AI Is Propping Up the Economy. The Fed Could Break It.

Bitcoin

Across the shows: 7 speakers in other episodes, 3 bullish and 4 bearish calls, past 30 days

• Andrew Sarna · Forthlane PartnersUncertain

  Bitcoin is a volatile, aspirational store of value whose long-term price depends on mainstream adoption.

  Sarna separates the store-of-value role from the technology and thinks Bitcoin will play any store-of-value role for crypto. Short term it will be very volatile, but over ten years the question is whether everyone holds it in brokerage accounts. Banks like Morgan Stanley now offer access, which would have sounded crazy five years ago. If adoption grows, prices probably go higher. If not, stay away. Bitcoin has been resilient through the rate rise since its roughly $60,000 bottom, though it remains speculative.

  “Bitcoin is going to be super volatile in the short term. But if I take a ten year time horizon, it's really a question of ten years from now, is everyone going to own this in their brokerage accounts?”at 35:43

NVIDIA

Across the shows: 17 speakers in other episodes, 17 bullish and 4 bearish calls, past 30 days

• Andrew Sarna · Forthlane PartnersBullish

  Nvidia and semi and memory stocks look very cheap on P/E, suggesting the market doubts AI buildout.

  Sarna points to Nvidia, semis and memory stocks trading at very low P/E multiples. He reads this as the market calling the bluff on whether AI infrastructure will keep being overbuilt, that is, the market doubts the AI capex story. He frames it as a central question for the market: whether this infrastructure build-out continues indefinitely. He also contrasts this with the fervour around private AI valuations.

  “You look at some of these stocks like Nvidia, like the semi stocks or the memory stocks and they look super cheap on a PE basis. And clearly the market is calling bluff.”at 38:06

Anthropic

Across the shows: 18 speakers in other episodes, 8 bullish and 33 bearish calls, past 30 days

• Andrew Sarna · Forthlane PartnersBearish

  Anthropic's planned compute spending looks crazy against its roughly $4B of 2025 revenue.

  Sarna cites a leaked S-1 for Anthropic that shows about $512 billion of planned compute spending next year, set against about $4 billion of revenue in 2025. He says this 'sounds kind of crazy'. He ties it to an index-level overhang: an almost unknowable question about whether AI spending will pay off. The argument is a skeptical one about the gap between AI infrastructure outlays and the revenue of the labs buying that capacity. He does not spell out a full mechanism, so the view is thin.

  “Anthropix S1 gets leaked and they're going to spend $512,000,000,000 on compute next year. And in 2025, they had $4,000,000,000 of revenue, which sounds kind of crazy.”at 38:50

AI infrastructure & data centers

• Andrew Sarna · Forthlane Partners

  Data-center backlash ahead of the election may stop hyperscalers overbuilding, which could lift their stock prices.

  Sarna is worried about the AI supply chain and the hyperscalers funding the CapEx trade. He argues that data centers are becoming more political, and the issue will become hotter as the next election cycle approaches. He notes the general population dislikes data centers and AI, and that this may be the only saving grace stopping hyperscalers from overbuilding and destroying capital, which could be very accretive to their stock prices.

  “But what may actually prevent the hyperscalers from overbuilding is, you know what, data centers are becoming more and more political. It's only going to become a more hot topic issue as we approach next election cycle.”at 10:27

• Andrew Sarna · Forthlane Partners

  AI demand is so strong that memory costs rising about 8x have not slowed hyperscaler capex.

  Memory costs have risen about 8x, yet hyperscalers have not slowed down. Instead they have raised their capex expectations, and Sarna cites a Goldman estimate of $1.2 trillion of spending next year. He uses this to argue that the old principle, that high prices choke off demand, does not hold for AI because demand is so voracious.

  “Memory costs have increased like 8x. And has that slowed the hyperscalers at all? No. They've only increased CapEx expectations.”at 16:36

US rates & the Fed

• Andrew Sarna · Forthlane Partners

  A Fed hike cannot fix oil-driven inflation, so further tightening risks overtightening the economy.

  Sarna was surprised the Fed delivered its first hike. Looked at logically rather than by CPI components, a hike cannot produce more oil, and oil is what is driving inflation higher. He worries the Fed will overtighten into a broadly lukewarm economy, where a slowdown could turn into a growth scare, and says a slowing economy is the only realistic way to bring yields under control.

  “Because if you approach it from a logical perspective rather than just looking at the components of CPI, it's not like a hike is going to print any more oil. And that's ultimately what is driving inflation higher.”at 2:07

US growth & recession risk

• Andrew Sarna · Forthlane Partners

  Beneath solid headline data, the US economy is lukewarm; only AI spending runs hot, masking weakness.

  Green shoots and decent headline strength look fine on the surface, but under the covers the economy is really being driven by the AI economy. Housing was already in poor shape and yields will pressure it further, and manufacturing is only propped up by initiatives. To illustrate AI spending, he cited a stat that Meta alone will spend more on AI capex this year than every other country's defense budget except the US's. Because the rest of the economy is lukewarm, an overtightening Fed could slow it sharply, and any hiccup in AI spending would leave the economy in a bad place, producing a growth scare.

  “But at the same time, once you look under the covers, it's actually really just the AI economy driving the broader economy.”at 2:37

US housing

• Andrew Sarna · Forthlane Partners

  Housing was already weak, and rising yields will pressure the sector even harder.

  When Sarna looks at cyclical sectors, housing was already in pretty poor shape before yields rose. Rising yields are only going to pressure the sector further, which he sees as a broader result of the Fed and yield dynamics behind the economy's lukewarm picture.

  “well, it was already in pretty poor shape. And now we're seeing yields that are only going to pressure that sector even harder.”at 2:49

Middle East & Iran

• Andrew Sarna · Forthlane Partners

  Iran holds maximum leverage until the midterms, so both sides are motivated to reach a deal before then.

  Sarna's read is that Iran's leverage peaks until the midterms and Trump will be motivated to get a deal done. He expects both sides want a deal before then, since after the midterms there is less motivation and speculation that the administration will try to finish the job. He ties the outcome to oil, which he says will dictate where yields go, and admits the timing is hard to call.

  “you know what? The Iranian's max leverage is up until midterms. Trump is gonna be motivated to get a deal done.”at 5:13

Market sentiment, positioning & volatility

• Andrew Sarna · Forthlane Partners

  A decade of zero rates built hidden leverage, so a blowup could trigger a 2024-style deleveraging.

  Sarna says a decade of zero interest rate policy incentivized the whole economy to take on more and more leverage, with pressure building in various pockets. Nobody knows where the leak will spring, and it is easy to end up with a blowup that triggers a major deleveraging event. He cites the 2024 yen carry trade unwind as an example of a pocket of the market with too much leverage hitting its breaking point. The administration will try to manage a soft landing, but that is easier said than done.

  “when you have a decade of zero interest rate policy, well, that incentivizes the entire economy to begin taking on more and more leverage. ... You have pressure building in these various pockets, but you don't know where the leak is going to spring.”at 6:49

US Treasuries & yields

• Andrew Sarna · Forthlane Partners

  Bonds work only as a short-term trade, since unsustainable debt and deficits make them unsafe to hold long term.

  Clients are asking about bonds for the first time in a long while, but Sarna sees them only as a trade. He cites debt-to-GDP and the deficit, and says neither current levels nor the trend are sustainable, so he has a hard time owning bonds long term. He sees a possible short-term bounce in yields, while noting real and nominal yields are at highs not seen since 2004 and 2007.

  “But the problem is, is I view it as only a trade. I don't think you look at the issues with debt to GDP and the deficit and no way can you think that trend is sustainable or current levels are even sustainable. So I have a real hard time owning bonds for the long term.”at 8:51

Gold & precious metals

• Andrew Sarna · Forthlane Partners

  Gold is the logical hedge against bonds if deficits are unsustainable, and should be higher within five years.

  If you think deficits are unsustainable, gold is the logical other side of a bond trade. Sarna asks whether gold will be higher in five years and answers that US deficits cannot realistically be cut. He notes gold has begun correlating more with equities, which he sees as a function of fiscal dominance: companies are also real assets that can raise prices, and both legs are propelled by fiscal policy. He also cites gold's divergence from real yields starting in 2002 as evidence the old rules have changed.

  “Gold is the logical other side of that trade if you think there's unsustainable deficits.”at 9:55

US stock market outlook

• Andrew Sarna · Forthlane Partners

  US stocks look expensive on cyclical P/E, but investors need AI exposure in case the AI forecasts come true.

  Sarna says that top down, things look pretty rich, especially on a cyclical PE ratio. But the market is being carried by an AI-driven economy, and if the predictions about AI come true, investors have to hold exposure to that. He has been surprised by how resilient equities are despite rising yields, which he rationalizes as the AI trade acting as an unstoppable force. He also notes a belief in both a Fed put and a Trump put as a backstop in a sell-off.

  “equities, the hard part is, yes, they look super expensive, especially if you're looking at a cyclical PE ratio. But at the same time, if these predictions do come to fruition about AI, you better have some exposure to that.”at 10:02

US deficit & government debt

• Andrew Sarna · Forthlane Partners

  The US cannot cut its deficit because entitlement reform is politically unpopular, so fiscal pressure persists.

  Sarna argues the US will not be able to cut its deficits. Elon Musk tried, but deficit reduction requires going after entitlements, and that is deeply unpopular, so no president is likely to do it. He adds that each 1% parallel upward shift in the yield curve adds about $120 billion to the US deficit every year, so the administration has strong reason to want yields lower.

  “Well, there's no way that The US is gonna be able to cut down their deficits. I mean, Elon Musk tried, but it's clear you have to go after entitlements and which president is going to go after entitlements? It's so deeply unpopular.”at 13:51

Copper & industrial metals

• Andrew Sarna · Forthlane Partners

  Copper will tighten over five years as grid build-out, electrification and no new mines push demand past supply.

  The US must invest heavily in the grid, the world is moving toward electrification, and no new mines are being built. Short term, copper will see noise and sensitivity to the economic cycle, but five years out Sarna expects the market to tighten further. He says it is hard to chase after a rally, so he sees pullbacks as an opportunity to add to the position.

  “Or we can look at something like copper where we know The US has to invest so much into the grid. You know the world wants to move towards electrification. We know they're not building any more new mines.”at 14:07

Commodities (broad)

• Andrew Sarna · Forthlane Partners

  Years of low prices starved commodity capex, so supply will tighten and prices should rise, not fall.

  After nearly two decades of a commodity bear market, companies in gold, copper and oil have neglected capex because prices stayed low, so no new supply came online. Sarna thinks the math will eventually play out and push prices higher. He rejects the idea that high prices will hurt demand enough to reverse them, arguing AI demand is too voracious and grid build-out needs are too large. He does expect short-term weakness tied to the economic cycle.

  “I think eventually the math just is going to play out where we have invested, whether it be the gold industry or copper or oil, we've just neglected so much CapEx because prices have just been so sustainably lower that they haven't incentivized any new supply coming online.”at 19:34

Oil & fuels

• Andrew Sarna · Forthlane Partners

  A proposed US diesel export ban would backfire, since Japan takes 37% of its diesel from the US.

  Diesel is almost 100% higher than last year, and the administration has floated a diesel ban. Sarna says that while it could help the US in the short term, the derivative effects are complicated. Japan imported 37% of its diesel from the US in July. Because Japan also owns a lot of Treasuries, choking it out of the diesel market would set off a chain reaction with many unintended consequences.

  “guess who imported 37% of their diesel from The US in July? Well, Japan. And unfortunately, Japan owns a lot of treasuries. So choking them out of the diesel market probably will have a chain reaction that would have lots of unintended consequences.”at 23:21

Power demand & the grid

• Andrew Sarna · Forthlane Partners

  US power shortfall through 2028 makes power stocks attractive without a parabolic price move.

  Sarna cites a recent Morgan Stanley report that the US power shortfall is expected to reach 57 gigawatts through 2028. He puts that in context: it is roughly ten New York Cities' worth of consumption, which raises the question of where the supply will come from. He links this to the copper trade and says some power producers have fallen from their highs, partly because of the political pushback on data centers. He argues that power prices are unlikely to go parabolic like memory prices, because power is more regulated. He concludes that power stocks can be attractively valued at current levels without that kind of spike, and presents power as a long-term opportunity.

  “You don't need to see with these stocks, you don't necessarily need to see that parabolic move for them to be attractively valued at these levels.”at 40:47

Commercial real estate

• Andrew Sarna · Forthlane Partners

  Senior housing is an undervalued play on the boomer demographic cliff, despite near-term rate pressure.

  Sarna describes a demographic cliff: more baby boomers are reaching the age where they will need care, so demand for senior living should rise. He says COVID hit the sector hard, which leaves it cheap, and he calls it one of the pockets of value Forthlane is looking at. He concedes that rates will weigh on the sector in the short term and that hoping for rate cuts is not a strategy. He still argues the demographic tailwind makes the opportunity work over the longer run.

  “COVID decimated that sector. So you sort of have two elements at play. But obviously, in the short term, rates are going to weigh on that.”at 41:23