During the $4.8-trillion-valuation chip supplier’s earnings call on Wednesday, analysts pressed Huang on whether major cloud customers—whose capital expenditures are nearing $700 billion a year—could keep up the pace. According to Huang, it’s a no-brainer. In the new AI-based economy, compute and revenue are essentially the same thing. Without the capacity to generate AI tokens, which are the small chunks of chatbot outputs in the form of words and text, cloud providers don’t have a way to meaningfully grow.

“I am confident in their cash flow growing,” said Huang, in response to a question. “And the reason for that is very simple.

“We have now seen the inflection of agentic AI and the usefulness of agents across the world and enterprises everywhere, and you’re seeing incredible compute demand because of it,” Huang continued. “In this new world of AI, compute is revenues. Without compute, there’s no way to generate tokens. Without tokens, there’s no way to grow revenues.”

So the hundreds of billions’ worth of capital expenditures now flow into AI, which eventually translates into growth, which translates “directly to revenues,” said Huang.

Nvidia offered AI investors a glimpse of a hairpin-turn recovery with its results for the fourth quarter and the full year of fiscal 2026, with results showing record revenue of $68.1 billion for the quarter, beating guidance by about $3 billion. Those numbers were up 20% from the third quarter and a whopping 73% from a year ago. 

Notably, the company released guidance for the first quarter of fiscal 2027 of $78 billion. Total supply-related commitments rose from $50.3 billion at the end of the third quarter to $95.2 billion at the end of the fourth quarter. In a statement, Nvidia said it has “strategically secured inventory and capacity to meet demand beyond the next several quarters.”

Going into results, investors were primed for any sign—a sigh, a hesitation, anything—that might indicate that its gross margins might be slipping further. Previous guidance had called for a 74.8% GAAP gross margin, which would signal a partial recovery, and Huang and chief financial officer Colette Kress have said the goal going into fiscal year 2027 is to hold margins “in the mid-70s.” 

On cue, investors kept a gimlet-eyed focus on those figures on Wednesday. And Nvidia did not disappoint. The company’s GAAP gross margin rose to 75%, beating guidance and up from 73.4% in Q3, and the non-GAAP gross margin clocked in at 75.2%. Nvidia’s stock rose more than 2% in the first phase of after-hours trading, though it quickly gave back much of those gains.

In all, GAAP net income was up 35% quarter over quarter and 94% year over year to roughly $43 billion. GAAP diluted earnings per share came in up 35% at $1.76 for the quarter and nearly double compared with fiscal 2025. Net income also saw a bump related to Nvidia’s investment in Intel stock. Non-GAAP income, which doesn’t include the Intel investment gains, came in at $39.6 billion. 

The Nvidia earnings results come amid a high-stakes backdrop of fears about AI overinvestment, in the form of eye-popping capital expenditures (capex) among hyperscalers including Amazon, Meta, Microsoft, Oracle, and Alphabet, which are locked in a frenzied AI race. A recent report from Moody’s flagged that some $662 billion in future data center lease commitments that have not yet begun remain off those companies’ balance sheets. 

“Computing demand is growing exponentially,” said Huang in a statement. “Enterprise adoption of agents is skyrocketing. Our customers are racing to invest in AI compute—the factories powering the AI industrial revolution and their future growth.”

For Nvidia, of course, a portion of that capex spending winds up in the company’s coffers to pay for its highly coveted—and premium-priced—chips. 

Full-year revenue also soars

For the full year, Nvidia revenues hit $215.9 billion, up 65% from last year; GAAP operating income was $130.4 billion, and net income was $120.1 billion. In comparison, in fiscal year 2025, which ended in January 2025, Nvidia posted $130.5 billion in revenue, more than doubling the year prior’s $60.9 billion. Net income for that year was $72.9 billion, and operating income more than doubled over the year before to $81.5 billion. Data center revenues for fiscal 2026 were $197.3 billion, up from $115.2 billion the previous year. 

Across fiscal year 2026, revenues rose each quarter, from $44.1 billion in Q1, to $46.7 billion in Q2, to $57 billion in Q3, and now to $68.1 billion in Q4.

Last quarter, Huang directly attempted to quash fears about frothiness in the market on the Q3 call with analysts. 

“There’s been a lot of talk about an AI bubble,” said Huang last quarter. “From our vantage point, we see something very different.”

He said the industry has undergone three structural platform shifts: from traditional CPUs to GPU-driven computing; from traditional machine learning to generative AI; and from generative AI to agentic AI. Each transition, on its own, justifies massive investments. Huang said the first two shifts were fully funded through cost reductions and revenue growth, while agentic AI is a new layer on top that will require investment. 

CFO Kress said last quarter that Nvidia had “visibility” to $500 billion in revenue from its Blackwell and Rubin offerings from the start of the 2025 calendar year through the end of the 2026 calendar year. Kress also said that Nvidia believes total AI infrastructure investment could reach $3 trillion to $4 trillion annually by 2029 or 2030.