copy citationCopy link

“that would be replaced with enterprises using these open weight models and the companies that would host that would be these hyperscalers.”

at 5:20

Part of Dean Pernas's 'commoditize your complements' argument: open-weight models are bullish for every layer of the AI stack except the frontier labs. If enterprises no longer pay frontier-lab token prices, usage rises (Jevons paradox) and the economics migrate to whoever controls the GPUs and hosts the inference — the hyperscalers. He names Microsoft, Google and AWS as the beneficiaries, explicitly setting Meta aside because it is not in the cloud business. He also rates Microsoft's and Google's AI capital allocation as more measured and rational than Meta's.

How sure

3 out of 5

Copy citationCopy link

“we believe their solutions are pretty much tailor-made for the rise of scale across in data centers”

at 19:18

The expression of Dean Pernas's data-centre interconnect theme. Smart Optics builds the hardware that converts data to light for transmission across fibre between data centres, plus the software to modulate it accurately, and modifies its hardware to interoperate with Broadcom, Cisco and Arista. He describes the software as best in class and says 'everyone in Silicon Valley knows them.' A smaller company, listed on the Oslo exchange for around four years.

What they own

Held in the Pernas portfolio.

How sure

3 out of 5

Copy citationCopy link

“And that benefits cloud providers because now the economics translates to whoever controls the GPUs. So GPU rental prices are going to increase as a result.”

at 4:30

Part of Dean Pernas's 'commoditize your complements' argument. He expects open-weight models to cut enterprise inference costs by around 90%, which raises usage rather than reducing spend, and pushes the economics to whoever owns and rents the GPUs — driving GPU rental prices higher. AWS is named alongside Microsoft and Google as the hosts that capture this, with Meta explicitly excluded as not being in the cloud business.

How sure

3 out of 5

Copy citationCopy link

“bandwidth needs are going to have to expand almost 15 times to accommodate that. And we believe that that is another theme that we believe is very attractive.”

at 9:54

Dean Pernas's second theme. He expects single-site data centres of one to three gigawatts to top out soon for two reasons: you physically cannot bring that much power to one site, and community pushback over emissions and water is producing state-level moratoriums. Hyperscalers will therefore have to bridge geographically separate sites into one training or inference fabric, which he estimates requires bandwidth between data centres to expand roughly fifteen-fold.

How sure

4 out of 5

Copy citationCopy link

“So you're getting a huge discount on a player that may be better positioned than some of these larger rivals to do very, very well given this platform shift.”

at 27:04

Deiya Pernas's cybersecurity pick, built on a claimed regime change in the threat model. Roughly 80% of recent attacks were identity-based, but frontier models now let attackers read thousands of lines of code and find exploitable vulnerabilities cheaply, so traditional hacking becomes the live problem again and CISOs are 'losing their minds.' That shifts spend toward exposure management — knowing where your assets are, which vulnerabilities actually matter, and fixing them — where he says Tenable has led since 2022 and has an Anthropic partnership. The valuation gap is the crux: Tenable at roughly 4x EV/sales against CrowdStrike and Palo Alto both north of 20x, with a large installed base of legacy vulnerability-scanning customers to upsell. He expects Tenable's growth to accelerate.

How sure

4 out of 5

Copy citationCopy link

“if AI can create the brain for robotics, given that the hardware is there, it's that hockey stick growth is pretty much incalculable over the next several years”

at 21:04

Dean Pernas's robotics play, and explicitly conditional on the robotics thesis landing. His argument is that humanoid hardware already exists but the 'brain' does not, and that high-fidelity physics simulation lets AI supply it — in which case humanoid robotics 'might be the largest industry in 10 years time.' Vishay Precision Group makes the strain gauges and sensors that go into humanoid hands, and he says it already has three of the largest customers in the space (he names Tesla and Figure), though still in prototype mode. He flags the confusion risk explicitly: this is VPG, not Vishay Intertechnology.

Time frame

next several years / 10 years

Only if

conditional on AI being able to create the 'brain' for robotics via high-fidelity physics simulation

What they own

Held in the Pernas portfolio; first discussed June 2025, tripled to June 2026, then roughly halved over the past month.

How sure

4 out of 5

Copy citationCopy link

“So that theme, we believe, is going to continue to sustain itself for the next three to five years until more alternatives can be bought to the market. So we believe that's still the strongest theme out there.”

at 9:54

Dean Pernas's strongest-conviction theme. US grid demand has been roughly flat near 4,000 TWh for twenty years as efficiency offset population growth, and AI is poised to take ~20% of US grid demand (~100GW) by 2030. The grid cannot deliver it: transmission is 'a two-lane highway which should be an eight-lane highway,' and interconnection queues run 18-24 months. So hyperscalers must bring their own power, and the US has natural gas pipelines almost everywhere to plug into. He expects this to persist three to five years until alternatives arrive, and argues open-weight models only accelerate it.

Time frame

next three to five years

How sure

5 out of 5

Copy citationCopy link

“Not a lot has to go right for there to be re-rating, but there is some hair on it with leverage.”

at 1:03:43

A cheap, hairy re-rating candidate rather than a quality compounder. Deiya Pernas values Paysafe on the sum of its parts: the digital wallets subsidiary alone accounts for 60-70% of enterprise value and he argues it could be sold at 9-10x earnings as a portable business. Operationally it is a turnaround — paying down debt, tightening operations, still growing at high single digits organically after asset sales. He also likes the gaming and prediction-markets orientation, a segment that Stripe and Adyen typically avoid, leaving room for a specialised processor. The stated risk is that the company is significantly over-levered.

Only if

re-rating case carries meaningful leverage risk on the balance sheet

Not sure

Yes, said with doubt

What they own

Held in the Pernas portfolio.

How sure

3 out of 5

Copy citationCopy link

“So SEND volumes growing 30%, 40% annually and revenues growing 25%.”

at 59:13

The second cross-border pick, framed as 'a Western Union killer' — a digitally native migrant remittance provider with cheaper fees and a better experience that has eroded Western Union's model. Deiya Pernas argues the moat is mindshare rather than price: migrants weight trust and brand heavily and are not simply shopping for the lowest cost, which lets Remitly's take rate decline only slowly. Send volumes are growing 30-40% annually against 25% revenue growth, and he sees that durability extending five to ten years as remittances keep digitising away from cash. Unlike Wise, roughly 90%+ of Remitly's transactions still move over Visa and Mastercard rails.

What they own

Held in the Pernas portfolio.

How sure

4 out of 5

Copy citationCopy link

“we're highly bullish on Upwork's staying power”

at 35:14

Deiya Pernas's most contrarian position. The market is selling Upwork on fear that AI replaces freelancers; he argues the platform's own data shows weakness only in sub-$500 low-value tasks — 'a lot of Fiverr's business' — while the overwhelming majority of Upwork's volume is higher-value work where small businesses want to talk to a person, often because they do not know what questions to ask. He treats the marketplace as its own economy subject to creative destruction: as one stratum of freelancers is automated away, new needs replace it, and AI makes the remaining experts more productive. Valuation carries the argument: a little over 1x EV/sales and a bit over 6x EV/free cash flow after stock comp, for a business he considers impossible to replicate and still growing — 'I don't know where you find a better bargain than that in this market.'

What they own

Held in the Pernas portfolio; also a heavy user of the platform, with spend growing over the years.

How sure

5 out of 5

Copy citationCopy link

“Leaving aside meta, it would be Microsoft, Google, AWS, the usual suspects.”

at 5:20

Part of Dean Pernas's 'commoditize your complements' argument: open-weight models are bullish for every layer of the AI stack except the frontier labs. Cheaper inference raises enterprise usage, and the economics accrue to whoever controls the GPUs and hosts the workloads — he names Microsoft, Google and AWS, explicitly excluding Meta as not being in the cloud business. He separately rates Google's AI capital allocation as more measured and rational than Meta's.

How sure

3 out of 5

Copy citationCopy link

“And so we just think that the trends are the trends are just as strong with with Capstone. However, the valuation is much more attractive.”

at 13:15

The main expression of the bring-your-own-power theme. Capstone makes micro-turbines (50kW to 2-3MW) against Bloom Energy's fuel cells on similar dimensions; Bloom trades at 15-20x revenue while Capstone is around 3x, and Dean argues the tailwinds are just as strong. The differentiator is time to market — one to three months versus 18-24 months for large turbines — which matters because 'if you don't have energy, you don't have anything.' The company runs at roughly 10% of a ~1GW/year three-shift capacity, and he sees three routes: smaller hyperscaler data halls of 20-30MW, thousands of edge data centres moving from CPUs to GPUs, and manufacturing that can no longer rely on the grid.

What they own

Held in the Pernas portfolio; initiated 2025-08-05, up ~545% since despite an ~50% drawdown from the May 2026 peak.

How sure

4 out of 5

Copy citationCopy link

“So those are two massive trends that we think are going to drive favorable unit economics for Stabilis.”

at 17:13

A recently initiated speculative position (~$75m market cap, sized 1-3%). Stabilis runs a 'virtual pipeline': liquefying natural gas, trucking it cryogenically, storing and vaporising it for data centres with no pipeline access — a long-consolidating niche where it is now one of few providers. Two drivers: a recently signed data-centre contract worth a couple of hundred million dollars over two years, and SpaceX as one of its largest customers, with launches expected to go from roughly 50 to a multiple of that over five years.

What they own

Held as a 'speculative' sleeve position (1-3% weight); ~$75m market cap.

How sure

3 out of 5

Copy citationCopy link

“It was so beaten up. It was such an easy call to make. Everybody is pricing this thing for complete decapitation. And it was very clear that the market was wrong on the probabilities.”

at 30:02

A software name Deiya Pernas argues was mispriced by screening rather than by fundamentals. Sprout screens poorly because growth has slowed while stock-based compensation stays very high, but he argues SBC should be treated like any other variable expense: if SBC as a percentage of revenue falls as growth normalises, GAAP earnings improve sharply on that alone. The market had priced the name for 'complete decapitation' on AI-disruption fear, which he judged wrong on probabilities. Noted on air as up roughly 50-60% since he discussed it in April.

What they own

Held in the Pernas portfolio.

How sure

3 out of 5

Copy citationCopy link

“WISE is another one where they're going to continue to defend their take rate because they have the best infrastructure around and nobody's done the work that WISE has done to build out their infrastructure to plug in to kind of different nations, central banks and be able to move money as fast as they can.”

at 59:13

One of two cross-border money-movement picks. Deiya Pernas argues correspondent banking is antiquated and expensive — around 5% average cost against roughly 2% through fintechs — and that Wise has built proprietary infrastructure plugging directly into national payment systems and central banks, effectively a global ACH that routes around correspondent banking and even around Visa and Mastercard rails. He frames Wise's falling take rate as deliberate loss-leading aggression rather than competitive pressure: 'they're the one causing the storm.' The structural tailwind is M2 growth of 5-6% a year that a payment company with a defensible scrape can compound on.

What they own

Held in the Pernas portfolio.

How sure

4 out of 5

Copy citationCopy link

“I think the more accurate way to think about it now is the e-commerce pie will grow and probably grow by a lot, and that's good for both of them.”

Episode: TIP855: Shopify (SHOP): The Big Winner of Agentic Commerce? w/ Shawn O'Malley & Daniel Mahncke

at 1:01:22

O'Malley no longer sees e-commerce as a zero-sum fight between Amazon and Shopify. With only about 20% of retail happening online, he accepts the argument that removing friction from buying and giving everyone an AI personal shopper will push a larger share of retail spending online, so the whole pie grows, and probably by a lot, which benefits both. He caveats that nobody yet knows exactly how LLMs and agents will change shopping, so either could gain more, or both could lose ground to a competitor that does not exist yet.

Not sure

Yes, said with doubt

How sure

3 out of 5

Copy citationCopy link

“On that note, I've been a bull on agentic commerce since covering some payment and e-commerce companies last year.”

Episode: TIP855: Shopify (SHOP): The Big Winner of Agentic Commerce? w/ Shawn O'Malley & Daniel Mahncke

at 35:15

Mahncke has been bullish on agentic commerce, AI assistants that search, compare and increasingly complete purchases for the shopper, since covering payment and e-commerce companies last year. Little happened for a while and he wondered if he had been too optimistic, but Meta's Muse launch has put it back on many people's radar. He notes Shopify's CFO's distinction that most activity today is LLM-driven discovery rather than agents buying on a budget, that the full vision is closer but not there, and that its timing in the near future is uncertain; he expects agents to matter most for high-consideration purchases rather than routine replenishment.

Not sure

Yes, said with doubt

How sure

3 out of 5

Copy citationCopy link

“And I think Meta is probably the only one that's pretty safe because they play ads when people are looking for entertainment and they're not anywhere in the buying process.”

Episode: TIP855: Shopify (SHOP): The Big Winner of Agentic Commerce? w/ Shawn O'Malley & Daniel Mahncke

at 46:44

Asking what agentic shopping does to the advertising businesses of the big platforms, Mahncke thinks Meta is probably the only one that is pretty safe: its ads run while people look for entertainment, outside the buying process, and people will keep going to Instagram for entertainment whether or not they shop through agents. Amazon and Google, by contrast, face a real threat to their core ad businesses, though he believes they will figure it out. Earlier in the episode he also said Meta has a huge advantage betting on AI.

How sure

2 out of 5

Copy citationCopy link

“But I just think that the market has taken things one step too far with that idea.”

Episode: TIP855: Shopify (SHOP): The Big Winner of Agentic Commerce? w/ Shawn O'Malley & Daniel Mahncke

at 1:06:48

Addressing the bear case that in a world where AI writes software merchants will vibe code their own Shopify, O'Malley says the market has taken the idea behind the so-called SaaS apocalypse sell-off, which hit names like Adobe, CoStar and Intuit, one step too far. Talking to real businesses, he thinks nobody is trying to vibe code essential functions: the savings are negligible next to the risks, a do-it-yourself stack loses the platform's scale benefits such as negotiated payment-processing fees and integrations, and when it breaks on a day like Black Friday the merchant is on their own. He hedges that the view may not age well.

Not sure

Yes, said with doubt

How sure

2 out of 5

Copy citationCopy link

“And so at best, it sounds like you're paying maybe a fair price for what is a very, very good business. And look, I think you can do a whole lot worse than that as an investor generally.”

Episode: TIP854: GE Aerospace: Chris Hohn’s Biggest Bet w/ Daniel Mahncke & Shawn O'Malley

at 1:10:27

Shawn O'Malley, co-host, argues GE Aerospace's earnings may be even more durable than Hermes's, because of contractual and regulatory lock-in and because passenger tastes play no part in engine choice, while demand for engines should keep rising as more people fly. With Ferrari and Hermes having traded above 30-40 times earnings, he thinks a 30x or higher exit multiple is not unreasonable and may prove conservative, and adds that the protected military business could warrant a higher multiple. He concedes the stock is clearly not on sale and that at best you are paying a fair price, but says you can do a whole lot worse than a fair price for a very good business. A mild, hedged positive view on the shares.

Not sure

Yes, said with doubt

How sure

2 out of 5

Copy citationCopy link

“I very much like the idea of having this long-term tailwind of more people traveling and using airplanes and aircraft to go around the world. I think that will be a trend that over the next 20, 30 years' time will be a massive tailwind.”

Episode: TIP854: GE Aerospace: Chris Hohn’s Biggest Bet w/ Daniel Mahncke & Shawn O'Malley

at 1:12:51

Daniel Mahncke expects growing air travel to be a massive long-term tailwind over the next 20 to 30 years, as more people spend time and money on travel and visiting other parts of the world rather than on luxury goods, which means much more spending on flights. He raises it while discussing GE Aerospace and the aviation parts suppliers that benefit from it.

Time frame

next 20, 30 years

How sure

3 out of 5

Copy citationCopy link

“But on the whole, it's sort of this upward trajectory of growth and growth above the cost of capital, which of course increases their, you know, cash flows and future earnings and the trajectory of their stocks.”

Episode: RWH074: True Wealth w/ Matt Ludmer

at 2:06:54

Ludmer's view is a long-horizon case for owning the global equity market rather than any part of it. The mechanism he gives is corporate survival: there are millions of companies, each has a cost of capital, and a company that does not beat its cost of capital does not stay in business, so a portfolio holding hundreds of companies across the globe captures an upward trajectory of growth above the cost of capital that feeds cash flows, future earnings and ultimately the stocks. Some names will be great and some terrible; the aggregate is what he is buying. He frames the path as what he attributes to Nick Murray -- temporary downs and permanent ups -- accepts the market can fall as much as 50 per cent in the short run, and says he will not take a portfolio from a client unwilling to stay ten years, because he needs enough rolls of the dice for the long-term pattern to show through the randomness. The rest of his framework is what he will not do: he argues manager selection, market timing and annual forecasts have poor records, that information is now too widely available for a differentiated view to add value, and that the practical expression is widely diversified index and Dimensional factor ETFs held at the systematic-risk level, stripped of company, industry and country risk. AI does not change the direction for him; his optimism rests on the people inside those companies using it to make them thrive.

Time frame

ten years and longer; he cites 1926-2025 data

What they own

Runs client portfolios in diversified index and Dimensional factor ETFs through his own firm and says he holds a portion of that third pot himself.

How sure

4 out of 5

Copy citationCopy link

“But needless to say, I think I see Airbnb and Reddit similarly to how you see Aritzia as being really compounders with a very long runway.”

Episode: TIP853: Aritzia (ATZ.TO): The Women's Apparel Brand Conquering the US, plus Nike, Lulu, and More w/ Kyle Grieve & Shawn O'Malley

at 6:03

O'Malley names Reddit alongside Airbnb as a holding in the show's intrinsic value portfolio that is not optically cheap, at about 37 times earnings, and that he nonetheless sees as a compounder with a very long runway. Reddit is the exception he flags to their habit of buying at lower multiples than the current one: its growth has been so explosive that it has grown into a lower multiple since they bought it. The implication he draws is the same as for Airbnb and for Grieve's Aritzia: with businesses like that, inaction tends to be the best course of action, a rising share price alone is not a reason to sell, and only a broken thesis is. Beyond the multiple and the growth rate, he gives no company-specific argument for Reddit in this segment.

What they own

Says Reddit is held in the show's intrinsic value portfolio and has grown into a lower multiple than when they bought it.

How sure

3 out of 5

Copy citationCopy link

“So it's my largest position, so it's fair to assume. I'm still very, very bullish on it. And after this research that I did to prepare for this call, nothing's really changed on that front. So I think they can continue increasing revenue in the mid to high teens of the next five years or so.”

Episode: TIP853: Aritzia (ATZ.TO): The Women's Apparel Brand Conquering the US, plus Nike, Lulu, and More w/ Kyle Grieve & Shawn O'Malley

at 1:19:51

Grieve has owned Aritzia since 2020, it is still the largest position in his personal account, and after re-researching it for this episode he says nothing has changed: he expects revenue to keep growing in the mid to high teens for the next five years or so. The growth plan requires nothing the company has not already done, in his reading — keep finding new locations, keep building brand awareness among women who have never heard of the brand, keep selling in boutiques and online. The positioning he calls everyday luxury sits between fast fashion and true luxury: multiple in-house brands under one design house, quality good enough that his wife still wears twenty-year-old leggings from their T&A line. The US is the runway: he cites talking to US-based women who liked the brand once they met it but had never heard of it outside New York, and reads that as open market share. New stores pay back in 12 to 18 months on management's numbers, roughly 1.5 million dollars of capex per boutique, and he expects capex to fall as the giant flagships and distribution centre spending roll off. The balance sheet carries no long-term debt, about 400 million of untapped liquidity and roughly 800 million of cash from operations last year; they repaid the COVID revolver inside the fiscal year and cleared the term loan in cash in 2021 rather than keeping free money on the books. No dividend and only a token buyback, which he thinks is right while the shares have not been cheap, though diluted shares still creep up about 1 per cent a year. On management, CEO Jennifer Wong is a 1987 shop-floor hire who has run every part of the business, and on his maths she clears Buffett's one-dollar test: about half a billion of earnings retained in her tenure against a market cap up more than 10 billion. His own criticisms are that capital allocation is not Wong's primary strength and that insider ownership outside founder Brian Hill, who still holds more than 15 per cent with voting control and takes a one-dollar salary, is thin. On value, with the stock down about a third, a reverse DCF tells him the market is pricing in about 12 per cent revenue growth; his base case of 15 per cent growth, margins edging from 20.6 to 21 per cent and a 15 times EV/EBITDA exit gives him about 215 dollars, which he says is about a 12 per cent return from here, and blending a bull case at 25 per cent probability against a bear at 20 with a 20 per cent margin of safety he gets an intrinsic value of about 148 dollars, which he says leaves the shares at about a 23 per cent discount. He is candid about the bias: he lives where the brand was founded, knows people who have worked there, researched the episode in Aritzia sweatpants, and is sitting on five times his cost. He also keeps the name out of the show's portfolio, because the hosts' record in retail has been poor, while holding it as his biggest personal position.

Time frame

the next five years or so

What they own

Says he has owned Aritzia since 2020, that it is still the largest position in his personal account, that he has trimmed it only slightly earlier this year, that the stock is a 5x on his average cost, and that he has sat through drawdowns of 30 to 65 per cent. Says he is fine keeping it out of the show's intrinsic value portfolio while holding it personally.

How sure

5 out of 5

Copy citationCopy link