OpenAI Signs $38 Billion Compute Deal with Amazon Web Services

OpenAI has signed a landmark $38 billion, seven-year agreement with Amazon Web Services (AWS) to expand its artificial intelligence (AI) computing capacity, marking one of the largest cloud infrastructure partnerships in tech history. The deal positions AWS as a key strategic partner in powering OpenAI’s next-generation models and broadens the company’s infrastructure base beyond its long-standing relationship with Microsoft.

Announced on November 4, 2025, the deal grants OpenAI access to hundreds of thousands of cutting-edge NVIDIA GPUs, including GB200 and GB300 models, alongside millions of CPUs hosted on AWS servers. The full deployment is scheduled to be completed by the end of 2026, with potential expansion into 2027 and beyond.

OpenAI CEO Sam Altman highlighted the deal as essential to scaling frontier AI systems that require massive, reliable compute power. AWS CEO Matt Garman emphasized that AWS’s infrastructure is designed to handle large-scale AI workloads, referencing clusters exceeding 500,000 chips already operating within its network.

This development follows OpenAI’s recent corporate restructuring, which officially ended Microsoft’s right of first refusal for computing services. While Microsoft remains OpenAI’s frontier model partner, with exclusive rights to certain intellectual property and Azure API integration, this new agreement signals OpenAI’s growing independence and diversification across multiple cloud providers.

Microsoft’s cumulative investment in OpenAI has reached $13 billion since 2019, with its stake valued at approximately $135 billion after OpenAI’s transition into a public benefit corporation. Despite that partnership, OpenAI’s latest move diversifies its cloud strategy while maintaining collaboration with Azure for specific AGI-related projects.

Under the AWS partnership, OpenAI will deploy advanced AI architectures optimized for performance and efficiency. Amazon EC2 UltraServers will host NVIDIA GPU clusters with low-latency interconnections designed to power both training and inference workloads, including the next generation of ChatGPT and enterprise-grade reasoning models.

Industry observers see the agreement as evidence of skyrocketing global demand for AI compute infrastructure. NVIDIA CEO Jensen Huang has noted that new AI reasoning systems require exponentially greater compute capacity than earlier models. McKinsey projects that global data center investment could reach $6.7 trillion by 2030 to meet these demands.

Following the announcement, Amazon’s stock surged to record highs, adding roughly $140 billion in market value. For AWS, securing OpenAI strengthens its competitive position against Microsoft Azure and Google Cloud in the escalating race for AI infrastructure dominance. AWS currently leads the global cloud market, reporting more than 20% year-over-year growth in its latest earnings report.

The deal, however, presents new complexities for Amazon due to its parallel investment in Anthropic, an OpenAI competitor. AWS has invested $8 billion in Anthropic and is building an $11 billion data center campus in Indiana dedicated to that company’s workloads. AWS officials clarified that OpenAI’s infrastructure will remain entirely separate from Anthropic’s Project Rainier systems.

OpenAI’s global compute expansion includes similar agreements with NVIDIA, Broadcom, Oracle, and Google, amounting to over $1.4 trillion in total commitments. Despite rapid revenue growth, projected to reach $20 billion by the end of 2025, the company continues to operate at a loss, reporting $12 billion in third-quarter deficits alone.

Analysts say the AWS deal cements OpenAI’s position among the few companies capable of operating at the trillion-parameter scale necessary for AGI development. It also underscores the increasing concentration of AI power within a handful of hyperscale cloud providers, AWS, Microsoft Azure, and Google Cloud, raising questions about market competition, fairness, and global AI governance.

The partnership further integrates OpenAI into Amazon’s broader AI ecosystem. Earlier this year, AWS introduced OpenAI’s foundation models on Amazon Bedrock, allowing enterprises to deploy OpenAI technology natively through the AWS platform.

Industry experts view this as a pivotal moment in AI’s evolution, where innovation is no longer just about algorithms, but about industrial-scale compute, infrastructure engineering, and global supply chain dominance.

OpenAI Launches ChatGPT Atlas, an AI Browser Poised to Redefine Online Search