Skip to content
    NVDA
    Earnings call· Oct 2025(Q3 FY26)

    NVIDIA CORP NVDA

    Nov 19, 2025 Source

    Executive summary

    NVIDIA Q3 FY26 — Record Revenue and Strong AI Demand

    NVIDIA delivered record Q3 FY26 results, driven by robust demand for its accelerated computing and AI platforms across hyperscalers and enterprises. The company highlighted strong momentum for Blackwell and anticipated ramp of Rubin, with significant revenue visibility into CY26. Despite geopolitical challenges impacting China sales, NVIDIA is expanding its ecosystem through strategic partnerships and investments, while managing supply chain resilience and preparing for continued growth.

    Highlights

    5

    • Total revenue reached $57 billion, marking a 62% year-over-year increase and a record sequential growth of $10 billion or 22%.

    • Data Center revenue hit a record $51.2 billion, up 66% year-over-year, with compute growing 56% and networking more than doubling.

    • NVIDIA has visibility to $0.5 trillion in Blackwell and Rubin revenue from the start of this year through the end of calendar year 2026.

    • Non-GAAP gross margins were 73.6%, exceeding the company's outlook.

    • Gaming revenue increased 30% year-on-year to $4.3 billion, driven by strong demand and Blackwell momentum.

    Concerns

    4

    • H20 sales were approximately $50 million in Q3, as sizable purchase orders did not materialize due to geopolitical issues and increased competition in China.

    • Inventory grew 32% quarter-over-quarter, and supply commitments increased 63% sequentially, indicating significant build-up.

    • The non-GAAP effective tax rate for Q3 was 17%, higher than the guidance of 16.5% due to strong U.S. revenue.

    • Input costs are on the rise for fiscal year 2027, posing a challenge to maintaining gross margins.

    Guidance & targets

    8
    CategoryTargetConfidence
    Total revenue
    $65 billion, plus or minus 2%
    high materiality
    High
    GAAP gross margins
    74.8%, plus or minus 50 basis points
    medium materiality
    High
    Non-GAAP gross margins
    75%, plus or minus 50 basis points
    high materiality
    High
    Gross margins
    Mid-70s
    high materiality
    Medium
    GAAP operating expenses
    Approximately $6.7 billion
    medium materiality
    High
    Non-GAAP operating expenses
    Approximately $5 billion
    medium materiality
    High
    GAAP and Non-GAAP other income and expenses
    Approximately $500 million income
    low materiality
    High
    GAAP and Non-GAAP tax rates
    17%, plus or minus 1%
    low materiality
    High

    Segment performance

    4
    SegmentRevenueYoYQoQMargin
    Data Center
    Record Q3 revenue, driven primarily by the GB300 ramp for compute and robust double-digit growth across NVLink, Spectrum-X Ethernet, and Quantum-X InfiniBand for networking.
    Compute growth: 56% YoYNetworking growth: more than doubled
    		$51.2 billion66%——
    Gaming
    Driven by strong demand as Blackwell momentum continued. End market sell-through remains robust, and channel inventories are at normal levels heading into the holiday season.
    		$4.3 billion30%——
    Professional Visualization
    Another record quarter, with growth driven by DGX Spark, the world's smallest AI supercomputer built on Grace Blackwell.
    		$760 million56%——
    Automotive
    Primarily driven by self-driving solutions, including partnership with Uber for Level 4 ready autonomous fleet.
    		$592 million32%——

    Operational metrics

    22
    Non-GAAP gross margins
    73.6%exceeded outlook
    Q3 FY26

    Increased sequentially due to data center mix, improved cycle time, and cost structure.

    GAAP gross margins
    73.4%
    Q3 FY26

    Increased sequentially due to data center mix, improved cycle time, and cost structure.

    Non-GAAP operating expenses
    11%up sequentially
    Q3 FY26

    Growth driven by infrastructure compute, higher compensation and benefits, and engineering development costs.

    GAAP operating expenses
    8%up sequentially
    Q3 FY26

    Growth driven by infrastructure compute, higher compensation and benefits, and engineering development costs.

    Non-GAAP effective tax rate
    17%higher than 16.5% guidance
    Q3 FY26

    Higher due to strong U.S. revenue.

    Inventory
    32%grew quarter-over-quarter
    Q3 FY26

    Preparing for significant growth ahead.

    Supply commitments
    63%increased sequentially
    Q3 FY26

    Preparing for significant growth ahead.

    Blackwell and Rubin revenue visibility
    $0.5 trillion
    CY26

    Visibility for Blackwell and Rubin platforms.

    AI infrastructure build estimate
    $3 trillion to $4 trillion
    annual

    Estimated annual AI infrastructure build.

    Hyperscaler aggregate CapEx
    $600 billionmore than $200 billion higher relative to start of year
    2026

    Aggregate CapEx for top CSPs and hyperscalers.

    OpenAI weekly user base
    800 million
    recent

    OpenAI's reported weekly user base growth.

    OpenAI enterprise customers
    1 million
    recent

    OpenAI's reported enterprise customer growth.

    Anthropic annualized run rate revenue
    $7 billionup from $1 billion at the start of the year
    annualized

    Anthropic's reported annualized run rate revenue.

    AI factory and infrastructure projects
    5 million
    aggregate

    Aggregate GPUs for announced AI factory and infrastructure projects.

    H20 sales
    $50 million
    Q3 FY26

    Sizable purchase orders never materialized due to geopolitical issues and competitive market in China.

    Blackwell Ultra MLPerf training performance
    5xthan Hopper
    recent

    Time to train in MLPerf training results.

    Blackwell DeepSeek-R1 performance
    10xvs H200
    recent

    Also 10x lower cost per token versus H200.

    OpenAI AI data centers
    10 gigawatts
    target

    Target capacity for AI data centers in strategic partnership.

    Anthropic compute commitment
    1 gigawatt
    initial

    Initial compute capacity commitment with Grace Blackwell and Vera Rubin Systems.

    Unilever content creation acceleration
    2x
    recent

    Using AI and digital twins, also cut costs by 50%.

    Salesforce engineering productivity increase
    30%
    recent

    After adopting Cursor for co-development.

    KSA GPU agreement
    400,000 to 600,000
    over 3 years

    Agreement with KSA for additional GPUs.

    Industry KPIs

    5
    MetricValueDetails
    Ai data center revenue$51.2 billionUSD
    Fab capacity utilizationFully utilized
    Inventory channel inventory32%%
    Node platform ramp scheduleRubin platform
    End market segment revenue mixData Center: $51.2 billion; Gaming: $4.3 billion; Professional Visualization: $760 million; Automotive: $592 millionUSD

    Product announcements

    6
    ProductTypeDetails
    Spectrum-XGSlaunch
    NVLink Fusion with Fujitsuexpansion
    NVLink Fusion with Intelexpansion
    Arm NVLink IP Integrationexpansion
    Blackwell wafer production on U.S. soilmilestone
    NVIDIA Hyperion L4 robotaxi reference architecturelaunch

    Deals & partnerships

    10
    AWS and HUMAINExpanded partnership for AI accelerator deployment

    Expanded their partnership, including the deployment of up to 150,000 AI accelerators, including our GB300.

    xAI and HUMAINJoint development of GPU data centers

    Announced a partnership in which the two will jointly develop a network of world-class GPU data centers anchored by the flagship 500-megawatt facility.

    OpenAIStrategic partnership for AI data center build-out and investment opportunity

    Working on a strategic partnership focused on helping them build and deploy at least 10 gigawatts of AI data centers. Also have the opportunity to invest in the company.

    AnthropicDeep technology partnership for model optimization and compute commitment

    Establishing a deep technology partnership to support Anthropic's fast growth, optimize Anthropic models for CUDA, and optimize future NVIDIA architectures for Anthropic workloads.

    FujitsuStrategic collaboration for NVLink Fusion integration

    Announced a strategic collaboration in October to integrate Fujitsu's CPUs and NVIDIA GPUs via NVLink Fusion, connecting large ecosystems.

    IntelCollaboration for custom data center and PC products using NVLink

    Announced a collaboration to develop multiple generations of custom data center and PC products, connecting NVIDIA and Intel's ecosystems using NVLink.

    ArmIntegration of NVLink IP for CPU SoCs

    Arm announced that it will be integrating NVLink IP for customers to build CPU SoCs that connect with NVIDIA.

    PalantirSupercharging oncology platform with NVIDIA AI

    Partnering to supercharge the incredibly popular oncology platform with NVIDIA CUDA-X libraries and AI models.

    UberScaling Level 4 ready autonomous fleet

    Partnering to scale the world's largest Level 4 ready autonomous fleet built on the new NVIDIA Hyperion L4 robotaxi reference architecture.

    TSMCBlackwell wafer production in U.S.

    Partnership to produce the first Blackwell wafer on U.S. soil.

    Capital programs

    6
    xAI's Colossus 2 data centerannounced

    Benefit: Gigawatt scale

    Described as the world's first gigawatt scale data center.

    Lilly's AI Factoryannounced

    An AI Factory for drug discovery, described as the pharmaceutical industry's most powerful data center.

    AWS and HUMAIN AI accelerator deploymentannounced

    Benefit: Up to 150,000 AI accelerators (including GB300)

    Expanded partnership between AWS and HUMAIN for deployment of AI accelerators.

    xAI and HUMAIN GPU data center networkannounced

    Benefit: Flagship 500-megawatt facility

    Joint development of a network of world-class GPU data centers, anchored by a flagship 500-megawatt facility.

    OpenAI AI data centersannounced

    Benefit: At least 10 gigawatts

    Strategic partnership focused on helping OpenAI build and deploy AI data centers.

    Anthropic compute capacityannounced

    Benefit: Up to 1 gigawatt

    Initial compute commitment with Grace Blackwell and Vera Rubin Systems as part of a deep technology partnership.

    Risks & headwinds

    3
    China Sales ImpactQ3 FY26

    H20 sales approximately $50 million in Q3 FY26

    Mitigation: Committed to continued engagement with the U.S. and China governments and advocating for America's ability to compete globally.

    Rising Input CostsFiscal Year 2027

    Input costs are on the rise

    Mitigation: Working to hold gross margins in the mid-70s through cost improvements, cycle time, and mix.

    Supply Chain ConstraintsOngoing

    Power, financing, memory, and foundry are all issues and constraints

    Mitigation: Extensive planning with supply chain partners, established partnerships in land and power, and focus on delivering best value per TCO and performance per watt.

    What to watch in Q4 FY26

    5

    Blackwell/Rubin Revenue Visibility

    End of CY26
    Current$0.5 trillion
    TargetIncreased beyond $0.5 trillion

    Why it matters

    Indicates continued strong demand and NVIDIA's ability to capture the growing AI infrastructure market.

    We currently have visibility to $0.5 trillion in Blackwell and Rubin revenue from the start of this year through the end of calender year 2026.

    Q&A highlights

    7

    Confirming the $0.5 trillion Blackwell/Rubin revenue forecast through CY26 and asking about potential upside given new deals.

    Colette Kress confirmed the $0.5 trillion forecast is on track and noted that new deals like KSA (400k-600k GPUs over 3 years) and Anthropic are net new opportunities, suggesting potential upside beyond the initial forecast.

    “Yes, that's correct. We are working into our $500 billion forecast. And we are on track for that as we have finished some of the quarters, and now we have several quarters now in front of us to take us through the end of calendar year '26. The number will grow.”

    asked by Joseph Moore · answered by Colette Kress

    2 min read6 chapters

    Detailed Narrative

    01

    AI Infrastructure Demand and Market Opportunity

    NVIDIA projects $0.5 trillion in Blackwell and Rubin revenue visibility through calendar year 2026, driven by exponential growth in accelerated computing, powerful AI models, and agentic applications. The company estimates the annual AI infrastructure build will reach $3 trillion to $4 trillion by the end of the decade. Demand continues to exceed expectations, with clouds sold out and NVIDIA's GPU installed base fully utilized across new and previous generations.

    02

    Platform Shifts and NVIDIA's Unique Position

    Jensen Huang highlighted three simultaneous platform shifts: the transition from CPU general-purpose computing to GPU accelerated computing, generative AI replacing classical machine learning in hyperscale infrastructure, and the rise of agentic AI systems. NVIDIA's singular architecture is uniquely positioned to address all three transformations, offering a unified platform for diverse computing needs from cloud to edge and across all phases of AI.

    03

    Blackwell and Rubin Platform Momentum

    Blackwell gained significant momentum in Q3, with the GB300 contributing approximately two-thirds of total Blackwell revenue and seeing seamless production shipments to major cloud service providers. The Rubin platform is on track to ramp in the second half⚖️ of 2026, promising an 'x-factor' performance improvement and leveraging NVIDIA's mastered rack architecture for a fast ramp. The company's annual 'X factor' performance leap increases performance per dollar and drives down computing costs.

    04

    Networking Business Dominance

    NVIDIA's networking business generated $8.2 billion in revenue, up 162% year-over-year, with contributions from NVLink, InfiniBand, and Spectrum-X Ethernet. The company is winning in data center networking, as the majority of AI deployments now include its switches. Hyperscalers like Meta, Microsoft, Oracle, and xAI are building gigawatt AI factories with Spectrum-X Ethernet, highlighting the platform's flexibility and openness.

    05

    Strategic Ecosystem Expansion and Investments

    NVIDIA is actively expanding its CUDA AI ecosystem through strategic investments and deep technical partnerships with frontier AI companies such as OpenAI and Anthropic. These collaborations aim to optimize AI models for CUDA and future NVIDIA architectures, ensuring optimal performance and efficiency. The company's investment thesis focuses on expanding CUDA's reach and securing a share in the success of these consequential, once-in-a-generation companies.

    06

    Geopolitical Impact and Supply Chain Resilience

    H20 sales in Q3 were approximately $50 million, impacted by geopolitical issues and increased competition in China, leading to sizable purchase orders not materializing. NVIDIA remains committed to engaging with governments to advocate for its ability to compete globally. The company is also focused on building resiliency and redundancy in its global supply chain, partnering with TSMC for U.S. wafer production and working with other partners to grow its U.S. presence.

    AI-generated summary of the company’s earnings call. Not investment advice.