The AI Trade’s Two Bears
TREE NEWS reports: US hyperscalers just delivered their sixth consecutive quarter of accelerating cloud revenue growth, yet Intel suffered its worst single-day plunge in half a century, falling 29%. Caught in the middle, hedge funds spent October dumping anything with “AI” in the name. The market is grappling with two contradictory fears: insufficient power to sustain AI buildout, or an imminent glut of compute. Both narratives argue for lower valuations, and in a crowded trade, either can trigger a selloff.
What Happened
Last week, cloud giants reported stellar numbers: Azure grew 48%, AWS 39%, and Google Cloud 103%. Combined cloud revenue for Amazon, Alphabet, and Microsoft accelerated to 54% year-over-year in Q3. But Intel’s warning of a severe PC demand correction spooked investors, and the stock cratered 29% on Friday. Meanwhile, hedge funds deleveraged at the fastest pace since 2015, with systematic long-short strategies down 2.17% for the week—the worst since December 2023. Net exposure is now at the 7th percentile historically, and individual stocks have seen seven straight weeks of net selling.
Market Implications
The debate between power scarcity and compute oversupply is circular, but it has real consequences. If power is the bottleneck, GPU rental prices stay high, but growth for “picks and shovels” stocks flattens. If compute becomes abundant, rental economics collapse, undermining chip returns. Either way, multiples compress. Goldman Sachs’ Rich Privorotsky notes that strong demand is undeniable, but the return on the next trillion dollars of investment is highly uncertain. Hyperscalers are expected to spend $1.2 trillion on capex by 2027, yet their combined free cash flow turned negative in Q3, burning $31 billion. This means the next wave of investment will be funded by debt markets, where AI-related off-balance-sheet debt has already ballooned to $4.1 trillion.
Equities are pricing a sharp growth slowdown and stalled AI capex, yet earnings show no sign of spending restraint. This disconnect suggests the selloff is more about positioning than fundamentals. Total leverage remains elevated at 314% (97th percentile), so further volatility is likely. However, with sentiment washed out and seasonal tailwinds ahead, conditions for a year-end squeeze are building. UBS warns of record short positions, and Goldman’s derivatives desk sees signs of healing across seven markets. Still, any rally could be a selling opportunity given the uncertain return on AI investment.
Key Takeaways
- AI trade is caught between fears of power scarcity and compute glut; both lead to lower valuations.
- Hyperscaler demand remains strong, but negative free cash flow and rising debt raise questions about capital returns.
- Hedge funds have aggressively deleveraged, but total leverage remains high, leaving room for more volatility.
- Year-end rally possible due to washed-out sentiment and seasonal factors, but risks remain.
