Nvidia recently announced a significant financial performance for its fourth quarter, reporting a revenue of $68 billion, marking an impressive 73% increase compared to the same period last year. This surge in revenue has been largely attributed to the robust demand for accelerated computing solutions prevalent among cloud providers, enterprises, AI model developers, and governmental entities. A notable segment of this growth has come from the data center revenue, which soared by 75% to $62 billion—an increase of 22% from the previous quarter. The company’s leadership attributes this trend to a growing reliance on GPU-based infrastructure, driven primarily by advancements in generative and agentic artificial intelligence.
During the earnings call, CEO Jensen Huang emphasized that the economics of AI are increasingly tied to computational capacity and efficiency. He stated, “In this new world of AI, compute equals revenues. Without compute, there’s no way to generate tokens, and without tokens there’s no way to grow revenues.” Furthermore, he highlighted that real-time inference is becoming a pivotal component in the monetization strategies of cloud services, amplifying the need for performance per watt as companies ramp up their production workloads.
Addressing potential uncertainties in future spending from hyperscalers, Huang expressed optimism for sustained demand. He noted a pivot towards agentic workloads, suggesting that this shift marks a significant turning point for compute demand, with expectations for rising cash flows in these areas. While hyperscalers account for just over half of Nvidia’s data center revenue, there has been a notable acceleration in growth from enterprises, AI model creators, and sovereign customers.
Nvidia also showcased its ecosystem’s strengths, particularly through its CUDA platform, which maintains compatibility across multiple generations of GPUs. Huang mentioned, “We’re the only accelerated computing platform in every cloud, and CUDA makes the entire installed base better with every generation.” The company presented improvements in inference performance and cost efficiency stemming from software optimizations that enhance older hardware capabilities.
Additionally, the demand for sovereign AI solutions has surged, with revenue exceeding $30 billion for the fiscal year as numerous nations invest in building national AI infrastructures. Huang also discussed the innovative prospects of early GPU deployments in space, citing imaging and on-board processing applications as promising avenues.
Looking forward, Nvidia’s Rubin platform, intended to optimize model training using one-fourth the GPUs while significantly lowering inference costs, is on track for release in the latter half of the year. CFO Colette Kress confirmed that the company has secured long-term supply commitments to ensure it meets the anticipated demand.
Key highlights from the earnings call included Huang mentioning that Nvidia is nearing completion of a new partnership with OpenAI, particularly noting the advancements of GPT-5.3 Codex, which has been deployed on their Grace Blackwell platform. The company reported accelerated growth outside hyperscalers, with various sectors such as enterprises, government bodies, and telecommunications contributing increasingly to data center demand.
Furthermore, Nvidia outlined the integration of the Grok inference team, likening it to the strategic enhancement seen with Mellanox, anticipating new low-latency features as part of their full-stack platform approach. The company reiterated its strategic focus on long-term supply commitments and investment in ecosystem development, prioritizing these over immediate stock buybacks despite strong cash generation.
In terms of financial metrics, Nvidia closed the quarter with a gross margin of 75% on a GAAP basis and a free cash flow of $35 billion for the quarter, contributing to an annual total of $97 billion. Shareholder return strategies included $41 billion returned through buybacks and dividends. Looking ahead, Nvidia forecasts its revenue for the first quarter of fiscal 2027 to reach approximately $78 billion, signaling continued strong performance expectations.
