NVIDIA invented the GPU and now dominates AI computing with its H100 and B200 chips. Every major AI model - ChatGPT, Gemini, Claude - was trained on NVIDIA hardware. Data center revenue grew 400%+ in one year as every tech company raced to buy GPUs. Its CUDA software platform created a moat - AI researchers learn CUDA, not alternatives. NVIDIA is the picks-and-shovels of the AI gold rush.
Suppliers 14
Companies that sell to NVIDIA
- TSMCSemiconductors · $900BH100/Blackwell GPU fabricationsince 1998
- EatonIndustrial & Defense · $95BEaton powers NVIDIA AI supercomputer facilities
- SynopsysSemiconductors · $85BChip design EDA tools
- Cadence DesignSemiconductors · $80BSignoff and verification
- AmphenolIndustrial & Defense · $80BGPU server connectors
- TE ConnectivityIndustrial & Defense · $46BServer connectors
- Vertiv HoldingsTechnology · $35BLiquid cooling for AI racks
- AnsysTechnology · $28BNVIDIA uses Ansys to simulate GPU chip design
- Monolithic PowerSemiconductors · $22BPower management for GPUs
- TeradyneSemiconductors · $20BTeradyne tests every NVIDIA GPU before packaging
- JFrogTechnology · $4BJFrog manages NVIDIA AI model and software artifacts
- SiTimeSemiconductors · $3BSiTime precision clocks in NVIDIA data center GPUs
- MYR GroupIndustrial & Defense · $3BMYR builds power infrastructure for NVIDIA AI facilities
- FormFactorSemiconductors · $2BFormFactor tests NVIDIA GPU wafers pre-packaging
Customers 31
Companies NVIDIA sells to
- MicrosoftTechnology · $2.8TH100/H200 GPUs power Azure AI
- AlphabetTechnology · $2.1TGPUs for Google Cloud TPU supplement
- AmazonConsumer & Retail · $2.0TGPUs for AWS Trainium/Inferentia
- MetaTechnology · $1.4T350,000+ H100s for AI infra
- TeslaAutomotive · $800BDojo AI training chips
- WalmartConsumer & Retail · $700BAI GPUs for Walmart's supply chain
- JPMorgan ChaseFinance · $600BGPU clusters for JPMorgan's AI
- VisaFinance · $560BGPUs for Visa fraud detection
- MastercardFinance · $470BGPUs for Mastercard Decision Intelligence
- OracleTechnology · $450BOCI AI cloud GPU clusters
- TencentTechnology · $400BNVIDIA GPUs for Tencent Cloud AI workloads
- ByteDanceMedia & Telecom · $330BGPUs for ByteDance's AI models
- ToyotaAutomotive · $280BToyota autonomous driving chips
- AlibabaTechnology · $220BNVIDIA GPUs in Alibaba Cloud AI clusters
- xAITechnology · $200BGPUs for the Colossus supercomputer in Memphissince 2024
- NintendoConsumer & Retail · $100BCustom NVIDIA chips powering Switch and Switch 2since 2017
- Dell TechnologiesTechnology · $90BGPUs for Dell AI Factory servers
- OpenAITechnology · $80BGPUs powering GPT model training
- VolkswagenAutomotive · $70BAutonomous driving AI
- Mercedes-BenzAutomotive · $60BNVIDIA DRIVE chips and AI software for Mercedes automated driving
- CloudflareTechnology · $45BGPUs for Cloudflare Workers AI edge inference
- NetEaseTechnology · $40BNVIDIA GPUs power NetEase game servers
- Hyundai MotorAutomotive · $40BNVIDIA Blackwell GPUs for Hyundai's AI factory (autonomy and robotics)
- BaiduTechnology · $35BNVIDIA GPUs for Baidu ERNIE AI training
- Sea LimitedTechnology · $35BNVIDIA GPUs for Sea's Garena game servers
- RobloxTechnology · $22BNVIDIA GPUs power Roblox 3D game rendering
- HPETechnology · $22BGPUs for HPE ProLiant and Cray servers
- Li AutoAutomotive · $18BAutonomous driving platform
- NIO IncAutomotive · $12BADAS computing chips
- XPeng IncAutomotive · $10BDrive Xavier chips
- DeepSeekTechnology · privateDeepSeek trained its models on NVIDIA H800 GPUs
Partners 29
Strategic partnerships, joint products and alliances
- ExxonMobilEnergy · $490BGPU simulation for reservoir modelling
- SalesforceTechnology · $290BNVIDIA models in Agentforce
- ToyotaAutomotive · $280BNVIDIA DRIVE autonomous platform
- SAPTechnology · $280BNVIDIA AI in SAP Business AI
- AdobeTechnology · $220BAdobe Firefly models on NVIDIA computing
- CiscoTechnology · $200BCisco Secure AI Factory with NVIDIA
- DisneyMedia & Telecom · $200BNVIDIA Omniverse at ILM
- CaterpillarIndustrial & Defense · $190BPhysical AI for construction and mining machines
- ServiceNowTechnology · $180BServiceNow integrating NVIDIA NIM AI microservices into Now Platform
- IBMTechnology · $175BNVIDIA AI Data Platform in IBM storage and cloud
- VerizonMedia & Telecom · $165BPrivate 5G with NVIDIA AI
- AmgenHealthcare · $160BAI drug discovery platform
- UberTechnology · $140BAI chips for autonomous fleet
- ARM HoldingsSemiconductors · $140BGrace CPU architecture
- Palo Alto NetworksTechnology · $120BNVIDIA AI in Precision AI
- CrowdStrikeTechnology · $75BNVIDIA AI in Falcon
- PalantirTechnology · $60BAI Enterprise joint deployment
- AutodeskTechnology · $52BGPU-accelerated design and simulation
- SnowflakeTechnology · $50BAI data cloud integration
- DatadogTechnology · $40BGPU fleet monitoring
- Super Micro ComputerTechnology · $30BH100/H200 GPU server systems
- AnsysTechnology · $28BGPU-accelerated simulation
- Pure StorageTechnology · $20BPure Storage accelerates AI workloads with NVIDIA
- MongoDBTechnology · $18BGPU-accelerated vector search in MongoDB Atlas
- Mistral AITechnology · $14BBuilding AI computing capacity in Europe with NVIDIA
- Unity SoftwareTechnology · $10BOmniverse simulation platform
- WolfspeedSemiconductors · $4BSTMicroelectronics licenses Wolfspeed SiC technology
- SoundHound AITechnology · $4BAI inference platform
- Recursion PharmaHealthcare · $2BNVIDIA $50M+ investment in Recursion AI drug discovery
Owned by 1
Companies with a stake in NVIDIA
- BlackRockFinance · $130BLarge institutional NVDA position
Stakes owned 5
Companies NVIDIA owns or has invested in
- IntelSemiconductors · $180BNvidia bought ~4% Intel stake for $5B - strategic semiconductor betsince 2025
- OpenAITechnology · $80BNvidia $100B equity+supply deal - 4% stake in OpenAIsince 2025
- AnthropicTechnology · $60BNVIDIA invested in Anthropic (2025)since 2025
- CoreWeaveTechnology · $35BNvidia $6.3B committed to absorb CoreWeave unsold capacity through 2032since 2023
- PerplexityTechnology · $20BNVIDIA is an investor in Perplexitysince 2024
Competitors 3
Companies it competes with
- BroadcomSemiconductors · $800BAI chips: custom ASICs vs GPUs
- AMDSemiconductors · $250BAI GPU rivalry
- QualcommSemiconductors · $185BAI chips
Key relationships in depth
NVIDIA supplies H100, H200, and Blackwell B200 GPUs that power Azure's AI supercomputing clusters. Microsoft is one of NVIDIA's largest hyperscaler customers, ordering tens of thousands of GPUs per quarter to support Azure OpenAI Service, Azure ML, and Copilot infrastructure.
Why it matters: Azure's position as the exclusive cloud for OpenAI creates massive GPU demand that only NVIDIA can fulfill at scale. Every ChatGPT inference request and Copilot suggestion runs on NVIDIA hardware hosted in Azure. Microsoft's $13B investment in OpenAI is only as valuable as the compute it can supply.
Google Cloud uses NVIDIA H100 and A100 GPUs to supplement its proprietary TPU fleet. NVIDIA GPUs handle inference workloads and serve third-party AI developers on GCP who prefer the widely-adopted CUDA programming environment over Google's custom TPU software stack.
Why it matters: NVIDIA's CUDA ecosystem is the dominant AI development framework with 4M+ active developers. Google cannot sell AI cloud services to CUDA-native developers without stocking NVIDIA hardware. TPUs are optimal for Google's internal workloads but lack the broad third-party software support NVIDIA has built over 15 years.
AWS installs NVIDIA GPUs across its data center fleet while simultaneously developing competing Trainium (training) and Inferentia (inference) custom chips. AWS offers NVIDIA instances as a premium tier for AI workloads requiring CUDA compatibility while pushing proprietary chips for price-sensitive workloads.
Why it matters: AWS's $115B annual revenue (2024) depends on being a full-service AI cloud. Customers building AI models expect NVIDIA availability. AWS cannot maintain cloud leadership without offering state-of-the-art NVIDIA hardware, even as it invests heavily in alternatives.
NVIDIA & Meta
Meta deployed more than 350,000 NVIDIA H100 GPUs by mid-2024, making it one of the largest single NVIDIA customers in history. These GPUs train Meta's Llama large language models, power the AI recommendation systems driving Facebook and Instagram engagement, and support Meta's AI assistant across all its platforms.
Why it matters: Meta's ad revenue model, which generated $164B in 2024, depends on AI-powered content ranking and ad targeting. The more powerful the AI, the better the recommendations, the higher the engagement, and the more advertisers pay. Meta's 40% increase in ad prices in 2024 was directly attributable to AI improvements funded by NVIDIA GPU investment.
Relationships are compiled by VexMap from company filings, announcements and reporting, and last reviewed September 2026. Spotted something wrong? Tell us. Not investment advice.