The "chip influencer" Citrini, who was among the earliest to go long on AI hardware, is quietly exiting

Author:

Source: Wallstreetcn

Citrini Research, the viral chip research firm, was once one of the earliest and loudest bulls on AI hardware trades. Now, Citrini has significantly reduced its semiconductor exposure and even holds some short positions. Behind this shift is his judgment on a fundamental change in the financing structure of AI infrastructure: when incremental capital switches from the self-funded cash flow of hyperscalers to equity and debt markets, market sentiment itself becomes part of the fundamentals.

On October 21, Citrini founder James van Geelen recorded this conversation on the Monetary Matters podcast. At the time, semiconductor stocks had just suffered nine consecutive declines—host Jack noted this was the longest losing streak since 2006. James's judgment was that this was not enough to shake AI capex, but the nature of the risk had changed. He said: "If market sentiment turns sharply, that incremental capital will dry up—because this is no longer a company spending its own money, but something decided by the market."

The conversation covered why he rotated out of semiconductors, the logic behind long Meta and short Google, the implications of DeepSeek's roughly 80% inference gross margin, and what signals would make him go long chips again. At the same time, he disclosed that the bottleneck in new drug development is "scarce monkeys" rather than compute, and the real investment opportunity lies in clinical outsourcing, not AI concept stocks.

Tech Giants' Cash Flow Turns Negative, the Financing Structure Has Changed, and Sentiment Can Kill Fundamentals

To understand why Citrini reduced its AI hardware exposure, one must first understand what he calls "reflexivity" risk—this is the core mechanism behind the entire judgment.

In the first three to four years of the AI trade, hyperscalers paid for AI infrastructure with their own operating cash flow. This meant that no matter how market sentiment fluctuated—whether tariff shocks, yen carry trade unwinds, or any other panic—as long as companies like Meta and Microsoft kept spending, semiconductor companies' fundamentals had a "floor" beneath them. Sentiment could push share prices down, but it could not crush revenue.

Now this mechanism has failed, and the cash flow of tech giants is rapidly drying up or even turning negative. Citrini James's judgment is that nearly every incremental investment in AI infrastructure today comes from equity financing (whether venture capital or public markets) or debt financing.

This means AI can become more useful and usage can continue to expand, but if market sentiment turns sharply, that incremental capital will dry up—because this is no longer a company spending its own money, but something decided by the market.

This is what he means by "reflexivity": a sentiment downturn can directly cut off the funding chain, thereby affecting actual construction spending, not just depressing valuations.

He further quantified this pressure point. Supporting current infrastructure valuations requires AI labs' annual recurring revenue (ARR) to reach roughly $1 trillion by the end of 2028. Based on the current ARR base of about $13.6 billion, that requires roughly $50 billion in net new ARR every month.

If you are relying on (AI) labs' combined ARR reaching $500 billion to support your assumption of capital flowing back into infrastructure, then ending up at $300 billion would be a disaster.

James listed several factors dragging on ARR.

First, "the consumer side is good enough." His own experience using frontier models is that daily needs are "99% covered" by existing models, and stronger capabilities can hardly drive new paid demand.

Second, "price cuts have not translated into proportional usage growth." He cited an estimate from the Anthropic Economic Index: "A 10% reduction in task cost corresponds to only a 3% increase in usage."

In his interpretation, this data point suggests that the Jevons paradox (efficiency gains actually raise total consumption) may not apply to tokens in the short term. After enterprises build a workable AI workflow, call volume may even decline: perhaps token demand is actually higher when the product is not good to use, because you have to keep revising.

Third, open-source share is rising. He mentioned that open-source models' share of new ARR has, by his estimate, risen from about 1% a year ago to about 18%, with buyers paying per task as long as minimum capability is sufficient: they don't drive fighter jets to work; they drive cars.

He believes the capability gap between Chinese and American models still exists, so price shocks are not his biggest concern. His most important criterion is: this is not a question of whether AI is useful or whether AI will get better, but whether frontier spending is actually economical. He believes that in the short term, "white-collar work" must become that killer app and bear the vast majority of token demand over the next few years.

"Cheapest" Is Exactly What the Top of the Cycle Looks Like

The host pointed out that on an earnings basis, these tech companies now look cheaper than in 2023 because their profits have surged, so why reduce exposure instead? Citrini James applied the rules of cyclical stocks and answered:

That's the cycle. Cyclical industries always peak when they look cheapest on next year's expectations.

In other words, James believes cheapness is an illusion created by temporarily inflated earnings—cyclical stocks have always been like this. Cyclical stocks' earnings rise and fall with the business cycle. At the peak of the cycle, company profits surge, the denominator grows, and the P/E ratio calculated on "next year's expected earnings" becomes very low, making the stock look cheap. But that profit is peak-cycle profit; once the cycle turns down, earnings decline and the share price falls with them. So the moment when something "looks cheapest" is often precisely the moment when earnings are highest and the cycle is near its top.

The host added an analogy: no one buys an oil company when its profits are fattest and its P/E ratio is only 14x.

This logic explains his repositioning. He said that if he had maintained a heavy infrastructure portfolio in April and May, the overall portfolio "would probably be down 50%." And during the two-to-three-month semiconductor drawdown in September and October 2023, he was not in a hurry to find substitutes because valuations were low and downside was limited. Now it is different:

The downside, when a truly negative outcome or negative sentiment emerges, is very large.

He frames his actions as portfolio adjustment, not a bearish call on AI as a whole: he remains bullish on Nvidia and sees more asymmetric opportunities in what he calls "agentic utilities" (industries that actually use AI and benefit from AI adoption). He also said demand still exceeds supply for now; the problem is the future.

Long Meta, Short Google, Waiting for the Signal to Return to Chips

On specific positions, Citrini James's logic is equally clear.

The core thesis for being long Meta is the potential monetization space of Muse (Meta's AI assistant product). His calculation framework is: Meta's current average revenue per user (ARPU) is just over $20, with data assets covering Instagram, WhatsApp, and Facebook. "If Muse achieves any meaningful adoption, you can draw a direct line from that data to a 4x increase in ARPU by 2030." He said he turned strongly bullish this summer, and Muse's launch confirmed that judgment.

The logic for shorting Google comes from his concern about talent drain at DeepMind. He believes the market underestimates the negative impact of this wave of departures, and it is happening precisely in a window when the search business still faces structural threats. "I myself use Google search far less now. Will search be completely killed? I don't know, maybe not. But the risk remains, and at the same time they are experiencing talent drain." He said he has stop-losses in place but currently maintains the "long Meta, short Google" pair trade.

As for when to return to semiconductors, he gave clear trigger conditions: seeing sustained growth in paid usage and a re-acceleration in the second derivative of ARR growth that can genuinely support financing construction through customer cash flow. If current demand is actually still constrained by supply and this is merely a gap between a leap in model capabilities and new product launches, and once products launch users are more willing to pay for new capabilities, then his concern does not hold, and this pullback in infrastructure stocks is instead a buying opportunity.

His view on CoreWeave is relatively mild—he sees a default probability of about 60% but does not expect an actual default. His view on the path of an AI cycle downturn is that it "ends with a whimper rather than a bang": not a crash triggered by one company's default, but lab ARR expectations revised down from $400 billion to $360 billion, with that incremental difference spreading outward because it has already been fully priced in, dragging down the entire sector.

On DeepSeek, he cited data from DeepInfer (DeepSeek's internal inference framework): serving 155 trillion tokens per day, revenue of about $38 per GPU hour, while market rental cost for GPU hours is about $5 to $8, implying an inference gross margin of about 80% and a GPU payback period of about 81 days. "This means token pricing may have further downside, especially in non-frontier models."

Designing Scorpion Venom Antibodies with Claude: What AI Compresses and What It Cannot

Citrini James also disclosed a rare experiment: his team worked with Anthropic to use Claude to design 211 scorpion venom antibody candidate molecules in 64 hours, 17 of which successfully bound to targets in the lab. The result is both shocking and led him to a conclusion that is not friendly to AI bulls: AI is "commoditizing" drug design, but the real bottleneck is in the lab, in the monkeys, not in compute.

To figure out what AI can and cannot do in drug discovery, Citrini James's team did something unusual: they contacted Anthropic's life sciences team (Nathan Frey, Amir Chennisa, and Steven Molina) and, inside Anthropic's office, used its internal research model to spend 64 hours designing scorpion venom antibodies.

Throughout the process, Claude called about 300 sub-agents, generated 180 design plans, and ultimately submitted 211 candidate molecules for testing. Lab results showed that 17 molecules successfully bound to targets, 7 of which were against the specific Mexican scorpion venom CN2, and the best candidate, CN2B1, also formed competitive binding with a known neutralizing antibody in the expected region.

"AI generated these molecules, and 17 of them bound in the lab—that is quite shocking."

But James immediately offered what he considers the more important other side. What AI compressed was the time from idea to experimental protocol—a process that originally took months now takes only days. But synthesis, testing, and obtaining control data still took 39 days. "If data moves at the speed of life, the arrival of AI cannot make cells in a petri dish grow faster. That is where the bottleneck is."

He estimates that across the entire process from drug design to clinical trials, AI can compress about 20% of the time, while the other 80% is incompressible. This judgment leads directly to his investment conclusion: go long experiments, go long companies that make experiments faster, rather than going long AI drug discovery concept stocks.

He explicitly named Charles River Laboratories (CRL) as his highest-conviction long. About 60% of CRL's revenue comes from early discovery and safety assessment, it is at the trough of a cyclical recovery, and it has ample capacity to take on new orders, with each new contract flowing directly to profit.

He also pointed out a bottleneck the market has not fully priced: the monkey shortage. Non-human primates are the closest-to-human irreplaceable resource in preclinical testing, with a breeding cycle of about four years, making supply expansion extremely difficult. "This may be the most important supply constraint in preclinical testing."