Get Mike Santoli’s Market Memo in your inbox first. Subscribe here
“The market bottoms on bad news” is one of those rules of thumb that sounds smart and proves true often enough to maintain its good standing as a Wall Street cliché.
Did weeks of brutal punishment for AI stocks culminate last week, with a reversal and rally on Monday after a slew of unfavorable headlines? Last week, American voters strongly opposed data center construction in much of the country, and Anthropic lowered its target IPO valuation just before the sex and espionage scandal broke out against the man who embodies the entire AI economy: Nvidia’s Jensen Huang.
Despite Nvidia’s 4.3% decline after Monday’s 3.4% rise in the Nasdaq 100 index and 7% decline in the Philadelphia Semiconductor Index (SOX), the narrative that the market took its cue from the culmination of bad news after a nasty setback and staged a big rally is well-contained for now.
Based on purely tactical evidence, the technology and hardware complex was flushed out pretty well entering this week and was poised to rebound. Many trading desks reported a historically rare intensity of intragroup net selling by financial institutions following last Thursday’s risk purge. The Nasdaq 100 Index was oversold on several indicators, similar to the Iranian low on March 30th. The CBOE S&P 500 Volatility Index recorded the kind of nice spike that usually accompanies the passage of a storm, rising to 28 but then falling back below 20.
Laurent Laskowski, options strategist at Baird Strategas, said there were “a lot of elements of capitulation” in the recent Nasdaq sell-off and rise in risk aversion. “Record options trading, spikes in volatility, skew and correlation, inverted VIX curves, and forced liquidations… So the setup is that the market is seeing the left tail (of the bottom) contracting and the right tail (of the top) contracting.” gain weight ”
Several technical analysis gurus (including John Kolovos of Macro Risk Advisors) argue that the market has not yet provided enough of a “fat pitch” to rally aggressively. Maybe so. But only about 7 percent of all pitches in Major League Baseball are in the center of the strike zone, according to FanGraphs calculations. These fat pitches lead to about 23% of the home runs hit, but the other three-quarters of balls that go over the wall come from less-than-perfect setups near the edges of the zone.
For these reasons, and the fact that valuation demands are much lower, I wouldn’t be too vehemently opposed to this tech-heavy bull market continuing for some time. Last week, I noted that while Big Tech companies have not yet responded to moderate stress conditions, they are approaching one, saying, “For a bull market built on the AI story to continue moving forward, the AI story will almost certainly need to get back on track. Extended periods in which the equal-weighted S&P 500 outperforms the market-cap weighted version are typically known as bear markets.” Monday’s surge could be a start, but there’s still a lot to prove.
However, this AI-driven part of the market has recently shown itself to be unworthy of investors’ trust, and there are enough signs of structural damage in the bull market that confidence cannot be restored anytime soon.
In a late June column, I asked, “Was May 14th a critical moment for this market, at least tactically?” This day, about six months ago, was the day the S&P 500 index soared to the 7,500 level for the first time. Since then, the index has never reached escape velocity, and in fact just yesterday crossed 7500 again.
On May 14, the much-touted AI chip startup Cerebras went public, with its stock more than doubling to a high of $386, amid reports that President Trump and the Chinese leader had made progress toward ending the Iran war. Since then, Cerebras has plummeted to $97, the semiconductor group is down 17%, and the Iran conflict continues.
SpaceX’s IPO a month later was another clear culmination of New Era’s technology enthusiasm. It may not have been a significant peak of AI excitement, but any subsequent market movement will probably be similar to what we’ve seen since then.
Perhaps a half-year calculation period in which the market refused to pay 30% annualized earnings growth for two consecutive quarters is enough to build a new foundation built on more modest expectations. I have focused on the late-cycle upheavals that spook markets in the form of peak earnings growth, peak capital spending, peak index concentration, Federal Reserve tightening, and a resurgence in credit. This all explains the months-long stagnation in the market, which continues to stagnate.
But it’s important to always be open to the possibility that these are just mid-cycle disruptions, as Morgan Stanley Chief Investment Officer Mike Wilson has argued.
But first, the tape needs to show progress beyond a massive one-day comeback. Another rule of thumb is that a rebound rally needs to be at least above the 20-day moving average to prove it’s more than just a reflexive dead-cat bounce. This would require the S&P 500 to rise 1.3%, the Nasdaq 100 to rise 3.7% and the SOX to rise 9% from Monday’s close.
Without pretending to have any clear idea of how Nvidia’s various investigations will impact the company’s sales, leadership, or long-term profitability in China, I can safely say that this scandal has come at a very opportune time for the market.
Persistent concerns that current profits are unsustainable and the group remains vulnerable to a slowdown in data center construction have pushed valuations to multi-year lows. This allowed other semiconductor stocks to soar in a single day as Nvidia lost about $200 billion in market capitalization on $80 billion in trading volume.
Of course, the overall S&P 500 is also visibly undervalued as semiconductors drive up expected earnings and investors refuse to assess the group’s sustainability.
The problem is that hyperscalers spend so aggressively that less revenue shows up as free cash flow. As this chart comparing valuation gauges shows, the market appears to be significantly cheaper in free cash flow than in earnings.
– This entertaining and enlightening review of James Glick’s “The Telephone: A New History” published in The New Yorker inevitably draws parallels between the advent of telecommunications and the current build-out of AI. Massive privately funded infrastructure developments, destabilizing new habits of interaction with technology, and persistent calls for government regulation in exchange for near-monopoly power. The book cites a 1909 AT&T advertisement that called the telephone “an instrument of civilization…one of the most powerful influences in extending human intelligence.”
– Mark Zandi, economist at Moody’s Analytics, shares a useful thread suggesting that AI is adding more jobs than it is hurting the U.S. economy so far. But the main reason is austere: most of the profits come from building data centers and creating wealth for people who own stocks that rely on AI. Cold comfort for job seekers in the information economy.
-Last week’s election, which led to the Democratic takeover of Congress, drew comparisons in size and message to the 1994 “Contract with America” Republican landslide. The election of Bill Clinton’s first term highlighted deep social rifts and economic instability.
In retrospect, many see the ’90s as a halcyon era of consensus and prosperity, but the decade between the ’94 election backlash, constant corporate downsizing, fear-mongering over federal debt levels, the 1995 Oklahoma City bombing and discomfort with technological change took a toll on many of today’s narrative beats. This Chicago Fed article, “The Decline in Job Security in the 1990s: Displacement, Insecurity, and Their Impact on Wage Growth,” can fill in the blanks for those who don’t get it.
