Nvidia reported record fourth-quarter revenue of $68.13 billion, representing a 73% increase from the prior year and 20% sequential growth. This surge was primarily fueled by the Data Center segment, which reached an all-time high of $62.31 billion as global demand for AI infrastructure accelerated. GAAP earnings per diluted share (EPS) nearly doubled year-over-year to $1.76, up 98% from $0.89 in the year-ago period. Gross margins remained exceptionally robust at 75.0%, reflecting significant pricing power and the successful ramp-up of the Blackwell architecture despite ongoing supply chain complexities.
Note: Nvidia's FY'25 ended on January 26, 2025. Q4 FY'26 ended on January 25, 2026.
Nvidia has successfully transitioned its product roadmap to an annual cadence, moving from the Hopper architecture to Blackwell and announcing the next-generation Vera Rubin platform. The company reported that Blackwell Ultra is now in full-scale production, delivering up to 50x better performance for agentic AI compared to the previous generation. Strategic partnerships have expanded significantly, including a multiyear agreement with Meta for large-scale deployment of millions of Blackwell and Rubin GPUs. Furthermore, Nvidia announced a massive partnership with OpenAI to deploy 10 gigawatts of systems for next-generation AI infrastructure, solidifying its role as the primary engine for the AI industrial revolution.
Below are key drivers of Nvidia's value that present opportunities for upside or downside to the current Trefis price estimate:
For additional details, select a driver above or a division from the interactive Trefis split for Nvidia at the top of the page.
Nvidia is the global leader in accelerated computing, providing the hardware and software stack essential for generative AI, high-performance computing, and graphics. The company operates through two primary reporting segments: Compute & Networking, which includes data center platforms and networking systems, and Graphics, which encompasses gaming and professional visualization. By leveraging its proprietary CUDA software platform and a rapid hardware release cycle, Nvidia has created a powerful ecosystem moat that makes it the preferred provider for hyperscalers, enterprises, and AI-native startups.
The Data Center segment is Nvidia's most valuable asset, contributing over 90% of total revenue and the vast majority of operating profit.
Nvidia's primary competitive advantage is its integrated full-stack approach, combining world-class silicon with the CUDA software ecosystem. This creates high switching costs for developers who have optimized their AI models for Nvidia's architecture. The company's ability to provide end-to-end systems, including GPUs, CPUs, and networking, allows it to capture a larger share of data center capital expenditure than any pure-play chip competitor.
The rapid transition to Blackwell and Blackwell Ultra has widened the performance gap over rivals. These systems are designed for trillion-parameter scale generative AI, offering significantly lower total cost of ownership (TCO) for cloud providers by reducing power consumption and increasing inference speed. This technological lead allows Nvidia to maintain gross margins in the mid-70% range even as competitors launch their own AI accelerators.
The industry is shifting from basic generative AI to "agentic" AI—systems that can reason and execute multi-step tasks—and "physical" AI, which involves robotics and autonomous systems. Nvidia is positioning its Grace Blackwell platform as the "king of inference" for these workloads. This trend expands Nvidia's reach from cloud training into real-time enterprise operations and industrial automation, representing a massive new addressable market.
Nvidia has significantly increased its capital return program, returning $41.10 billion to shareholders in fiscal 2026 through repurchases and dividends. With $58.50 billion remaining in its share repurchase authorization as of January 2026, the company is using its massive free cash flow to support its stock price and signal confidence in the long-term sustainability of the AI infrastructure build-out.
A notable strategic shift in 2026 is the diversion of wafer and memory supply away from Gaming toward high-margin AI Data Center production. For the first time in three decades, Nvidia did not launch a new gaming GPU refresh at CES, leading to supply constraints and price hikes in the consumer market. This trend confirms that management is prioritizing the AI boom over its legacy gaming business to maximize return on silicon.
