Invesight

NVIDIA

NVDAFY2026 · Based on 10-K

GPU design company dominating over 80% of the AI semiconductor market. Designs chips for data centers, autonomous driving, and gaming.

Market Cap

Revenue Growth

+66.0%

Operating Margin

60.4%

Revenue Structure

Compute & Networking
89.6%+67% YoY
Graphics
10.4%+57% YoY
Revenue Structure
SegmentRevenuePercentYoY Growth
Compute & Networking$193.5B89.6%67%
Graphics$22.5B10.4%57%

Customer Concentration

EstimatedNot explicitly stated in 10-K filing

Customer Concentration
CustomerRevenue %Estimated
Microsoft Corporation22%Yes
Amazon.com, Inc.14%Yes

Value Chain Related Stocks

#1
Taiwan Semiconductor Manufacturing Company LimitedTSMSupplier

Taiwan's leading foundry, manufacturing NVIDIA's GPUs, CPUs, and DPUs at advanced nodes (e.g., 4nm, 3nm) under fabless model

Source: Item 1, Business Description

#2
Samsung Electronics Co., Ltd.SSNLFSupplier

Alternative foundry partner for GPUs and supplier of DRAM and high-bandwidth memory

Source: Item 1, Business Description

#3
SK Hynix Inc.SKHYYSupplier

Primary supplier of HBM3/HBM3E high-bandwidth memory — critical for Blackwell and Rubin platforms

Source: Item 1, Business Description

#4
Micron Technology, Inc.MUSupplier

Supplier of HBM and GDDR6/GDDR6X memory — part of NVIDIA’s memory diversification strategy for data center and gaming GPUs

Source: Item 1, Business Description

#5
Hon Hai Precision Industry Co., Ltd.2317.TWSupplier

Foxconn — key contract manufacturer responsible for server system assembly and testing

Source: Item 1, Business Description

#6
Wistron Corporation3231.TWSupplier

Wistron — contract manufacturer for assembly and testing of servers and AI infrastructure equipment

Source: Item 1, Business Description

#7
FabrinetFNSupplier

Optical and precision assembly-focused contract manufacturer — produces network switch chips and optical communication modules

Source: Item 1, Business Description

#8
Microsoft CorporationMSFTCustomer

Cloud service provider (CSP) and AI supercomputer partner — deploying Blackwell-based DGX Cloud and GB200 in Azure AI infrastructure

majority of Compute & Networking revenue

Source: Item 7, MD&A; Item 1, Business Description

#9
Amazon.com, Inc.AMZNCustomer

AWS cloud customer and AI infrastructure partner — collaboration on Amazon Bedrock, Trainium/Inferentia integration, and NVIDIA AI Enterprise deployment

significant portion of Compute & Networking revenue

Source: Item 7, MD&A; Item 1, Business Description

#10
Google LLCGOOGLCustomer

Cloud and AI model development customer — uses Blackwell-based infrastructure for Gemini model training and inference

material

Source: Item 1, Business Description

#11
Meta Platforms, Inc.METACustomer

AI infrastructure customer — large-scale Blackwell deployment for Llama model training and inference

material

Source: Item 1, Business Description

#12
Groq, Inc.Partner

LLM inference optimization startup — entered multi-year non-exclusive license agreement with NVIDIA, requiring substantial engineering effort

Source: Item 1A, Risk Factors

#13
OpenAIPartner

AI model development company — investment and partnership agreement with NVIDIA under finalization as of February 2026

Source: Item 1A, Risk Factors; Item 7, MD&A

#14
Mellanox Technologies, Ltd.Partner

InfiniBand and high-speed networking company — acquired by NVIDIA in 2020 to strengthen DPUs and networking capabilities

Source: Item 1, Business Description

Competitors

Advanced Micro Devices, Inc.AMDHuawei Technologies Co. Ltd.Intel CorporationINTCAlibaba Group Holding LimitedBABAAlphabet Inc.GOOGLAmazon, Inc.AMZNBaidu, Inc.BIDUMicrosoft CorporationMSFTBroadcom Inc.AVGOQualcomm IncorporatedQCOMTesla, Inc.TSLAArista Networks, Inc.ANETCisco Systems, Inc.CSCO

Risk Factors

Escalating U.S. Government AI Semiconductor Export Controls

High

Since 2022: A100/H100 bans, 2025 H20/H200 licensing requirements, de facto exclusion from China’s data center market. $4.5B H20 inventory charge and unmonetized licenses. Inability to deliver a dual-approved product risks permanent competitive erosion in China.

Impact: Revenue loss, inventory impairments, expansion of competitors’ (Huawei, Alibaba) global ecosystems, Chinese government’s push for domestic alternatives, and potential long-term growth rate downgrade.

Supply Chain Concentration and Manufacturing Lead Time Uncertainty

High

Asia-heavy foundry and assembly footprint, >12-month lead times, reliance on non-cancellable prepayments. Demand forecasting errors risk shortages or excess inventory — FY2025 Blackwell low yields pressured gross margin.

Impact: Lost revenue opportunities, reputational damage, inventory write-downs and price reductions, higher operating costs, delays in U.S./Latin America supply chain diversification.

Sophisticated Cybersecurity and Operational Disruption Risks

High

State-sponsored hacking, AI-powered attacks, third-party supply chain vulnerabilities (e.g., GFN breach), geopolitical tensions in Israel, Taiwan, and Korea. Insider threats and open-source data quality issues also present.

Impact: Legal liability, eroded customer trust, IP theft, product launch delays, rising security enhancement costs, and expanded financial losses due to insurance coverage limitations.

Expanding Global Regulation and AI Accountability Risks

Medium

EU AI Act, U.S. state AI laws, Chinese antitrust investigations, GDPR/NIS2 compliance mandates. Enhanced AI model transparency, auditability, and accountability increase R&D and operational costs.

Impact: Higher regulatory compliance costs, delayed product launches, service restrictions in certain markets, and need to adapt AI software revenue models.

Customer and Partner Concentration Risk

Medium

36% of FY2026 revenue concentrated in two direct customers (MSFT, AMZN). Most sales are purchase-order-based, permitting customers to cancel or delay orders without notice.

Impact: Risk of sharp revenue decline if relationship with a key customer deteriorates, weakened negotiation leverage, and increased accounts receivable collection risk.

Accelerating Technology Transition Cycles and Shortening Product Lifecycles

Medium

Annual architecture transitions in data center (Blackwell → Blackwell Ultra → Rubin), transition-related inventory impairments, channel inventory rebalancing, and infrastructure readiness delays amplify demand volatility.

Impact: Reduced revenue stability, margin pressure, delayed customer adoption, and market share vulnerability due to differential competitor response speed.

Geopolitical Risk and Climate Change Impact

Medium

Israel conflict (impacting ~6,000 employees), Taiwan/Korea supply chain fragility, California wildfire-related power shutoffs, and tightening climate-related regulations.

Impact: R&D delays, logistics disruptions, HQ and data center outages, and increased CAPEX due to carbon taxes and energy efficiency mandates.

Growth Drivers

Surging Data Center AI Infrastructure Demand Driven by Blackwell and Rubin Platforms

FY2026 Data Center revenue up 68%, with Blackwell accounting for majority of segment revenue. Blackwell Ultra (GB300) volume shipments commenced; Rubin platform scheduled for H2 FY2027 production — optimized for agentic and physical AI.

Outlook: Continued >40% CAGR over next 3 years expected, fueled by expanding use cases: LLM training/inference, AI agents, robotics/automation, and omics.

Expansion of Software and Platform-Based Revenue Models

Growth in NVIDIA AI Enterprise licensing, NIM (NVIDIA Inference Microservices), NeMo, and AI Blueprints — SaaS-style AI software offerings. Ecosystem of 7.5M+ CUDA developers established.

Outlook: Diversification from hardware-centric to software- and service-centric revenue mix, with increasing share of high-margin (>85%) licensing and subscription revenue.

Cross-Industry AI Adoption Expansion in Automotive and Healthcare

DRIVE Hyperion-based autonomous driving solutions drove 39% automotive revenue growth. Clara healthcare platform and Omniverse physical AI simulation expand into vertical domains.

Outlook: Long-term contracts with automotive OEMs and Tier-1 suppliers, plus rising demand for digital twins and real-time simulation, support >25% CAGR in professional verticals over next 5 years.

Competitive Position: Market share Data Center AI Accelerators: >80% (est.), TOP500 Supercomputers: >78%, Green500: 9/10, Full-stack moat: proprietary CUDA software ecosystem + world-leading GPU/CPU/DPU silicon + high-bandwidth interconnects (NVLink, InfiniBand) + domain-specific AI stacks (DRIVE, Clara, Omniverse) + 7.5M+ developer community — creates immense switching costs and ecosystem lock-in.

Investment Insights

FY2026 · Based on 10-K

NVIDIA is no longer just a GPU vendor—it is the foundational infrastructure provider for the AI era, delivering an end-to-end 'operating system' for accelerated computing. While explosive Blackwell demand and the Rubin platform’s technical leadership fuel near-term growth, the de facto exclusion from China’s data center market—driven by U.S. export controls—poses a structural, long-term threat to growth rates and global market share. The $4.5B H20 inventory charge is not merely an accounting item but a symptom of an irreconcilable policy gap: no product can simultaneously satisfy U.S. export thresholds and Chinese performance expectations. Conversely, NVIDIA’s 7.5M+ CUDA developers, dominance in 9 of the top 10 Green500 systems, and deep vertical integrations (DRIVE, Clara, Omniverse) create an ecosystem moat that rivals cannot easily replicate. For investors, the pivotal question is no longer 'How fast will AI infrastructure grow?' but rather 'How long will the U.S.-China tech cold war persist?'—and this geopolitical timeline matters more for long-term enterprise value than quarterly margins.

Related Value Chains

Explore other value chains in similar sectors