Nvidia Soars $440 Billion in a Day, Then Loses Most Gains Fast

Nvidia had a wild week. Shares jumped 8.74% on August 27, a rare post-earnings pop for a company that’s basically made a habit of beating Wall Street and then watching its stock slide anyway. The rally added roughly $440 billion in market value in a single session. Then it evaporated — mostly.

By the following day, the stock had dropped 4.57%, closing at $217.55. About $250 billion of those gains were gone by week’s end. Fast money, fast exit.

The Earnings Numbers Were Genuinely Big

Nvidia’s quarter wasn’t close. Revenue hit $96.2 billion — up 106% from the same period a year earlier. The company then projected $108 billion for the next quarter. Those aren’t numbers you see often from a company already at this scale. And yet, for most of the past year, Nvidia’s post-earnings pattern ran the same way every time: beat the number, stock falls. Traders had basically priced in perfection, and perfection kept not being enough.

Before the August 27 surge, Nvidia shares had already slid from $225.30 to $208.48 over seven sessions between August 14 and August 24. The stock was leaking heading into the print. Then the report dropped, shares briefly hit $227.98, and for one day it looked like the pattern had finally broken.

It hadn’t really broken. Just paused.

Tom Lee’s Take and the Skeptics Waiting in the Wings

Fundstrat’s Tom Lee had an interesting read on what drove the unusual move. He said investors pulled money out of other positions to chase Nvidia’s momentum — a kind of funding problem where the gains came partly at the expense of other holdings. Lee also pointed to something worth sitting with: Nvidia’s price-to-earnings ratio is unusually compressed for a company growing this fast. It’s trading at 27.5 times past earnings, but only 18 times expected earnings. That gap means analyst estimates are rising faster than the stock price is. Lee’s argument is basically that Nvidia’s financial story is outrunning the market’s willingness to price it in.

Salesforce had a strong week too, for what it’s worth. The software company rose 22.58% on its own earnings results, which actually outpaced Nvidia’s single-day gain. Some of the week’s tech enthusiasm clearly spread beyond chips.

More context: Bitcoin Surges Past $80,000, but $617 Billion Still in Losses

But not everyone’s buying the bull case. Jay Goldberg at Seaport Research Partners kept a sell rating on Nvidia. His argument isn’t that the demand isn’t there — it clearly is. The problem, per Goldberg, is supply. Nvidia can’t just flip a switch and produce more chips. Existing commitments eat up capacity, and that ceiling limits how much the company can actually capitalize on surging interest in AI infrastructure. Demand doesn’t matter much if you can’t fill the orders.

Politics Is Now Part of the Nvidia Story

There’s a newer wrinkle that’s probably underappreciated. Opposition to new data center construction has become a legitimate political issue. Data centers are noisy, power-hungry, and increasingly unpopular in some communities. That’s turned into an election-season talking point, and it’s not going away. For Nvidia, whose entire growth thesis depends on AI infrastructure expanding at scale, political friction around data center permitting and construction is a real operational risk — not a hypothetical one.

The company’s chips power the servers that fill those data centers. If the centers don’t get built, or get delayed, that’s a demand problem that has nothing to do with Nvidia’s product quality.

So you’ve got a company reporting 106% revenue growth, projecting $108 billion next quarter, trading at a compressed earnings multiple — and still facing a ceiling on upside because of chip supply constraints and a political environment that’s turning against the infrastructure AI needs to run. That’s a strange combination of factors to hold at once.

More context: Dallas Fed Warns Tokenized Deposits Could Lead to $700 Billion Credit Crisis

Lee’s point about analyst estimate revisions is worth returning to. Nvidia keeps beating, estimates keep rising, and the stock keeps not fully reflecting those revisions. The price-to-earnings compression he flagged is either a buying opportunity or a sign that the market knows something the bulls don’t. Probably both, depending on the week.

The $250 billion that vanished after the initial pop is a pretty clean summary of where investor confidence actually sits right now — interested, not convinced.