Nvidia Q1 FY2026: $81.6B Revenue, $75.2B Data Center Record, $43B Startup Stakes, $80B Buyback

Q1 revenue reached $81.6B, up 85% YoY and 20% quarter-over-quarter, beating expectations

Wall Street consensus had targeted approximately $78.9B–$80B. The beat confirms AI infrastructure spending among major tech companies remains robust and sequentially accelerating.

Data center revenue hit a record $75.2B in Q1, the dominant driver of total results

Data center now constitutes the overwhelming majority of Nvidia's business. CFO Colette Kress noted Blackwell architecture has been adopted by every major hyperscaler, cloud provider, and model maker.

Net profit hit $58.3B, up 211% YoY; adjusted EPS of $1.87 beat the $1.75 analyst estimate

211% profit growth on a large base. Nvidia's entire quarterly profit was approximately $2B just three years ago — a 29x increase.

Q2 revenue guidance of $91B exceeded Wall Street's ~$87B prediction by roughly 4–5%

Analyst consensus ranged from $86B to $87.29B. Strong forward guidance signals growth trajectory remains intact into fiscal Q2 FY2026.

Nvidia authorized $80B in new share repurchases and raised its quarterly dividend from $0.01 to $0.25 per share

The company ended Q1 with $38.5B remaining under its prior buyback authorization. The dividend increase — 25x the prior level — signals management's confidence in sustained cash generation.

Nvidia's privately held startup stakes nearly doubled from $22B to $43B within a single quarter

$18.5B in new purchases of non-marketable equity securities during Q1, versus just $649M the prior quarter. This figure excludes the $30B OpenAI commitment made in February and investments in publicly traded companies like Corning and IREN, as well as future commitments not yet closed.

Jensen Huang disclosed a significant pending compute buildout with Anthropic, describing prior coverage as 'largely zero'

Huang told investors the capacity coming online for Anthropic this year and next 'is going to be quite significant,' marking a new major customer relationship that was not previously public.

Vera Rubin, Nvidia's next-gen architecture built for agentic AI, launches in Q3 2026 with production ramping through H2

Huang said Vera Rubin is 'off to a tremendous start' and will surely be more successful than Grace Blackwell. The company expects to be supply-constrained throughout Rubin's lifecycle — consistent with the pattern seen with Blackwell.

Nvidia restructured its reporting from five segments to two: Data Center and Edge Computing

The previous five segments were Data Center, Gaming, Professional Visualization, Auto, and OEM & Other. Data Center breaks into Hyperscale/AI Clouds and Industrial/Enterprise subgroups; Edge Computing covers PCs, consoles, workstations, wireless, robotics, and automotive.

Chinese export restrictions had no material impact on Q1 earnings; future China revenue remains uncertain

CFO Kress stated H200s have been approved for U.S. export but Nvidia has 'yet to generate any revenue' from China and is uncertain whether imports will be permitted.

Nvidia is now the most valuable publicly traded company in the world

AMD, Broadcom, Micron, and TSMC all have significant exposure to AI infrastructure spending, but Nvidia's scale and margin profile are in a class apart.

Demand constraint has replaced demand uncertainty as Nvidia's primary business challenge

No pause in Blackwell purchases while awaiting Rubin. Nvidia's biggest problem is now fulfilling orders — a constraint Huang expects to persist into the Vera Rubin generation.

Nvidia's inferencing market share is growing quickly, expanding its relevance beyond training workloads

Huang noted on the investor call that Nvidia's share of AI inferencing is 'growing very quickly,' suggesting the company is capturing value across the full AI compute stack, not just model training.