Nvidia Offers $100 Billion to Secure OpenAI’s Chip Purchases

Nvidia has made two landmark moves that underscore its dominance in the AI era: a $5 billion investment in Intel and a $100 billion commitment to support OpenAI’s infrastructure ambitions. These decisions highlight Nvidia’s strategy not just to lead in AI hardware, but to shape the entire ecosystem by securing key partnerships and ensuring long-term demand for its technology. The Intel deal, finalized after months of technical collaboration, sees Nvidia purchase $5 billion in Intel stock and co-develop CPUs with NVLink Fusion chiplets. This enables Intel Xeon processors to connect directly to Nvidia’s HGX and NVL72 server systems—previously only possible with Nvidia’s own Grace Arm CPUs. The integration will allow hyperscalers and cloud providers to build custom AI systems using Intel’s Xeon CPUs while leveraging Nvidia’s high-speed interconnects. Intel confirms it will continue its GPU roadmap, including integrated Arc and future “Jaguar Shores” accelerators, but now with a strategic partnership that strengthens its position in AI workloads. The move also signals a potential boost for Intel Foundry, which needs high-volume customers to justify its advanced 18A and 14A processes. More striking is the $100 billion agreement with OpenAI. Rather than a loan, Nvidia is investing in OpenAI in exchange for equity—likely around 2% based on OpenAI’s $500 billion valuation. The commitment covers at least 10 gigawatts of AI compute capacity, equivalent to the peak power draw of New York City. The first gigawatt is expected online by late 2026, powered by Nvidia’s next-gen Vera Rubin platform. This system, expected to deliver 3.6 exaflops per rack at FP4 precision, will require up to 5 million GPUs—more than Nvidia’s entire annual output. The investment will be disbursed incrementally as infrastructure is deployed, creating a self-reinforcing cycle: OpenAI uses the funds to buy Nvidia hardware, which fuels Nvidia’s revenue and cash flow. This model echoes past vendor financing schemes, but with critical differences. Unlike telecom-era loans, this is equity—no repayment risk. With over $570 billion in cash and projected $970 billion in free cash flow, Nvidia can absorb any shortfall. Analysts see this as a strategic bet to lock in OpenAI as a long-term anchor customer amid rising competition from Google’s TPUs, AMD, and startups like Groq. It also counters OpenAI’s own efforts to develop custom chips with Broadcom. The deal complements, but does not replace, OpenAI’s earlier “Stargate” project—its $500 billion plan with Oracle and SoftBank. Oracle is rumored to be building and managing the physical infrastructure, while Nvidia supplies the AI hardware. The new agreement appears to clarify OpenAI’s path forward, though details on coordination remain sparse. Market reaction was swift: Nvidia’s stock rose nearly 4%, adding $1.7 trillion to its market cap. The partnership reinforces Nvidia’s role not just as a chipmaker, but as the central nervous system of the AI economy. By investing in OpenAI and integrating with Intel, Nvidia is building a moat around its ecosystem—ensuring that the most ambitious AI projects depend on its technology. As the AI race intensifies, these moves signal that Nvidia isn’t just selling hardware; it’s shaping the future of computing itself.