OpenAI Strikes $38 Billion Cloud Deal with Amazon as AI Infrastructure Race Heats Up

OpenAI Secures Major Cloud Deal with AWS to Boost Computing Power
Partners with AMD, Nvidia, and Oracle to Strengthen Infrastructure
Microsoft, Nvidia, and Google Also Expand AI Infrastructure Investments

Artificial intelligence (AI) startup OpenAI has signed a major cloud computing deal with Amazon Web Services (AWS), further expanding its network of infrastructure partners after recent collaborations with AMD, Nvidia, and Oracle. The move underscores OpenAI’s aggressive push to secure additional computing capacity — a trend increasingly mirrored across the broader AI industry.

OpenAI–AWS Partnership

On the 3rd (local time), OpenAI announced a new seven-year cloud computing agreement with Amazon Web Services (AWS) worth 38 billion dollars. Under the deal, OpenAI will gain immediate access to AWS’s Amazon EC2 UltraServer infrastructure, equipped with hundreds of thousands of Nvidia GPUs. AWS also plans to build a dedicated infrastructure environment exclusively for OpenAI.

The additional computing power will be used to support a wide range of operations — from powering ChatGPT’s inference services to training next-generation models. OpenAI CEO Sam Altman said, “Scaling the next generation of AI requires stable, large-scale computing. Our collaboration with AWS will help bring advanced AI capabilities to more people.” AWS CEO Matt Garman added, “AWS’s optimized computing resources are uniquely positioned to support OpenAI’s massive AI workloads. Our infrastructure will serve as the backbone of OpenAI’s AI ambitions.”

Analysts view the deal as a sign that OpenAI is beginning to move away from Microsoft’s long-standing dominance and toward more independent operations. Since 2019, OpenAI has relied heavily on Microsoft’s Azure cloud services, backed by roughly 13 billion dollars in cumulative investment from Microsoft. However, after OpenAI finalized its transition into a nonprofit governance structure late last month, it decided not to grant Microsoft priority access or exclusive rights to its cloud computing contracts.

OpenAI’s Aggressive Push to Expand AI Infrastructure

OpenAI is accelerating efforts to secure computing infrastructure through partnerships with multiple key technology companies beyond AWS. A notable example is its agreement with AMD, signed on the 6th of last month, to jointly build large-scale AI infrastructure. Under the deal, OpenAI plans to deploy up to 6 gigawatts (GW) worth of AMD GPUs for next-generation AI systems. In return, AMD has granted OpenAI warrants for up to 160 million shares, exercisable upon meeting specific performance milestones. The first phase of GPU deployment will begin in the second half of 2026 with AMD’s 1 GW Instinct MI450 series, with capacity expanding gradually thereafter. AMD expects the partnership to generate several billion dollars in additional revenue.

Earlier, OpenAI also announced a strategic alliance with Nvidia. Starting in the second half of 2026, the two companies plan to build a new AI model training system using Nvidia’s CPU–GPU hybrid platform Vera Rubin. The project will require new data center hardware and roughly 10 GW of power capacity. Nvidia is expected to invest up to 100 billion dollars to support OpenAI’s infrastructure buildout.

Oracle has likewise entered into a massive cloud supply agreement with OpenAI. According to a Wall Street Journal report in September citing sources familiar with the matter, OpenAI will purchase approximately 300 billion dollars’ worth of computing resources from Oracle over five years starting in 2027 — one of the largest cloud deals ever in the AI industry.

OpenAI is also participating in the U.S. megaproject Stargate, a large-scale AI infrastructure initiative launched jointly with Oracle and SoftBank. The four-year, 500 billion dollar project aims to build several multi-GW data centers across the U.S., providing hundreds of thousands of GPU-level compute units. As part of the plan, OpenAI, Oracle, and Related Digital will begin construction early next year on a data-center campus in Saline Township, Michigan, with capacity exceeding 1 GW.

AI Infrastructure Investment Boom Sweeps the Tech Industry

OpenAI is far from alone in ramping up AI infrastructure spending. According to TechCrunch on the 3rd (local time), Microsoft recently signed a large-scale infrastructure agreement with cloud computing company Lambda. Backed by Nvidia, Lambda will supply Microsoft with tens of thousands of Nvidia GPUs, though the exact contract value has not been disclosed. Lambda CEO Stephen Balaban stated, “Lambda and Microsoft have been working together for over eight years, and this agreement marks an important milestone in building AI supercomputers.”

Nvidia has also invested 2 billion dollars in Elon Musk’s AI startup xAI as part of a 20 billion dollar funding round. The company’s move follows its major GPU supply deal with OpenAI, signaling an aggressive expansion of its partnerships with leading AI developers. One market analyst noted, “Nvidia’s strategy is a long-term investment in its own future — as AI firms grow, their expansion inevitably drives more demand for Nvidia’s GPUs, which are essential for training and deployment.”

Google, meanwhile, continues to lead in AI capital spending. The company spent 52.5 billion dollars on AI-related initiatives in 2024 and has twice raised its investment targets this year — first announcing a 75 billion dollar plan in February, then increasing it to 85 billion in July. During its third-quarter earnings call on the 29th, Google again revised the target upward to between 91 and 93 billion dollars, underscoring the escalating competition for AI infrastructure dominance.