Stock Markets March 2, 2026

Big Tech and Startups Pour Tens to Hundreds of Billions into AI Infrastructure

A string of multibillion-dollar commitments - from chips to data centers and licensing - underscores booming demand for compute and cloud capacity

By Leila Farooq INTC
Big Tech and Startups Pour Tens to Hundreds of Billions into AI Infrastructure
INTC

Nvidia, hyperscalers, chipmakers and investors have announced a wide array of multi-billion-dollar arrangements to expand artificial intelligence compute, storage and model capabilities. The deals span direct chip and cloud purchases, equity investments, long-term contracts for data center capacity, and licensing agreements for content and model use. These commitments highlight how companies across the technology stack are directing large amounts of capital and contractual commitments into AI infrastructure and services.

Key Points

  • Major players across semiconductors, cloud providers, data centers and AI firms have signed multibillion-dollar agreements to secure compute, chips and capacity.
  • Deals include long-term cloud purchase commitments, direct chip supply contracts, equity investments and licensing arrangements for content and technology.
  • Sectors most impacted include semiconductors, cloud infrastructure, data-center operations and media/content licensing.

Nvidia will invest $2 billion each in photonic product makers Lumentum and Coherent to support those companies’ research, development and U.S. manufacturing operations, the companies said.

Below is a compiled list of significant recent AI, cloud and semiconductor agreements that together represent many billions of dollars of capital and long-term commitments across chip suppliers, cloud providers, data-center operators and AI firms.


OpenAI-related arrangements

  • Amazon and OpenAI - Amazon is considering an investment of around $10 billion in OpenAI, although those discussions remain described as very fluid.

  • Disney and OpenAI - Walt Disney will invest $1 billion in OpenAI and will license characters from Star Wars, Pixar and Marvel for use in OpenAI’s Sora AI video generator. Under a three-year licensing agreement, Sora and ChatGPT Images will begin producing videos featuring licensed Disney characters such as Mickey Mouse, Cinderella and Mufasa early next year. The licensing excludes talent likenesses and voices.

  • Broadcom and OpenAI - OpenAI has partnered with Broadcom to develop its first in-house artificial intelligence processors, reflecting the startup’s expanding relationships to secure computing power for its services.

  • AMD and OpenAI - AMD agreed to supply AI chips to OpenAI under a multi-year arrangement that also gives OpenAI the option to acquire up to roughly 10% of the chipmaker.

  • Nvidia and OpenAI - Nvidia is set to invest up to $100 billion in OpenAI and to supply the startup with data-center chips, a deal that gives Nvidia an ownership stake in OpenAI. OpenAI is already a major Nvidia customer.

  • Oracle and OpenAI - Oracle is reported to have signed a cloud arrangement with OpenAI under which OpenAI is expected to purchase $300 billion in computing power from Oracle over approximately five years.

  • CoreWeave and OpenAI - CoreWeave signed a five-year contract worth $11.9 billion with OpenAI in March, prior to the Nvidia-backed startup’s initial public offering.

  • Stargate datacenter project - Stargate is a joint venture including SoftBank, OpenAI and Oracle announced as a project to build data centers. That project was described as an effort in which the companies would invest up to $500 billion to support AI infrastructure, a figure announced publicly in January.


Meta-related deals

  • Meta and AMD - AMD will sell up to $60 billion worth of AI chips to Meta, with the social media company able to purchase as much as 10% of AMD. AMD will supply six gigawatts’ worth of chips to Meta, beginning with one gigawatt of the forthcoming MI450 flagship hardware in the second half of this year. Meta also plans to buy central processors, including a customized variant for its needs.

  • Meta and Manus - Meta is acquiring Chinese startup Manus to accelerate integration of agentic AI tools into platforms such as Facebook, Instagram and WhatsApp. The transaction was reported to value Manus between $2 billion and $3 billion, though financial terms were not disclosed officially.

  • Meta and CoreWeave - CoreWeave has signed a $14 billion contract to supply compute capacity to Meta.

  • Meta and Oracle - Oracle is in talks with Meta for a multiyear cloud computing agreement worth about $20 billion to secure access to computing power.

  • Meta and Google - Google and Meta struck a six-year cloud computing deal reportedly worth more than $10 billion.

  • Meta and Scale AI - Meta acquired a 49% stake in Scale AI for about $14.3 billion and brought in Scale’s CEO, Alexandr Wang, to play a prominent role in its AI strategy.


Nvidia-related deals beyond the Lumentum and Coherent investments

  • Nvidia and Groq - Nvidia has agreed to license chip technology from startup Groq and to hire Groq’s CEO Jonathan Ross and other engineers. Reports indicate Nvidia agreed to acquire Groq’s assets for $20 billion.

  • Microsoft, Nvidia and Anthropic - Microsoft will invest up to $5 billion and Nvidia up to $10 billion in Anthropic. Under that agreement, Anthropic pledged $30 billion to run its workloads on Microsoft’s cloud and committed up to 1 gigawatt of compute powered by Nvidia's advanced Grace Blackwell and Vera Rubin hardware. Anthropic will also collaborate with Nvidia to improve chips and AI models.

  • Nvidia-backed investor group and Aligned Data Centers - An investor consortium that includes BlackRock, Microsoft and Nvidia is acquiring U.S.-based Aligned Data Centers, which operates nearly 80 facilities, in a transaction worth $40 billion.

  • Nvidia and Intel - Nvidia will invest $5 billion in Intel, which will give Nvidia roughly a 4% stake in Intel after the issuance of new shares.

  • CoreWeave and Nvidia - CoreWeave signed an initial order with Nvidia worth $6.3 billion. That agreement guarantees that Nvidia will purchase any cloud capacity that CoreWeave does not sell to customers.


Google-related infrastructure and licensing moves

  • Google and Texas - Google will invest $40 billion in three new data centers in Texas through 2027. One site will be in Armstrong County in the Texas Panhandle, and two will be in Haskell County near Abilene. Google is also continuing investment at its Midlothian campus and the Dallas cloud region, which are part of its global network of 42 cloud regions.

  • Google and Windsurf - Google hired several staff from AI code generation startup Windsurf and agreed to pay $2.4 billion in license fees to use some of Windsurf’s technology under non-exclusive terms.


Other notable arrangements

  • Nebius Group and Microsoft - Nebius Group will provide GPU infrastructure capacity to Microsoft in a deal valued at $17.4 billion over five years.

  • Intel and SoftBank Group - SoftBank Group is providing Intel with a $2 billion capital injection, making SoftBank one of Intel’s top-10 shareholders.

  • Tesla and Samsung - Tesla signed a $16.5 billion deal to source chips from Samsung Electronics; Tesla’s CEO stated that Samsung’s new chip factory in Texas would make Tesla’s next-generation AI6 chip.

  • Amazon and Anthropic - Amazon invested $4 billion in Anthropic, doubling its stake in the OpenAI competitor known for the Claude chatbot.

  • SoftBank and DigitalBridge - SoftBank Group will acquire digital infrastructure investor DigitalBridge Group in a transaction valued at $4 billion.


Context and implications

The compiled items above include direct chip supply agreements, licensing deals for model and content use, equity investments in AI startups, multiyear cloud purchase commitments and large-scale data center investments. Together they represent substantial capital flows across semiconductor manufacturers, cloud providers, data-center operators and AI model developers. Several agreements include options to purchase equity stakes in suppliers, multi-year purchase commitments measured in tens or hundreds of billions of dollars, or long-duration contracts for compute capacity. Other deals include integration of corporate content into generative AI products under defined licensing terms.

These arrangements span a range of business models - from hardware supply and customization, to cloud procurement and long-term capacity contracts, to licensing of intellectual property and character rights for AI-driven content generation. They involve established technology companies, startups, and investor groups, often overlapping across multiple agreements.


Summary takeaway

The recent wave of multibillion-dollar transactions highlights the accelerating capital commitments being directed toward AI compute, cloud and data-center capacity, along with licensing and strategic partnerships to support model development and deployment.

Risks

  • Significant concentration of capital commitments creates exposure to execution risk in data-center buildouts and long-term cloud procurement - impacting data-center operators and cloud providers.
  • Large multi-year purchase commitments and equity stakes heighten counterparty and concentration risk for chipmakers and buyers if demand dynamics shift.
  • Licensing deals for character and model usage introduce commercial and operational constraints tied to content, potentially affecting the media and streaming sectors.

More from Stock Markets

Casablanca market slides as Moroccan All Shares falls 4.21% to six-month low Mar 2, 2026 Barclays Reprices Alternatives: Downgrade for Blue Owl, Upgrade for StepStone Mar 2, 2026 U.S. Faces Hurdles in Shifting Congo Minerals Supply Chain Away from China Mar 2, 2026 Athens market slips as banks, telecoms and household names lead losses Mar 2, 2026 Tel Aviv Stocks Rally as TA 35 Closes at Record High, Led by Energy and Financial Names Mar 2, 2026