AI Bull Market Turning Point: Leverage Liquidations, Computing Power Glut — Why This Macro Analyst Turns Fully Bearish

Original Author: Geo Chen (Fidenza Macro)

Original Translation: TechFlow

Introduction: Geo Chen of Fidenza Macro liquidated all of his AI semiconductor and infrastructure positions in June this year. Now, he presents his complete bearish thesis. His analytical framework spans the Korean leveraged ETF liquidation, the impact of Chinese models on standard closed-source AI, the inflection point where compute supply shifts from shortage to surplus, and the risk of the Fed losing credibility under stagflationary pressures—a must-read for any investor holding AI-related assets.

March 2000. March 2008. January 2020. December 2022. Certain months stand out vividly in my memory—they were all inflection points before violent market turmoil. I believe that, in hindsight, this month will be remembered the same way.

In June this year, I liquidated all of my AI semiconductor and infrastructure positions and moved to cash. Afterwards, I took a break, enjoying a summer away from the markets, expecting a quiet, rangebound season with few opportunities.

The outcome was completely unexpected.

Everything that has happened over the past month has increasingly convinced me: the bull market in equities has peaked. I am generally an optimist, so this conclusion was not reached lightly. Unfortunately, developments in the AI sector, the war in Iran, and Fed policy are combining to create stagflationary conditions, with more turbulence ahead. In this article, I will break down my bearish logic point by point.

Why the AI Bull Market Is Over

AI is the leader of this bull market and the primary driver of GDP growth. Without AI, the bull market lacks support. The strong earnings of hyperscalers and semiconductor companies have underpinned this rally, so some argue: as long as earnings grow and fundamentals remain solid, the bull market can continue. But the reality is that stock prices reflect capital flows and market narratives; earnings are just one of many drivers. Most bull markets peak before earnings decline, and some even end before analysts begin cutting estimates.

Many bull markets have an overshoot phase: low-quality capital pushes the market to new highs, often in a parabolic pattern. Low-quality capital refers to groups with asymmetric information, price insensitivity, and unsustainable buying power. The overshoot phase is usually only clear in hindsight, but if you can identify low-quality buyers driving the final leg up, you have a chance to spot the overshoot in real time. This was the framework I used when exiting the crypto market in August 2025 — when I judged that MicroStrategy and crypto treasury companies were providing exit liquidity for the cycle top.

In this AI rally, the low-quality capital has primarily been Korean retail investors, who heavily bought 2x or 3x leveraged ETFs on SK Hynix, the KOSPI, and other memory stocks. This capital spilled over, inflating valuations of companies across the AI supply chain benefiting from supply bottlenecks. These leveraged ETFs created hundreds of billions of dollars in additional buying power, but this force is simply unsustainable. The hedging demand from these products forced market makers to accumulate large negative gamma exposure, compelling them to buy on up days and dump heavily on down days, and the resulting volatility triggered liquidations and margin calls.

Citi estimates that the fallout from the leveraged product sell-off has so far erased $38.7 billion, and data circulating online suggests 1.2 million accounts were liquidated—meaning roughly 1 in 30 Korean adults was wiped out. The extreme speculation among Korean retail investors exceeds anything I have seen in other bull markets. A sense of financial nihilism has driven many Korean retail investors to go all-in, amplifying their positions with leverage, simply because they feel they arrived too late and must catch up:

Recently, a post appeared on the workplace community Blind telling the story of someone who suffered massive losses in a margin call. The poster wrote: "SK Hynix and Samsung Electronics kept rising, but I felt I entered too late and panicked." He added: "After the market close, I went all-in with my entire assets of 170 million won plus 200 million won in unsettled margin, totaling 370 million won, and the next day the Korean stock market crashed, causing massive losses."

— From Business Korea

And after all this pain, crowding in momentum stocks remains at elevated levels.

Chart: S&P 500 momentum leader crowding (J.P. Morgan, currently at 93.3%, near the previous peak from July 2026). Source: J.P.Morgan

Parabolic bull markets almost always end with prolonged bear markets. The more extreme the sentiment, prices, and leverage during the ascent, the worse the hangover afterwards. Once margin is called, that capital is impaired and rarely returns. For the bull market to reach new highs, this impaired capital must somehow be replaced by new capital and entirely new narratives—a repair process that takes a long time, if it happens at all.

Those waiting for a new narrative to reignite the AI bull market are likely to be disappointed. If anything, the narrative has only worsened over the past few weeks. Chinese AI lab Moonshot AI released Kimi K3, a model that has outperformed Claude Fable on Code Arena.

Chart: Frontend Code Arena rankings, Moonshot AI Kimi-K3 leads (1,679 points), surpassing Claude Fable 5 (1,631 points). Source: Arena

Alibaba followed in Moonshot AI's footsteps, releasing Qwen 3.8, a large open-weight model with up to 2.4 trillion parameters. Market sentiment is also shifting toward open-source models, with NVIDIA's Jensen Huang being the latest heavyweight in the AI space to publicly endorse open source.

Global macro guru Louis Gave once said: "When China enters, profits run away." This proved true in electric vehicles and solar panels, and now the market fears that competition from China will commoditize intelligence. All signs point in the same direction: the cost of intelligence is converging toward the cost of the compute required to serve it. Users can obtain near-parity performance from open-source models at a fraction of the cost of closed-source models, with better data privacy and no lock-in, making it increasingly difficult to justify paying a premium for OpenAI and Anthropic.

Cheaper intelligence is good for end users, but it is terrible news for closed-source AI labs (OpenAI, Anthropic, Google) that have committed to massive compute leasing or procurement expenditures. As margins and market share erode, their ability to raise capital at higher valuations weakens, which in turn undermines their capacity to fulfill compute commitments. OpenAI's decision to delay its IPO until next year is likely due to a lack of confidence in achieving a $1 trillion valuation. SpaceX falling to $112, down 27% from its IPO price, may also be further dampening their IPO prospects. Because OpenAI and Anthropic engage in circular transactions with other participants across the ecosystem, they have become a single point of failure for the entire AI industry.

The Coming Compute Glut

AI bulls point out that AI compute and components such as memory and optical networking remain in shortage, but this logic is flawed. Every commodities trader knows that when supply shocks and bottlenecks are felt most acutely, that is often the top of the bull market. By the time supply and demand rebalance, the bull market has usually fully reversed. In many cases, shortages actually turn into gluts, leading to prolonged bear markets.

Given the scale of compute that emerging cloud providers and hyperscalers have already committed to or broken ground on, I would not be at all surprised to see a compute glut emerge a year or two from now.

A shift in compute from shortage to surplus can easily happen in parallel with token consumption and AI model revenue continuing to grow rapidly. The efficiency gains in hardware and AI algorithms are outpacing the rate at which users are increasing token consumption, causing total token expenditure to decline from June levels. Silicon Data's token spending index shows that overall token spend peaked in June and has been trending lower since.

Chart: SDLLMTK index (token spending index), peaking in June and declining since. Source: Silicon Data

What would a compute glut look like? Abandoned data centers, broken commitments, and in some cases debt defaults. It could get ugly. Corporate bond spreads for data centers and hyperscalers are signaling that the astronomical investment in compute is becoming an increasingly dangerous business decision.

Chart: AI data center bond spreads widening (Hut 8, QTS, Meta, etc.). Source: Bloomberg

Chart: AI ecosystem credit risk surging (5-year CDS basis points, SPCX soaring). Source: Bloomberg

The stock market is no longer rewarding hyperscalers that announce increased AI capital expenditures, but they are ignoring this signal and continuing to ramp up spending.

Chart: Consensus estimates revisions for capital expenditures, FY2024–FY2026. Source: Bloomberg

Google announced it would raise its 2026 AI capital expenditure from $195 billion to $205 billion, and its stock fell 7% on the day.

I know this bearish scenario is hard to imagine, but recent history offers plenty of reminders. When the Strait of Hormuz was blockaded in April, almost no one expected oil prices to fall back to $70 so quickly. When silver traded at $120 in January, few thought it would drop to $55 within a year. In 2021, almost no one believed the high-flying growth stocks of that bull market would lose 80% to 90% the following year. Shortages can quickly flip into gluts, and positioning can turn just as fast.

Iran — The Next Endless War

I previously believed the Iran war would have limited long-term impact on equities, but my view has changed. The conflict is evolving into a fitful quagmire. Iranian hardliners have no intention of giving up their two trump cards: the stockpile of weapons-grade uranium and control over the Strait of Hormuz. Seizing both would require a protracted ground war, and even then, the odds of success are uncertain. The war also has the potential to become a proxy war between the US and China.

The US Department of Defense estimates the war has so far cost American taxpayers $37.5 billion, but this is likely an underestimate, as it does not account for the economic costs or the future spending required to replenish equipment and ammunition to pre-war levels. Trump dragging the nation into a costly, endless war without Congressional approval will go down in history as one of the most telling examples of the breakdown of American democracy.

Over the past six years, the world has experienced four inflation shocks (COVID-19, the Russia-Ukraine war, Trump's tariffs, and the Strait of Hormuz blockade). Each one led to tighter monetary policy and significant market drawdowns. The Iran war may be the most persistent stagflationary force among them, as it affects global energy and commodity supplies while simultaneously pushing up government financing costs.

The Fed Under Warsh

Kevin Warsh is attempting to overhaul the Fed's approach to measuring inflation, responding to it, and communicating with the public, all while under the dual pressures of a supply shock and the bursting of the AI bull market bubble. It is like trying to replace every part of an airplane while flying through a storm.

With inflation running above target and fiscal deficits widening, Warsh faces two bad choices. He could tighten early and flatten the yield curve, but that risks triggering a recession. Or he could delay tightening and let the long end of the bond market do the work. So far, he appears to have chosen to delay.

At yesterday's FOMC meeting, the Fed had an opportunity to validate Warsh's hawkish rhetoric with a rate hike, but they did not. The bond market reacted with a sharp bearish steepening, a signal that the Fed's credibility in controlling inflation is eroding. Joseph Wang noted that Warsh promises price stability and a 2% inflation target on one hand, while on the other hand modifying how inflation is measured in ways he cannot publicly disclose. Without a clear framework, bond investors have no anchor for their expectations, and volatility will rise in a way the stock market will struggle to digest.

The long end of the yield curve broke above 5.2%, hitting a two-year high. This technical breakout signals a new phase in the US Treasury bear market, adding new headwinds for equities.

Chart: US 30-year Treasury yield weekly, breaking above 5.2% to a two-year high. Source: Bloomberg

Disclaimer: The content of this newsletter is for informational and educational purposes only and does not constitute professional financial advice, investment recommendations, or a solicitation to buy or sell any securities or instruments. This newsletter is not a trading signal service, and the author strongly advises readers not to follow his trades without relevant market experience and research. The author of this newsletter is not a registered investment advisor or financial planner. The information presented in this newsletter is based on the author's own research and experience and should not be considered personalized investment advice. Any investment or trading decisions you make based on the content of this newsletter are at your own risk. Past performance does not guarantee future results. All investments carry the risk of loss, and any trades or strategies discussed in this newsletter are not guaranteed to be profitable or suitable for your specific circumstances. The author assumes no liability for any actions taken or not taken based on the content of this newsletter. The author is not responsible for any losses, damages, or liabilities arising from the use or misuse of the information provided.