NVIDIA and OpenAI have signed a letter of intent for a partnership that commits up to $100 billion in investment to deploy at least 10 gigawatts of NVIDIA chips for OpenAI's AI infrastructure. This move addresses the escalating demand for compute power amid intensifying competition for chips and energy resources. The first phase targets rollout in the second half of 2026, with full details to be finalized soon.
This agreement underscores the capital-intensive reality of scaling AI systems, where compute infrastructure forms the backbone of model training and deployment. For investors, it signals NVIDIA's deepening entrenchment in the AI supply chain, while OpenAI gains a pathway to sustain its growth beyond existing partnerships like Microsoft. The deal arrives as global data center power demands are projected to double by 2030, driven by AI workloads.
The partnership builds on a decade-long collaboration between the two firms. NVIDIA provided the hardware for OpenAI's early breakthroughs, including the DGX supercomputers that powered initial model development. Today, OpenAI serves over 700 million weekly active users across enterprises, small businesses, and developers, with products like ChatGPT driving widespread adoption. Yet, the path to artificial general intelligence requires exponential increases in compute capacity—far beyond current capabilities.
Under the letter of intent, NVIDIA will invest progressively, allocating up to $10 billion per gigawatt deployed. This structure ties funding to milestones, mitigating risk for both parties while ensuring aligned incentives. The initial gigawatt will leverage NVIDIA's Vera Rubin platform, an upcoming architecture designed for next-generation AI workloads. Subsequent phases will involve co-optimization of OpenAI's software stack with NVIDIA's hardware, positioning NVIDIA as the preferred provider for OpenAI's AI factory expansion.
Key executives highlighted the strategic alignment in statements accompanying the announcement.
💬 “NVIDIA and OpenAI have pushed each other for a decade, from the first DGX supercomputer to the breakthrough of ChatGPT. This investment and infrastructure partnership mark the next leap forward—deploying 10 gigawatts to power the next era of intelligence.” — Jensen Huang, founder and CEO of NVIDIA
💬 “Everything starts with compute. Compute infrastructure will be the basis for the economy of the future, and we will utilize what we’re building with NVIDIA to both create new AI breakthroughs and empower people and businesses with them at scale.” — Sam Altman, cofounder and CEO of OpenAI
💬 “We’ve been working closely with NVIDIA since the early days of OpenAI. We’ve utilized their platform to create AI systems that hundreds of millions of people use every day. We’re excited to deploy 10 gigawatts of compute with NVIDIA to push back the frontier of intelligence and scale the benefits of this technology to everyone.” — Greg Brockman, cofounder and president of OpenAI
For NVIDIA shareholders, the deal reinforces the company's dominance in AI accelerators, where it holds over 80% market share. The $100 billion commitment, while substantial, is phased over years and funded through operational cash flows—NVIDIA generated $28 billion in free cash flow in its last fiscal year. This investment secures long-term revenue from chip sales, maintenance, and software licensing, potentially adding billions to annual topline as OpenAI scales. Post-announcement trading showed NVIDIA shares up 2.3% in early sessions, reflecting market approval of the locked-in demand.
OpenAI benefits from diversified compute sources, complementing its Microsoft Azure integration and recent deals with Oracle and SoftBank. The Stargate project—a $100 billion joint venture with Microsoft for U.S.-based data centers—now gains a hardware anchor, reducing reliance on single suppliers. However, execution risks remain: supply chain bottlenecks for advanced nodes and regulatory scrutiny over energy use could delay timelines.
The broader investment landscape reveals stark challenges in AI infrastructure. Data centers worldwide consumed 460 terawatt-hours of electricity in 2022, equivalent to Japan's annual usage; by 2030, AI-driven demand could push this to 945 terawatt-hours. A single gigawatt data center rivals the output of a large nuclear plant, and 10 gigawatts would require grid upgrades costing tens of billions. Forecasts indicate AI training alone may need 50 gigawatts of new capacity by 2027. Investors in utilities (e.g., NextEra Energy) and power infrastructure (e.g., Eaton) stand to gain, as hyperscalers retrofit facilities for high-density AI racks.
Competition sharpens the focus. Microsoft, OpenAI's primary backer, invests $100 billion in Stargate but faces internal pressures from Azure's 30% AI-related capacity utilization. Google, through DeepMind, competes directly on models while building its own TPU-based clusters; a recent cloud deal with OpenAI hints at hedging strategies. Meta prioritizes open-source models with custom chips, aiming for 600,000 H100 equivalents by year-end, while Amazon's AWS trains models on Trainium hardware. These efforts underscore a bifurcation: closed ecosystems like OpenAI-NVIDIA versus integrated stacks from Google and Meta.
For portfolio managers, the Nvidia-OpenAI pact elevates compute as a non-negotiable moat. Allocate to semiconductor leaders like NVIDIA (NVDA) for growth exposure, but balance with energy plays to hedge volatility from power constraints. Diversify into AI software via Microsoft (MSFT) or enterprise adopters like Salesforce, where inference costs—projected at $1 trillion annually by 2028—drive monetization. Avoid overconcentration; AI hype has inflated valuations, with NVIDIA trading at 50x forward earnings.
Risks include geopolitical tensions over chip exports and potential antitrust probes into NVIDIA's market power. Forward-looking statements in the letter note uncertainties in technology roadmaps and deployment. Yet, the deal's scale positions both firms to capture value in a market where generative AI investments hit $33.9 billion in 2024, up 18.7% year-over-year.
This partnership quantifies the AI buildout's enormity: $100 billion for 10 gigawatts translates to $10 billion per gigawatt, a benchmark for future deals. Investors should monitor quarterly updates on deployment and capex burn, as execution will dictate returns in this capital-heavy sector.