When Jensen Huang said “Blackwell is sold out; we have entered a virtuous cycle of AI,” the entire tech world shook. After releasing its FY2026 Q3 results, the king of AI compute once again declared with a suffocating string of figures: in the AI era, Nvidia is the rule-maker. Yet, unexpectedly, the stock fell as much as 5% after hours following the report. Behind this—anxiety over lofty valuations, or worry about an AI bubble? This article will dissect this “most explosive in history” earnings report in depth and analyze its far-reaching impact on rivals, downstream manufacturers, and the entire AI compute industry.
Earnings Breakdown: The Numerical Shock Behind $57 Billion
On November 19 (ET), Nvidia announced its FY2026 Q3 results for the quarter ended October 26, 2025. The numbers were so strong they virtually reset every record.
Revenue and Profit
- Total revenue: $57.006 billion, up 62% year over year, up 22% quarter over quarter
- Gross margin: GAAP 73.4%, non-GAAP 73.6%
- Net income: $31.91 billion, surging 65% year over year, up 21% quarter over quarter
- Net margin: as high as 55.9%, astonishing profitability
- EPS: $1.30, above market expectations of $1.25–$1.26
This performance exceeded the average analyst expectations surveyed by FactSet and temporarily quieted earlier doubts about an AI bubble. Meanwhile, the gross margin held at an eye-popping level—down 1.2 percentage points year over year but up 1 point from last quarter. In the capital-intensive chip industry, such a margin is tantamount to a “money printer.” Nvidia is highly optimistic in its guidance for FY2026 Q4, expecting revenue to hit another record at $65 billion (±2%). Its GAAP and non-GAAP gross margins are projected to climb to 74.8% and 75.0%, respectively.
Business Structure
- Data center: $51.2 billion, up 66% year over year, up 25% quarter over quarter, nearly 90% of total revenue
- Gaming & AI PC: $4.3 billion, up 30% year over year, down 1% quarter over quarter
- Professional visualization: $760 million, up 56% year over year
- Automotive & robotics: $592 million, up 32% year over year
From a contribution perspective, the core driver of the company’s high-speed growth remains the GPU product line serving AI infrastructure. While pushing technological change, this business also brings significant economic return. As the key growth engine, the data-center segment continues to contribute the lion’s share. Founder and CEO Jensen Huang emphasized in the report that demand for the new Blackwell-architecture chips is “off the charts,” and cloud service providers’ GPU inventories are sold out—reflecting that market demand for high-performance computing products is far from saturated. In other segments, although consumer graphics card prices keep rising, profit contribution is relatively limited; professional visualization remains solid; and the automotive and robotics segment—long a strategic bet—has begun to show commercial value.
For Rivals: Squeezing Competition, a Widening Gap
Nvidia’s dominance amounts to a “dimensionality-reduction strike” for competitors such as AMD, Intel, and Broadcom.
AMD’s dilemma: Although the MI300X matches the H100 on paper, Nvidia’s CUDA ecosystem has formed a positive flywheel—the more developers use CUDA, the stronger the ecosystem, and the higher the barrier for new entrants. When Nvidia’s single-quarter R&D spend hits $4.71 billion (+39% YoY), AMD’s full-year R&D is only $5 billion. In this arms race, AMD is already a step behind.
Intel’s disappointment: In the AI training-chip market, Intel’s Gaudi series is virtually absent. More fatal still, via the bundled Grace CPU+GPU strategy, Nvidia is eroding Intel’s share of the data-center CPU market. CEO Pat Gelsinger once boasted “five nodes in four years,” but on the AI compute battlefield, Intel has become a supporting actor.
Broadcom/Marvell’s ASIC route: These vendors focus on custom AI chips, but Nvidia’s “general-purpose GPU + software ecosystem” approach has greater economies of scale. When Nvidia’s quarterly revenue breaks the $50-billion barrier, the entire ASIC market combined is still less than one-third of that. Jensen Huang’s subtext is clear: general-purpose computing is the future; bespoke is a dead end.
Capital markets voted with their feet: after the report, AMD fell 2.4% after hours, Intel fell 1.8%, and although Nvidia pulled back in the short term, its year-to-date gain still exceeds 200%.
For Downstream Manufacturers: Sweet Dependence, Risky Symbiosis
Nvidia’s strength makes downstream players both love it and fear it.
For tech giants, Nvidia has become an unavoidable “compute tax.” Microsoft Azure, Google Cloud, and AWS are all hoarding H100s and Blackwells. At the same time, Google Cloud and Microsoft Azure are building super compute clusters with hundreds of thousands of GPUs. Nvidia seems to have become an indispensable link in these giants’ business ecosystems. Meta, OpenAI, and others are spending billions of dollars every quarter on Nvidia chips—about half of their capital expenditures. More awkwardly, their in-house AI chips—like Google’s TPU and Amazon’s Trainium—still lag Nvidia in performance and can only partially substitute in inference and other scenarios.
For AI startups, compute costs are a chokehold. Star startups like Anthropic and Cohere spend most of their financing on Nvidia GPUs. Nvidia also invests in some AI startups—such as Inflection AI—to lock in customers further. Although compute-rental prices have fallen somewhat, they remain the biggest cost item for AI startups.
Enterprise customers are sinking into path dependency. From Tesla’s FSD to Bloomberg’s financial models, from Adobe’s creative tools to ServiceNow’s enterprise AI, virtually all large-scale AI applications are built on CUDA. Migration costs are extremely high, and customer lock-in is extremely strong.
Industry Impact: AI Compute May Enter an “Ultra-Monopoly” Era
With single-quarter revenue topping $57 billion and R&D at $4.7 billion, Nvidia has formed a flywheel of “high revenue → high R&D → stronger products → higher market share.” For competitors to break this cycle is as hard as reaching the sky.
Nvidia is not merely selling chips; it sells a three-in-one combo of “hardware + software + ecosystem.”
Hardware: GPUs, NVLink, InfiniBand networking
Software: CUDA, TensorRT, Triton Inference Server
Ecosystem: DGX Cloud, AI Enterprise partner programs
This package keeps gross margins at or above 73%, far surpassing traditional chip vendors. Nvidia’s market share in AI training chips already shows clear monopolistic characteristics. With Blackwell’s rollout, Nvidia’s share in AI training chips may climb further.
Risks and Controversies: Can High Growth Last?
Despite stellar results, the after-hours drop exposed market worries. The Q4 revenue guide of $65 billion beats expectations, but can the growth rate be sustained? Wall Street fears an “S-curve” inflection in AI demand. Moreover, if AI application rollouts proceed more slowly than expected, a destocking cycle could be triggered.
Second, the China market issue is particularly acute. Jensen Huang noted in October that, due to U.S. export controls, Nvidia’s share of China’s high-end AI-chip market fell from 95% to 0%. Colette Kress stated that, affected by geopolitical issues and intensified competition in China, large purchase orders did not materialize this quarter.
A quarterly revenue of $57 billion, $51.2 billion from data centers, and 13,000 Blackwell samples delivered—behind these numbers lies the footnote of an era: the rules of the AI compute game are defined by Nvidia. For rivals, it’s the cruel reality of squeezing competition; for downstream players, it’s a sweet yet dangerous deep binding; for the industry, it’s an accelerated concentration and a winner-takes-most Matthew effect.
However, history tells us: there is no eternal hegemony, only periodic disruption. When everyone believes “this time is different,” a turning point may be quietly brewing. The only certainty is that, on the AI compute battlefield of 2026, Nvidia will still be the absolute king wielding the “Blackwell” cudgel.
[Disclaimer]: The above content reflects analysis of publicly available information, expert insights, and BCC research. It does not constitute investment advice. BCC is not responsible for any losses resulting from reliance on the views expressed herein. Investors should exercise caution.