OpenAI and Amazon Web Services have closed a cloud computing deal valued at 38.000 million (approximately €33.000 billion) over the next seven years. The collaboration will allow the creator of ChatGPT to execute and scale AI training and operations on AWS infrastructure, effectively ending Microsoft's exclusivity as OpenAI's primary cloud provider.
The news had an immediate effect on the market, with Amazon's shares advancing by around 5%–6% after the announcementBeyond the stock market movement, the agreement responds to the growing need for computing power for increasingly complex models and positions AWS as an additional pillar in the OpenAI technical ecosystem.
Details of the agreement and technological scope
According to a statement from both companies, OpenAI will gain access to hundreds of thousands of Nvidia's latest generation GPUs Hosted in Amazon data centers, with the capacity to scale to massive clusters for training and inference. The roadmap envisions configurations of up to 500.000 interconnected chipslow-latency networks and servers optimized for large AI workloads.
On a technical level, the infrastructure will combine GPUs from the family Nvidia GB200 and GB300 with high-performance instances on AWS to accelerate both model tuning and production deployment. This design aims to maximize the efficiency and performance in projects that demand enormous volumes of data and computing power.
OpenAI emphasizes that a foundation of AI is needed to drive the next generation of AI systems. massive, reliable and secure computingand that the alliance expands that ecosystem with an additional top-tier provider. AWS, for its part, emphasizes that its breadth of services and immediate capacity availability position the platform in a Privileged position to support OpenAI workloads.
The operational schedule is starting gradually: OpenAI is already running critical tasks on AWS and anticipates significant expansions before the end of 2026, with room to grow in subsequent years if demand requires it.
Impact on the market and on major suppliers
The move reshapes the cloud partnership landscape: OpenAI adds AWS as a strategic partner and It reduces your dependence on Azure, while maintaining parallel agreements with other infrastructure actorsIn recent months, the company has closed major deals with chip manufacturers such as Nvidia and AMD and with specialized data center providers.
Among the known contracts are infrastructure agreements with Oracle infrastructures and with emerging operators such as CoreWeaveIn addition to investment plans that, when combined, raise OpenAI's committed computational capacity to unprecedented levels. Various estimates place the total effort at trillion dollar levels In the medium term, this pace has also raised concerns about a possible overheating of the sector.
In parallel, Amazon is reinforcing its position as a leader in cloud services with AWS, bolstering a business that continues to invest massively in data centers and networks for AI. Demand for accelerated computing continues to outpace supply, with Nvidia GPUs becoming the most sought-after resource for training and deploying advanced generative models.
For the semiconductor market, the agreement adds visibility to the traction of generative AIMany analysts predict that investment in chips and memory for AI workloads will continue to grow, driven by hyperscale projects such as the one announced by OpenAI and AWS.
Implications for Spain and the European Union
For companies in Spain and the EU, the alliance expands access to cloud capacity for AI backed by a global provider and with a presence in European regions. This can accelerate automation projects and the development of AI-assisted software, advanced analytics and generative services in sectors such as finance, health, retail or industry.
Issues such as data residencySecurity and regulatory compliance will remain key in Europe, especially with the rollout of the AI ​​Act. The availability of large-scale computing resources on AWS will make it easier for European customers to scale pilots to production while maintaining controls. governance and privacy adequate.
Calendar and key figures
The contract has a seven-year validitywith immediate use of AWS infrastructure and substantial expansion planned between now and 2026 lateThe total amount is 38.000 million, and includes access to large Nvidia clusters and an architecture optimized for training and inferring with the most demanding models.
The operation consolidates AWS as a central component of OpenAI technical frameworkWhile the startup secures the necessary computational muscle to support its roadmap in generative models and new AI capabilities.