AI rivals like OpenAI, Nvidia, and Oracle are collaborating to build ‘Stargate’—but a Yale expert says it violates 135 years of antitrust law

President Trump announced the Stargate Project, a $500 billion joint venture led by OpenAI, Oracle, SoftBank and joined by Microsoft, Nvidia, Arm and Abu Dhabi‑backed MGX, pledging $100 billion immediately to build an unprecedented array of AI data centers (projects totalling 7–10 gigawatts including a 1.2GW site in Abilene, Texas) to drive US AI infrastructure capacity; OpenAI and partners have outlined roughly $400–$500 billion in infrastructure commitments across multiple states. Yale researcher Madhavi Singh warns the consortium concentrates power at the critical infrastructure layer—where Nvidia dominates GPUs and hyperscalers control cloud services—and argues Stargate risks violating the Clayton and Sherman Acts by entrenching competitors, raising prices, limiting choice and blunting innovation. With Congress and regulators largely silent or supportive and few public legal challenges, the venture could materially reconfigure supply dynamics and pricing power in chips and cloud services, creating systemic competition and policy risks for enterprise AI buyers and the broader tech ecosystem.

Analysis

On Jan. 21, 2025 the White House unveiled “Stargate,” a $500 billion AI infrastructure joint venture led by OpenAI, Oracle and SoftBank with partners Microsoft, Nvidia, Arm and Abu Dhabi‑backed MGX; OpenAI said the new company will deploy $100 billion immediately and build projects totaling roughly 7–10 gigawatts (including a 1.2 GW site in Abilene, Texas) across multiple U.S. states. The announcement frames a behemoth data‑center build‑out and redefines who supplies and operates cloud and GPU capacity: Oracle, OpenAI and Nvidia are described as close collaborators for procurement and ops, Microsoft as a technology partner/tenant, while governance and ownership shares remain unclear.

Yale researcher Madhavi Singh flags acute competition risks at the infrastructure layer where market concentration is already high—three firms control ~70% of cloud services, Nvidia holds an estimated 80–95% of GPUs and TSMC ~60% of chip production—arguing Stargate could violate the Clayton and Sherman Acts by reducing future head‑to‑head competition, raising prices and blunting innovation. Regulatory and congressional response has so far been muted or supportive, increasing near‑term execution visibility for partners but leaving substantial legal and policy overhang.

Market signals in the article indicate mixed implications: consortium members that supply chips and datacenter capacity (NVDA, ORCL) stand to capture outsized demand and pricing power in the near term, while enterprise cloud customers and independent challengers face higher input risk and potential erosion of competitive options; key short‑term indicators to watch are capital deployments, contract terms, exclusivity clauses and any formal antitrust inquiries.

Key Decisions for Investors

• Tactically maintain exposure to infrastructure beneficiaries (NVDA, ORCL) to capture near‑term revenue from large capex contracts, but size positions modestly and use hedges because the venture creates meaningful regulatory overhang
• Monitor hard regulatory triggers—FTC/DOJ statements, state AG actions, congressional hearings, and the publication of Singh’s paper—and reduce gross exposure or widen hedges if formal investigations or enforcement actions emerge
• Insist on transparency: watch for disclosures of ownership, governance, exclusive supply agreements and customer pricing terms; adjust positions if the consortium secures exclusive access to GPUs or cloud capacity that could entrench pricing power
• Consider relative‑value trades or option hedges (long select infrastructure suppliers versus underweight or short cloud hyperscalers and enterprise software providers) to express a view that input‑cost pass‑through will favor suppliers over buyers if cartel‑like dynamics materialize