“Hopper demand remains strong, and the anticipation for Blackwell is incredible,” Jensen Huang beamed, rhapsodizing in the press release accompanying Nvidia’s fiscal 2025 Q2 results, released to a veritable investor frenzy on Wednesday afternoon in the US.

There’s no sense burying the lede. They beat. On everything. Revenue for Q2 was $30.04 billion, up 123% YoY. The gains are obviously decelerating, but somehow “slower” seems like a misnomer. Consensus was looking for $28.86 billion on the top-line for last quarter.

Data center revenue rose 154% YoY to $26.3 billion, $1.2 billion ahead of estimates. Not that anyone cares, but gaming revenue was $2.9 billion, also a beat. Adjusted gross margin was 75.7%, operating income of $19.94 billion was more than a billion higher than expected, non-GAAP EPS of ¢68 beat by a few pennies and expenses managed to undershoot.

Crucially, the revenue guide was ahead of the Street, albeit not by a huge margin. Nvidia sees $32.5 billion in sales for the current quarter, plus or minus 2%. Analysts were looking for $31.85 billion there. The adjusted margin guide was in-line at 75%, as was the expense outlook at $3 billion.

Don’t ask me if all that’s “good enough.” I have no idea. And neither, frankly, does anybody else. The market will make that determination collectively.

Nvidia has recorded so many enormous single-session gains and losses over the course of the last 15 months that it’s scarcely worth trying to chart them anymore. Suffice to say $100 billion days are just par for the course. The last two earnings reports were followed by $200 billion sessions.

Gun to my head, I’d tentatively suggest the initial release was as much of a non-event as it could’ve been in the context of earnings reports from the most important company on the planet. There were no real surprises that I could see, and the numbers left the impression that the company’s momentum is intact, even as the revenue growth rate continues to recede now that Nvidia’s lapping its own insane comps.

It’s possible the shares see a relatively muted post-earnings move on Thursday. Note the emphasis. Anything other than a $150 billion swing would arguably constitute a pedestrian gain/loss. That said, if the after-hours selloff holds, Nvidia would suffer a very large value destruction event tomorrow.

The real question going forward is what reports of Blackwell delays mean for the outlook beyond the current quarter. “Blackwell samples are shipping to our partners and customers,” Huang said, before the call. Color from CFO Colette Kress said the company “executed a change… to improve production yield.” The production ramp for the chip is “scheduled to begin in the fourth quarter and continue into fiscal 2026.”

Speaking to investors, Kress said Nvidia should see “several billion dollars” of Blackwell revenue in Q4. Demand outstrips supply, she went on. Separately, she noted that data center revenue in China fell versus pre-export controls levels, but grew sequentially and was a “significant contributor” to total sales.

However investors decide to treat the Blackwell remarks, the numbers Nvidia posted on Wednesday were fine. To be clear: They’re a lot more than “fine,” but the bar’s so high for Nvidia that it’s very difficult, even for industry specialists, to determine what counts as a “beat.” If you’re Nvidia, beating company analysts’ estimates is just one hurdle. You also have to clear the bar investors set, and they (investors) aren’t always rational.

Of course, Huang’s fine with indulging the excitable masses, often employing difficult-to-quantify — and in some cases wholly nebulous — superlatives to describe demand . “[G]lobal data centers are in full throttle,” he said Wednesday, on the way to reiterating, for the three trillionth time, that “generative AI will revolutionize every industry.”

For good measure, the company said it’ll buy back $50 billion more in shares.