I am not a financial professional, and nothing in this article should be construed as financial advice. Please consult your own financial advisor before making any investment decisions, and read my full disclaimer at the bottom of this post.
After yesterday’s absolute bloodbath in the Nasdaq, which fell more than 5%, I wanted to follow up on a few things I’ve learned watching these types of market events unfold over the years, what I’m seeing this morning, and why I think the next several days could be just as important as yesterday’s selloff itself.
While yesterday may have been the day the market finally began acknowledging some of the risks I’ve been writing about since August, I suspect we’re only beginning to find out exactly how much damage has been done. And, as usual, the market isn’t going to make that process particularly easy because sure*, *there’s nothing wrong with releasing *positive *guidance out to 2030 or whatever Oracle did last week, but when it comes to *negative *news that’s material *now…*well, no one wants to piss in the punch bowl.
The first thing I want to note today is that usually, volatility tends to beget volatility…and I’m talking about directionally agnostic volatility. When you get a move as violent as yesterday’s, there’s nothing wrong with expecting similarly violent moves in the opposite direction. In fact, as I write this, Nasdaq futures are already up nearly 2% this morning.
This is not surprising to me. I feel like it’s exactly what happens when markets begin repricing risk in a hurry. Shorts cover, dip buyers show up, traders who were caught offsides scramble to reposition, and investors who have spent the last several years being conditioned to buy every decline convince themselves that yesterday was another golden opportunity to pick up their favorite technology stocks at a discount.
Maybe it was, but I would caution against declaring a violent bounce with an actual mid to longer-term bottom. Some explosive rallies in market history have taken place during sustained bear markets. When positioning becomes extreme and liquidity deteriorates, markets can rip higher with extraordinary speed without changing the underlying trend. A 2% bounce after a brutal selloff doesn’t necessarily mean the selling is finished any more than one terrible day means every stock in the market is suddenly worthless.
Both things can be true: the market can rally violently, and the broader trend can remain decidedly lower. In fact, I would expect precisely that type of action if we’re entering the prolonged repricing of the AI trade that I’ve been warning about. The important thing, in my opinion, isn’t whether the Nasdaq manages to bounce today. It’s whether the fundamental assumptions that supported its previous valuations are still intact.
And I think those assumptions are still only beginning to come under serious scrutiny.
Then there’s the coming policy response. It’s already happening in Taiwan, where authorities announced plans to unveil stock market stabilization measures later this afternoon, while also imposing new short-selling restrictions aimed at supporting the equity market.
It will soon happen in the US if stocks keep falling. The nature of these responses is that they’ve gone from one-time emergency measures to being deployed faster and faster with every increasingly minor bout of market volatility. The US has become so thoroughly conditioned to expect policy intervention at the slightest tick lower in equities that even the most modest selloffs now trigger demands for extraordinary measures.
This is why Jeremy Siegel was screaming for an emergency 75 bps rate cut the last time markets experienced even a modest bout of volatility. It’s Policy Response for Cowards 101. Or, as I titled it a couple of years ago:
We’re barely a day removed from yesterday’s carnage, and already we’re seeing governments consider or implement measures designed to arrest the decline. I suspect that if U.S. markets continue falling at anything resembling yesterday’s pace, it won’t take particularly long before we begin hearing similar discussions here.
During the 2008 financial crisis, the SEC temporarily banned short selling in 799 financial stocks on September 19, ostensibly to restore confidence and prevent disorderly trading. It didn’t exactly fix the underlying problem. The ban remained in effect until October 8, and over those 14 trading days, financial stocks fell more than 12%, according to a subsequent New York Fed study.
Morgan Stanley was an especially striking example: despite short interest in the stock collapsing from 45.3 million shares on September 15 to just 18.4 million by September 30, its share price still plunged roughly 40% over that same period. By October 6–8, with the ban still in place, the S&P 500 financial sector had fallen another 17.8% in just three trading sessions.
And when the restrictions finally expired, the selling continued: on October 9, Morgan Stanley plunged another 26% to $12.45, Bank of America fell 11% to $19.63, and the broader financial sector dropped 11.7% in a single session. The intervention had done little to restore confidence, much less repair the underlying balance sheets.
Subsequent research from the New York Fed found that the restrictions not only failed to arrest the decline but also reduced market liquidity and increased trading costs across equity and options markets by more than $1 billion.
The lesson was straightforward: banning short sellers doesn’t eliminate the fundamental reasons investors want to sell. It merely interferes with price discovery while allowing policymakers to pretend they’re addressing the crisis.
I bring this up because I think it’s important to distinguish between policymakers trying to stabilize prices and policymakers actually fixing the problems responsible for those prices falling. They are not the same thing, and if the underlying issue today is that investors have wildly overestimated the economic returns available from hundreds of billions of dollars in AI infrastructure spending, restricting short sellers isn’t going to suddenly make those investments profitable.
It may produce a bounce and buy some time, but it doesn’t change the math.
It also broke yesterday that “major investors are increasingly skeptical Anthropic’s blockbuster initial public offering can proceed at the expected valuation of $1.5 trillion,” according to the Wall Street Journal.
Imagine that..it turns out that *somewhere *under the $5 trillion dung heap of liquidity the Fed shat out during Covid there may actually be some limit to the amount of money investors have, and are willing to pay for an AI company based on the promise of extraordinary profits at some unspecified point in the future. Who could have seen that coming?
Just a few weeks ago, I wrote about the potential ramifications of disappointing public-market debuts for OpenAI and Anthropic, specifically the companies whose valuations and investment theses have become increasingly intertwined with the continued expansion of the AI ecosystem.
The point wasn’t simply that one or two highly anticipated IPOs might disappoint. It was that those IPOs could become important price-discovery events for an entire industry that has spent years marking itself higher on private transactions, strategic investments, and increasingly aggressive assumptions about future demand.
If public-market investors begin questioning whether Anthropic deserves a $1.5 trillion valuation, what does that imply for the companies financing its growth, the infrastructure providers counting on its continued expansion, or the venture investors and strategic partners whose own valuations have been supported by the assumption that the AI ecosystem will continue growing exponentially?
These are not isolated questions. They’re all connected, and once investors begin questioning the valuations at the top of the food chain, those questions have a tendency to work their way down.
But there’s something else I think is even more important this morning: yesterday was the crash…but the next several days are when we start finding out who actually got hurt. What do I mean?