Amazon AMZN
Amazon is Anthropic's largest strategic investor and primary cloud partner through AWS. It has committed billions to Ant...
Cloud and AI infrastructure.
Where models actually run: hyperscale cloud, GPU clouds, and the server builders wiring up the racks for AI workloads.
Alphabet (Google) GOOGL
Alphabet has committed up to $40B to Anthropic and supplies Google Cloud capacity plus custom TPU accelerators for train...
Microsoft MSFT
Microsoft is OpenAI's largest backer and primary infrastructure partner through Azure, making it Anthropic's main compet...
Oracle ORCL
Oracle is a Stargate operations partner hosting massive AI training clusters, with multi-year cloud commitments tied to ...
CoreWeave CRWV
CoreWeave is a GPU-focused cloud built for AI training and inference, with direct capacity deals to Anthropic, Microsoft...
Super Micro Computer SMCI
Super Micro Computer builds AI server racks and liquid-cooled systems that hyperscalers and GPU clouds deploy at scale. ...
Dell Technologies DELL
Dell Technologies builds enterprise and AI server infrastructure used by every company and cloud provider building out A...
Where AI actually runs.
Very few companies build their own AI data centers. Most rent capacity from a cloud: the three hyperscalers (Amazon Web Services, Microsoft Azure and Google Cloud), Oracle, and specialist GPU clouds such as CoreWeave. Dell and Super Micro sit one step earlier in the chain, assembling the servers and racks that fill those data centers.
The hyperscalers are also the AI labs' biggest backers. Amazon and Alphabet own large stakes in Anthropic, Microsoft owns about 27% of OpenAI, and Amazon invested in OpenAI too in 2026. Each pairs its stake with a cloud contract, which makes these stocks the closest listed exposure to the private labs.
The newer players.
AI demand created room for companies that do nothing but rent out GPUs. CoreWeave went public in March 2025 and has multi-billion-dollar contracts with OpenAI. Oracle, long a database company, signed an agreement for about $300 billion of capacity for OpenAI over five years from 2027, which turned it into one of the most AI-dependent large stocks. Private rivals such as Nscale, Lambda and Fluidstack have signed tens of billions in contracts with Anthropic, and some are preparing their own listings.
What moves these stocks.
Cloud revenue growth and backlog, the contracted revenue not yet recognized, are the clearest signals of AI demand. Capital spending is the other side: Amazon, Microsoft, Alphabet and Meta together planned roughly $725 billion for 2026, and investors watch whether revenue grows fast enough to justify it. For server builders, the mix of AI servers and their thin margins matter more than headline growth.
Risks to weigh.
Spending could run ahead of demand, leaving clouds with expensive capacity and rising depreciation. Several of the newer players rely on one or two AI labs for much of their future revenue and on debt to build. Power and grid connections increasingly limit where and how fast data centers can be built.