To The Moon

Here we go again: Nvidia just walked into earnings and flipped the “AI bubble” narrative on its head.

On Nov. 19, Nvidia reported fiscal Q3 (ended Oct. 26) revenue of $57.0 billion and diluted EPS of $1.30—both above Wall Street’s expectations. Data-center sales, the heart of its AI business, hit a record $51.2 billion. The company guided next quarter’s revenue to about $65 billion, plus or minus 2%. Shares popped roughly 4–5% on the news as traders exhaled. 

Two reasons: demand and direction. CEO Jensen Huang said Blackwell chip sales are “off the charts,” and the company sees accelerating compute needs for both training and inference—wonky words that simply mean building AI models and then running them everywhere. He pushed back hard on the “AI bubble” talk, saying Nvidia sees “something very different” from a speculative mania, pointing to real orders and deployments across cloud providers and industry. 

There’s also concentration and confidence. Reuters noted 61% of revenue came from four big customers—a reminder that hyperscalers are still the main buyers. That’s a risk if any one slows spending, but it also shows just how central Nvidia is to their road maps. Gross margins stayed in the mid-70s, signaling pricing power even as volumes surge. 

First, the ripple effect. When Nvidia clears the bar, chip land usually rallies—and that’s what we saw. The VanEck Semiconductor ETF (SMH) and the broader Philly Semiconductor Index bounced after the print, while names like Broadcom caught a bid alongside Nvidia. It’s the classic “rising tide” day for AI hardware. 

Second, the scoreboard matters. Year over year, Nvidia’s revenue rose 62%, while data-center sales climbed 66%. Those aren’t bubble-ish vibes; they’re “customers are still buying as fast as we can build” vibes. And the $65 billion guidance says the pipeline isn’t running dry yet. Still, markets are forward-looking—for a stock this widely owned, any hint of slower orders or a margin wobble can swing hundreds of billions in market cap. 

Third, the questions to keep in your back pocket:

• Can supply keep up? Lead times and energy constraints are real bottlenecks for AI buildouts. If those pinch, ship dates slip. (Nvidia flagged power and capital availability as industry constraints.) 

• Customer mix risk. Heavy reliance on a few hyperscalers cuts both ways: it accelerates growth—and concentrates it. 

• Competition watch. AMD is pushing hard with new data-center GPUs, and custom silicon from hyperscalers is a slow-burn threat. None of that dents tonight’s beat, but it shapes the long game. 

Bottom line: Nvidia didn’t just beat; it reset the conversation. Huang’s “not a bubble” message lands differently when it’s paired with $57 billion in quarterly sales, 73%-plus gross margins, and guidance that tops the Street. The market heard it—and chip stocks moved in unison. If you care about the AI trade, Nvidia’s results remain the bellwether to watch.