When a company is already the most valuable in the world, it takes a lot to surprise anyone. But Nvidia (NVDA 1.00%) CEO Jensen Huang managed to do it last week, and he did it with a single word.
Closing out the chipmaker’s fiscal first-quarter earnings call, Huang put any concerns about the demand environment to rest.
“Demand has gone parabolic,” he said. The reason, in his telling, was simple: agentic AI — systems that can reason, plan, and carry out tasks on their own rather than just answer a prompt — had finally arrived and started doing real work.
It’s a loaded word. A parabolic curve doesn’t just rise; it bends more steeply as it climbs. And paired with what Nvidia actually reported, “parabolic” may end up being one of the more consequential descriptions of the AI cycle this year.
The question it leaves hanging is the one investors keep circling back to: are we near the top of this build-out, or in the middle?
Based on Huang’s description, as well as Nvidia’s continued business acceleration, there’s probably still plenty of room to run.
Image source: Getty Images.
A quarter that backs up the word
The numbers gave Huang’s word some teeth.
Revenue in Nvidia’s fiscal first quarter of 2027 (the period ended April 26, 2026) reached $81.6 billion, up 85% from a year earlier. That figure matters less in isolation than in motion: revenue growth had decelerated for much of the prior year, then turned and accelerated again. Growth came in at 73% in the prior quarter and 85% in the most recent one.
The main driver, of course, remains Nvidia’s AI-driven data center business, where revenue climbed 92% year over year to $75.2 billion — itself an acceleration from 75% growth in the prior quarter.
Profit growth truly was parabolic. The chipmaker’s non-GAAP (adjusted) earnings per share rose 140% year over year to $1.87, helped by both soaring revenue growth and year-over-year gross margin expansion.
Then there was Nvidia’s capital return, which marked an aggressive step up. Nvidia raised its quarterly dividend from a token $0.01 per share to $0.25 — a 25-fold increase — and authorized another $80 billion in share repurchases on top of the roughly $39 billion it had left. Indeed, the company returned about $20 billion to shareholders during the quarter alone, against free cash flow that approached $49 billion.
Today’s Change
(-1.00%) $-2.14
Current Price
$212.11
Key Data Points
Market Cap
$5.1T
Day’s Range
$211.90 – $217.80
52wk Range
$135.40 – $236.54
Volume
7.3M
Avg Vol
166.4M
Gross Margin
74.15%
Dividend Yield
0.02%
The bet behind the buyback
And this parabolic momentum looks like it’s going to continue.
Management reiterated its view that Blackwell and the coming Vera Rubin platform together represent about $1 trillion in revenue visibility from 2025 through calendar 2027. To meet it, Nvidia is committing enormous sums ahead of the orders: total supply commitments, including inventory and prepaid purchases, swelled to about $145 billion, and inventory alone rose to $25.8 billion from $21.4 billion in just three months.
Additionally, chief financial officer Colette Kress added a smaller but telling detail on the call: the rental price of an older H100 chip in the cloud has risen about 20% so far this year.
Given this backdrop, it may not be surprising to hear that Nvidia’s guidance calls for revenue in the current quarter to rise about 95% year over year — and this assumes no data center compute revenue from China at all.
Ultimately, Nvidia has front-loaded tens of billions in commitments on the conviction that this looks much more like the middle of the build-out, not a peak. If that read is right, today’s price — with shares near $217 as of this writing, off an all-time high set earlier in May, and trading at a price-to-earnings ratio of about 33 — may look reasonable in hindsight.
But there are risks. If hyperscalers digest their spending and pull back just as that supply lands, the curve could bend the other way. Additionally, some of Nvidia’s biggest customers have their own ambitious custom silicon programs, designed specifically to address AI needs. If these programs gain significant momentum, it could threaten Nvidia’s dominance.
