• Nvidia (NVDA) and OpenAI's potential $100 billion investment deal, announced as a non-binding letter of intent in September 2025, remains under discussion despite reports of stalling.
  • Nvidia CEO Jensen Huang pushed back over the weekend against claims of uncertainty, calling them "nonsense" and affirming plans to invest "a great deal of money" without exceeding $100 billion, per CNBC.
  • The deal aims to deploy at least 10 gigawatts of Nvidia systems starting in late 2026, with the first 1GW on the Vera Rubin platform targeted for H2 2026.

A High-Stakes AI Partnership in Focus

Nvidia Corp, the world's most valuable company with a $4.5 trillion market cap achieved in October 2025, and OpenAI, which serves over 700 million weekly active users with AI models like ChatGPT, are navigating intense scrutiny over their proposed $100 billion investment deal. According to people familiar with the matter, discussions are ongoing, though recent Wall Street Journal reporting cited uncertainty and Huang's concerns over OpenAI's business discipline and competition from Google and Anthropic. Huang's weekend remarks, however, sought to quell such speculation, emphasizing the partnership's continuity.

Efforts to finalize the agreement have hit a snag, with analysts flagging sustainability issues. Daniel Windsor notes that OpenAI generates approximately $10 billion in annual revenue but faces profitability challenges, as each gigawatt of compute costs $50 billion with a five-year lifespan and static revenue per gigawatt despite efficiency gains. Without a deal, OpenAI could face constraints in scaling its superintelligence ambitions, potentially slowing investments and raising compute token prices amid falling efficiency-driven costs.

Nvidia's stock rose 62% over the past 12 months despite a Monday dip after the stalling report, with analysts bullish post-CES 2026 on AI demand and scaling. The company's Vera Rubin platform, with six new chips launching H2 2026, is central to the deployment plans. OpenAI, which partners with Microsoft (MSFT), Oracle (ORCL), SoftBank, and others for AI infrastructure, has emphasized benefiting humanity via AGI, but investors face dilution and debt risks in such megadeals.

Broader trends include Oracle raising $50 billion in debt/equity for AI clients like OpenAI and Nvidia, signaling investor skepticism, with Oracle shares down 50% from a September 2025 peak. The partnership builds on a decade of collaboration from Nvidia's DGX supercomputers to ChatGPT, but Nvidia's November filing stressed no guarantees on finalizing investments. Short-term, details are expected to finalize soon, while long-term, analysts predict sustained AI demand and Nvidia leadership, though Windsor warns of potential investment slowdowns and supply constraints unless economics improve.

Correction: An earlier version misstated the timeline for the first gigawatt deployment; it is targeted for H2 2026, not late 2026.