Key points: An AI bull market is unlikely to be led higher by names like Walmart or Eli Lilly. JPMorgan technical strategist Jason Hunter is advising clients to stay defensive until the market shows a clear capitulation bottom signal. Compared with other tech giants making massive AI investments, Apple's once-key valuation advantage has now disappeared.
If the bull market is going to charge ahead, the leadership will not come from Walmart, Berkshire Hathaway, and Eli Lilly.
Over the past few weeks, money has poured furiously into traditional defensive quality names. This has created an almost perfectly opposite hedge against the selloff in semiconductors and AI infrastructure spending stocks.
Those three trillion-dollar companies are the only non-big-tech members of the trillion-dollar club. Over the past month, their combined market value has risen by $550 billion, nearly offsetting the market value wiped out from Alphabet.
The Nasdaq 100 and the S&P 500 Low Volatility ETF (SPLV) have been converging steadily. The picture looks symmetrical, but it reflects frequent tactical rotation among investor groups, with money ping-ponging between different factors.
This seesaw between sectors has kept the S&P 500 from any significant pullback for months, but the index has remained stuck in a range since May — and since May, the market has been through two earnings seasons, with annual corporate profit growth reaching 30%.
A week ago I wrote that the AI theme had reached a make-or-break point: either it recovers, or it loses its leadership status and drags the broader market down with it.
So far, even though the sector has entered a moderately oversold range and some hyperscalers have delivered superficially strong quarterly results, AI has still not recovered, and the market's tolerance for it is steadily declining.
When the market is under pressure and investors begin reassessing the long-term bullish case, share prices often need to probe lower to attract committed buyers. At the same time, the market needs to frustrate dip buyers in order to eliminate complacency.
We cannot yet judge whether we have reached that stage, but we should be closer to it.
The Nasdaq 100 closed Monday exactly at its 200-day moving average, down nearly 11% from its high. An early Monday rebound attempt failed, and the Philadelphia Semiconductor Index plunged another 3.8%, closing below its late-July low, just after a certain macro hedge fund had completed a full liquidation.
JPMorgan technical strategist Jason Hunter noted: "The market has not followed seasonal patterns — the typically bearish seasonal pressure usually fades by mid-October, but instead, in the late-summer and early-autumn trading, there was an extremely violent wave of month-end selling pressure." He advises clients to stay defensive "until the charts show a clear bottoming signal, or until there are clear signs of panic selling."
This is a time-tested playbook, but historically many deep pullbacks have ended without a full-blown capitulation.
The resilience of non-tech stocks has won considerable recognition. Many sectors have benefited from the "quality asset" rotation; falling oil prices and the 10-year Treasury yield dropping back below 5% have also given the neglected cyclical consumer sector some breathing room.
For active fund managers, the highly concentrated, highly tech-dependent S&P 500 offers plenty of room to operate.
But I still hold to one view: a bull market built on the AI narrative, if it is to keep rising, almost certainly must rely on the AI leadership getting back on track.
Generally speaking, if the equal-weighted S&P 500 consistently outperforms the market-cap-weighted version, that often signals a bear market environment.
Morgan Stanley data shows that although the weight of AI-related constituents in the S&P 500 has fallen from its peak, it is still close to 45%.
Historical charts show that during this rally, the AI sector also experienced turbulence before reclaiming leadership. But this test is trickier: each previous AI buildout wave had a clear bottleneck, giving investors assets to position for.
In 2023-2024, GPUs were scarce, so the market piled into Nvidia; when cloud computing power, electricity, and cooling facilities became scarce resources, hyperscalers and AI-related industrial companies were revalued; then memory chips became the bottleneck, and Micron and SanDisk shares soared.
And as Nvidia's Jensen Huang said, the current constraint is capital itself. That means there is no longer a clear asset to bet on; large tech companies' free cash flow turning negative has also raised the cost of capital across the market.
Multiple feedback loops between the real economy and financial markets are intertwined, with correlations in disarray, further eroding investor confidence.
Treasury yields no longer track oil prices and are now mainly driven by AI capex expectations, and recently the two have moved lower together. The AI sector's decline simultaneously weakens upward inflation momentum and the stock market wealth effect; compared with a month ago, market expectations for a Fed rate hike have cooled.
All of this looks like an elaborate trap, and for participants in the previously crowded trade, it has already become reality.
Does this mean the AI mania is over? As I have repeatedly mentioned, the peak of this cycle roughly occurred around the SpaceX IPO in June. We cannot make a definitive call, but even if that is how it ends, by this point the tape would look exactly like this.
However, I also do not rule out the possibility that market prices have already fully digested most of the known bad news. I have previously argued that this cycle's valuation peak is already in the past. The S&P 500's forward P/E has fallen from 23 times a year ago to 18.5 times; the Nasdaq 100's 20 times P/E, relative to the S&P 500, is at a nine-year low valuation premium.
The market seems to be pricing in companies' "earnings being at a high level and set to decline in the future," but at least for now valuation risk has been released somewhat compared with recent months.
Sentiment in the AI sector has fallen to a trough, and fund positioning in tech stocks has dropped below neutral. The two important stage lows on March 30 and July 29 this year both came near month-end; and the current oversold reading in the Nasdaq 100's relative strength index is exactly consistent with those two points.
Therefore, be wary: the market may stage a relief rebound, and if the midterm elections pass and traders' taut nerves are released, the rebound could also turn into a larger-scale move, carrying into the year-end stretch.
Sentiment indicators: Latest view from a market analyst: "Like valuation, sentiment itself is not a timing tool. But right now it is almost hard to find bullish voices; financial conditions are tightening, tech earnings are deteriorating one after another, and multiple bearish factors are reinforcing each other."
Wall Street view summary: Investor Cliff Sosin joined an interview to discuss the business realities and investment implications of the new phase of AI adoption. He compared the economics of AI infrastructure to internet broadband operators; he argued that most of the productivity gains from AI will translate into consumer surplus rather than corporate profits — an insightful point.
Monday was my birthday, and my colleague Sara Eisen wanted to know how I would write a birthday message to myself. I had her try ChatGPT, and it generated this in a financial-writing style: "Another year older, but the long-term trend remains bullish. Underlying internal metrics may have deteriorated slightly, but no decisive breakdown signal has appeared yet. The number of candles on the cake has reached a historic high, and valuation concerns have lingered for years without interrupting the upward trend. At this stage, the question is not that I am getting older — that is largely already priced in — but whether the gain in wisdom can continue to offset the inevitable contraction in the multiple of energy. Keep cautiously optimistic in the short term, retain upside options; family and friends form downside support, and no emergency Fed rescue is needed. Happy birthday to me. The long-term bull market remains intact."
Closing commentary: Since AI capex anxiety erupted several months ago, money has surged into defensive sectors, which has also been a core driver of Apple's significant outperformance versus the Nasdaq 100. In the six months before Monday's earnings release, Apple's excess return relative to the Nasdaq 100 was close to 20 percentage points.
Apple did not follow the crowd in spending heavily to expand AI computing power, preserving abundant free cash flow, making it the perfect hedge against the "capex trade," and its own valuation was pushed to a 15-year high.
But now, with Apple's stock price continuing to rise while peers' shares have fallen sharply, Apple's relative advantage in free cash flow yield has narrowed.
Comments