OpenAI’s original $38 billion agreement with Amazon Web Services, announced November 3, 2025, was a multi-year commitment for AWS computing capacity, including hundreds of thousands of NVIDIA GPUs. OpenAI said it planned to use NVIDIA GB200 and GB300 clusters in Amazon EC2 UltraServers for ChatGPT inference and training next-generation models, with deployment targeted before the end of 2026. In 2026, the companies announced a further $100 billion expansion over eight years, including approximately two gigawatts of AWS Trainium capacity. Those are announced commitments and targets—not proof that all capacity has been delivered or that it has produced a particular performance result.
What was the original $38 billion AWS deal?
OpenAI and AWS announced the agreement on November 3, 2025. OpenAI described it as a $38 billion multi-year partnership whose scope would continue growing over seven years. AWS was to provide access to computing infrastructure for OpenAI workloads, including hundreds of thousands of NVIDIA GPUs and the ability to scale to tens of millions of CPUs for agentic workloads. OpenAI’s announcement does not establish that all of this capacity was already installed or in use.
The stated uses included serving ChatGPT responses, known as inference, and training next-generation models. That makes the deal an infrastructure supply arrangement: AWS provides computing capacity, while OpenAI uses it to run and develop AI systems.
What changed in the 2026 expansion?
In 2026, OpenAI and Amazon announced an expansion of the existing agreement by $100 billion over eight years. The companies said OpenAI would consume approximately two gigawatts of AWS Trainium capacity, spanning Trainium3 and next-generation Trainium4. OpenAI linked the capacity to Stateful Runtime Environment, Frontier, and other advanced workloads. OpenAI’s expansion announcement and Amazon’s announcement describe the companies’ plans; the figures are not independent measurements of delivered compute.
The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →#1 Best Overall
- PLEASE NOTE: Exporting an NVIDIA RTX Pro 6000 GPU outside the US requires strict adherence to the U.S. Export Administration Regulations (EAR) and issuance of an export license from the Bureau of Industry and Security (BIS). Compliance and Know Your Customer (KYC) screening may be required as a condition of order acceptance. [NVIDIA Blackwell Streaming Multiprocessor] The new SM features increased processing throughput, and new neural shaders that integrate neural networks inside of programmable shaders | DLSS 4: Multi Frame Generation ensures ultra-smooth frame pacing for lifelike simulations.
- [Double-Flow-Through Design] The RTX PRO 6000 Blackwell features a double-flow-through cooling design, optimizing efficiency and airflow to sustain peak performance under 600W power loads. | [5th Gen Tensor Cores] Deliver up to 3X the performance of the previous generation and support for FP4 precision for faster AI model processing times with reduced memory usage, enabling local fine-tuning of LLMs and generative AI | [4th Gen Ray Tracing Cores] Double the ray-triangle intersection rate of the previous generation to create photoreal, physically accurate scenes and immersive 3D designs with RTX Mega Geometry, which enables up to 100X more ray-traced triangles.
- [PCIe Gen 5] Support for PCIe Gen 5 provides double the bandwidth of PCIe Gen 4, improving data-transfer speeds from CPU memory and unlocking faster performance for data-intensive tasks like AI, data science, and 3D modeling. | [GDDR7 Memory] With 96 GB of GPU memory and 1.8 TB ps bandwidth, it can tackle massive 3D and AI projects, fine-tune AI models locally, explore large-scale VR environments, and drive larger multi-app workflows.
- [DisplayPort 2.1] Achieve unparalleled visual clarity and performance, driving high resolution displays at up to 8K at 240 Hz and 16K at 60 Hz. Increased bandwidth enables seamless multi-monitor setups while HDR and higher color depth support ensures superior color accuracy for precision work, such as video editing, 3D design, and live broadcasting.
- [Universal MIG] Divide a single RTX PRO 6000 Blackwell into multiple isolated instances, each with dedicated resources, allowing for concurrent execution of multiple workloads, optimized GPU utilization, and secure isolation of different applications or users. [WARRANTY] 3 YR Manufacturer's Warranty. Bulk OEM Packaging. Retail Packaging is NOT included.
The expansion also extends beyond supplying chips. Amazon said AWS would be the exclusive third-party cloud distribution provider for OpenAI Frontier, and that a joint Stateful Runtime Environment was planned for availability through Amazon Bedrock. The companies also described customized models for Amazon customer-facing applications. Trainium4 delivery was expected to begin in 2027, while product availability and timing remain forward-looking plans.
Where NVIDIA Blackwell fits
OpenAI’s 2025 announcement specifically named NVIDIA GB200 and GB300 GPU clusters running through Amazon EC2 UltraServers. Those are the Blackwell-related systems associated with OpenAI’s stated AWS compute plan. The announcement does not say that every Blackwell product AWS has discussed is part of OpenAI’s contract.
Rank #2
- Professional GPU with Blackwell Architecture
- Blackwell Architecture
- 24GB GDDR7 with PCIe 5.0 & Ray Tracing
- AI Workstation
Separately, NVIDIA and AWS described broader AWS plans for Blackwell infrastructure, including GB200 systems and B100 Tensor Core GPUs. NVIDIA’s description of the GB200 NVL72 says it combines 72 Blackwell GPUs and 36 Grace CPUs, connected by fifth-generation NVLink, with AWS networking and infrastructure. This is platform context, not a specification for every system OpenAI will use. See NVIDIA’s AWS collaboration announcement.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How to read the deal’s headline figures
| Figure | What the companies said it represents |
|---|---|
| $38 billion | OpenAI’s stated value of the original multi-year AWS agreement announced in 2025; it is a commitment, not a reported total of completed deployments. |
| Seven years | The period over which OpenAI said the original agreement’s scope would continue to grow. |
| Hundreds of thousands of NVIDIA GPUs | Compute access described in OpenAI’s 2025 announcement, including GB200 and GB300 clusters. |
| Tens of millions of CPUs | Potential scale OpenAI cited for agentic workloads under the original arrangement. |
| $100 billion over eight years | The 2026 announced expansion of the existing agreement. |
| Approximately two gigawatts | Trainium capacity OpenAI said it would consume under the expansion, spanning Trainium3 and Trainium4. |
| 72 Blackwell GPUs and 36 Grace CPUs | NVIDIA’s stated configuration for a GB200 NVL72 system; this is a separate platform description, not an OpenAI deployment count. |
These figures come from company announcements. The Associated Press report on the original deal placed it in the context of OpenAI’s broader need for computing capacity and large infrastructure commitments, and reported investor concerns about their scale and financing. Such concerns are commentary about risk, not a determination of OpenAI’s solvency.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Rank #3
- Form Factor: Plug-in Card
- Cooler Type: Active Cooler
- Maximum Power Consumption: 70W
- Length: 6.6
- Height: 2.7
What the deal does—and does not—show
The announcements show that OpenAI plans to source substantial AI computing capacity from AWS through both NVIDIA GPUs and Amazon-designed Trainium systems. They also show that the relationship is intended to support cloud distribution and other enterprise-facing services, not just model training.
They do not establish how much of the original capacity has been deployed, what share of OpenAI’s workloads will run on each chip family, or what performance the systems deliver in practice. No independent performance test or realized deployment total is reported in the cited sources. AWS and OpenAI’s announcements describe commitments, targets, and planned services; Amazon also notes that business conditions, including energy prices, resource and supply volatility, and customer demand, can affect forward-looking plans.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
