H200 has not yet generated revenue in China, while Nvidia's Q4 revenue surged 73%, Jensen Huang says: Computing power equals revenue!

On February 26, NVIDIA announced its Q4 FY2026 financial results.

The report shows that for the quarter ending January 25, 2026, NVIDIA’s revenue reached a record $68.1 billion, up 20% from the previous quarter and 73% year-over-year. FY2026 revenue totaled $215.9 billion, a 65% increase year-over-year.

This quarter, GAAP and non-GAAP gross margins were 75.0% and 75.2%, respectively. For FY2026, GAAP and non-GAAP gross margins were 71.1% and 71.3%.

Looking ahead to Q1 FY2027, NVIDIA expects revenue of $78 billion, plus or minus 2%. The outlook does not include revenue from data centers in China. The expected GAAP and non-GAAP gross margins are 74.9% and 75.0%, plus or minus 50 basis points, including a 0.1% impact from stock-based compensation.

Image source: Screenshot from NVIDIA’s official website

Before releasing this financial report, doubts about an AI bubble were widespread. According to Global Times citing foreign media, the Sittrini Research Company recently published the “2028 Global Intelligence Crisis” report, which describes the threat of artificial intelligence to the U.S. economy and the potential for a “global intelligence crisis.” While some experts believe the market overreacts to the extreme scenarios in the report, it clearly highlights that investors are now highly alert to the “AI bubble.”

During the earnings call, when asked about cash flow growth in AI companies, NVIDIA founder and CEO Jensen Huang emphasized his confidence. He believes the turning point for Agentic AI has arrived, with exponential growth in computing demand.

Huang stated that in this new AI era, computing equals revenue. With products like Codex, Claude Code entering commercial use, and the high market recognition of Claude Cowork and OpenAI Enterprise, various enterprise software developers (ISVs) are building agent-based systems on their platforms. The industry is at a critical turning point. The tokens generated now can improve customer productivity and generate profits for cloud service providers.

Regarding specific business segments, NVIDIA’s data center revenue in the fourth quarter hit a record $62.3 billion, up 22% from the previous quarter and 75% year-over-year, driven by accelerated computing and AI. Full-year revenue increased 68% to a record $193.7 billion.

NVIDIA CFO Colette Kress mentioned during the earnings call that the data center revenue growth in Q4 was mainly due to the continued strong ramp-up of Blackwell and Blackwell Ultra capacity. Due to high demand for NVIDIA infrastructure, even Hopper and many products based on the six-year-old Ampere architecture are sold out in the cloud.

Kress stated that since the launch of ChatGPT in FY2023, NVIDIA’s data center business has expanded nearly 13-fold. Looking ahead, she expects quarterly revenue to grow sequentially throughout FY2026, surpassing last year’s $500 billion Blackwell and Rubin revenue opportunities. She also emphasized that NVIDIA is prepared with inventory and supply commitments to meet future demand, including shipments extending into FY2027.

Additionally, gaming and AI PC business revenue in Q4 was $3.7 billion, up 47% year-over-year. NVIDIA said that after strong holiday demand, channel inventory naturally slowed. For the full year, this segment’s revenue grew 41% to a record $16 billion.

The professional visualization segment earned $1.3 billion in Q4, up 74% quarter-over-quarter and 159% year-over-year, driven by the special demand for Blackwell. The full-year revenue for this segment increased 70% to a record $3.2 billion.

Automotive and robotics revenue in Q4 was $604 million, up 2% quarter-over-quarter and 6% year-over-year, benefiting from continued adoption of NVIDIA’s autonomous driving platform. Full-year revenue for this segment reached a record $2.3 billion, up 39%.

Recently, NVIDIA has partnered with several tech giants including Anthropic, Meta, OpenAI, and xAI. During the call, Huang confirmed that NVIDIA is advancing cooperation agreements with OpenAI and is close to finalizing them. “This is a once-in-a-decade company, and we are honored to have worked with them since day one.”

Previously, NVIDIA and Meta announced a strategic partnership. Meta will deploy millions of Blackwell and Rubin series GPUs, NVIDIA CPUs, and Spectrum-X Ethernet for model training and inference.

At the same time, NVIDIA announced a partnership with Anthropic, investing $10 billion. Huang mentioned that Anthropic’s Claude Cowork platform is revolutionary, significantly expanding enterprise AI applications. The “ChatGPT moment” for agent-based AI has arrived.

Additionally, NVIDIA recently signed a non-exclusive licensing agreement with Grok to access its low-latency inference technology and onboard its core engineers. NVIDIA plans to leverage Grok’s technological innovations, similar to the successful integration of Mellanox, to further expand its architecture.

Regarding the Chinese market, Kress revealed that although the U.S. government approved a small number of H200 products for Chinese customers, no final revenue has been generated yet, and it’s unclear whether any products will be allowed to be imported into China in the future. “Our competitors in China have made progress with recent IPO support, and in the long term, they have the potential to disrupt the global AI industry structure. To maintain its leadership in AI computing, U.S. companies must attract every developer and become the preferred platform for all businesses, including Chinese companies. We will continue engaging with U.S. and Chinese governments and advocate for U.S. competitiveness globally.”