“I’m not dead yet!” shouted the AI trade.
Between a $350 billion selloff in Broadcom and a total bloodbath in Oracle, it was looking grim out there for a minute, but Micron was the unlikely savior for a suddenly dicey trade in all things AI- and semi-related.
I didn’t have “huge Micron rally” on my Bingo card this week, but we need to keep things in perspective. (Hold that thought, because as usual I’m burying the lede here.)
The figure below shows that at the intraday highs, the stock’s Thursday gain was indeed among the largest in years, which is naturally what financial media headlines picked up.
The catalyst was obviously the company’s guide which was, in a word, absurd relative to consensus. Current-quarter sales will be $18.7 billion taking the midpoint versus $14.4 billion expected, and profit should be $8.42, nearly double the average Street guess.
As Nomura’s Charlie McElligott put it Thursday, the magnitude of the beat has “only been seen once before in the chips space and that was the initial Nvidia earnings paradigm shift in May 2023.”
Management styles itself (not inaccurately) an “AI enabler” through the storage business. One of the company’s executives did his best Jensen Huang impression following the earnings release. “We are more than sold out,” he said, marveling at the size of “unmet demand” which is “substantially higher than supply for the foreseeable future.”
Another high-level team member took the hyperbole up a notch in remarks to Bloomberg, describing “the most significant disconnect between demand and supply that we’ve experienced in my 25 years in the industry.”
That’s all fine and good, but Micron was only a $300 billion company even after a — checks notes — 335% rally from the “Liberation Day” lows. By noon on Thursday, the “mammoth” one-day surge was a whole $30 billion in market cap terms. Whoop-de-do. That’s like Nvidia’s bid-ask.
Jokes aside, I don’t want to downplay Micron’s upbeat forecast, particularly given the sheer scope of the upside surprise on the guide. But it’s well established that the industry has a shortage of memory and that’s what Micron does. Of course, analysts were aware of that when they reckoned their forecasts, so demand’s plainly a lot more robust than industry watchers were inclined to believe.
We shouldn’t forget that if, as CEO Sanjay Mehrotra claimed, Micron can only fill two-thirds of orders on a good day, that means spending more on capacity, in this case 45% more in fiscal 2026 versus 2025, doing some simple math with the capex guide.
Anyway, don’t look the gift horse in the mouth, I suppose. McElligott summed it up well. “There’s still blood left in this AI stone,” he wrote Thursday, flagging “notable willingness to sell downside in some key AI equities proxies” following the correction, including Broadcom and Nvidia.




Or, are we just playing whack a mole in an overpriced graveyard? Time will tell…
I expect traders to continue to have a whole lot of fun with volatility in both directions all over the tech universe (and even expanding further into the picks & shovels and other realms collateral to the AI revolution …and then to collaterals to the collaterals; like everybiz.com 25 years ago!).
To me, it seems more like WS is trying to scare me into selling my shares of certain AI/tech companies, including Avgo.
Hock Tan has missed consensus EPS estimates only 5 quarters since January, 2015; all relatively small misses. He has more than proven that he knows how to guide the analysts to the correct range of sales/earnings.
With respect to the PE, I’d be more concerned if the Federal and state budgets were going to be balanced in 2026.
Was his name-o
Have to comment that the artwork for this one is superb.