Nvidia pulled in $81.6 billion in revenue for the quarter ending April 2026, up 85% from a year ago and beating analyst expectations. For the second quarter, the company forecasts $91 billion, a figure that does not include any chip sales to China.
But the headline figures were not the most significant thing CEO Jensen Huang said on the earnings call. The more consequential announcement was a processor that Nvidia is aiming at a market it has never competed in.
The Vera CPU and Why It Matters
During the earnings call, Huang described Vera as the “world’s first CPU, purpose-built for agentic AI,” marking a shift for Nvidia as it enters a market traditionally dominated by Intel and AMD. Vera was first introduced at Nvidia’s annual GTC conference in March and can be purchased as a standalone product or bundled with the upcoming Rubin GPU.
The distinction Huang drew between Vera and existing cloud processors is that traditional cloud CPUs are built to handle multiple application instances as quickly as possible, using many cores, while Vera is optimized to process tokens, the fundamental units of data in AI models, as fast as possible. This design makes it specifically suited for agentic AI systems, where AI agents execute tasks autonomously rather than simply generating responses.
Huang stated on the call that “every major hyperscaler and system maker is partnering with us to deploy it,” and that “the world is rebuilding computing for agentic AI and robotic physical AI.” He added that as billions of AI agents come online, each will need its own processing tools, drawing a direct parallel to how human workers rely on personal computers today.
Another Record Quarter, Driven Almost Entirely by Data Centers
Data Center revenue reached $75.2 billion in the quarter, rising 92% from a year ago and 21% sequentially, driven by adoption of Blackwell 300 products and strong demand for InfiniBand, Spectrum-X Ethernet, and NVLink solutions. Data center networking revenue reached $14.8 billion, up 199% from a year ago. These figures confirmed that the Blackwell architecture, which Nvidia began shipping in volume last year, is still in high demand across cloud providers and enterprise customers.
Nvidia also authorized $80 billion in additional stock buybacks and raised its quarterly cash dividend to $0.25 per share, up from just $0.01, with the payment scheduled for June 26. The company returned approximately $20 billion to shareholders in the first quarter through buybacks and dividends.
The scale of that capital return is a sign that management is confident in the company’s near-term cash flows. The Q2 forecast revenue of $91 billion excludes any Data Center compute revenue from China, meaning any policy shift on China shipments would represent additional upside not currently built into the forecast.
$20 Billion in Vera Revenue Projected for This Fiscal Year
Nvidia expects $20 billion in revenue from Vera chip sales by the end of this fiscal year, which Huang confirmed was not included in the company’s earlier $1 trillion estimate of Blackwell and Rubin AI chips between 2025 and 2027. He also described Vera as likely to become the second-largest sales contributor beyond that $1 trillion figure, and tagged the total addressable market at $200 billion.
Whether Vera performs at the scale Huang projects will depend on how quickly agentic AI workloads mature and how many customers choose Nvidia’s processors over those being developed in-house by the cloud giants.
