OpenAI And AMD6 Gigawatt has signed a long -term agreement with GPU distribution of capacity. In this context AMD Instinct MI450 The series GPUs will be used in the new generation of artificial intelligence infrastructure of OpenAI in the second half of 2026. The cooperation brings together AMD’s experience in the field of high -performance information processing with the goals of OpenAI, which develops large -scale artificial intelligence models. The agreement constitutes a framework that will carry the technical and commercial cooperation between the two companies to the coming years.
With this agreement signed, AMD to OpenAI 160 million shares right recognized. These rights will gradually apply to the completion of certain milestones. The first rights will be active with the introduction of 1 Gigawatt installation. In the ongoing process, new rights will be commissioned by reaching a total capacity of 6 Gigawatts. In addition, AMD will reach the designated share price targets will be part of the process. This structure allows both financial and technical goals to progress in parallel.
For the OpenAI side, this agreement includes not only the equipment supply, but also the process of expanding the production scale. The company needs a stronger information process infrastructure due to its recent increasing model complexity. Because, AMD Instinct The advantage of high memory bandwidth and low power consumption of GPUs comes to the fore. In this cooperation, Greg Brockman said that OpenAI is one step closer to OpenAI’s target of özmek increasing the calculation capacity ”goal.
AMD Instinct MI450 will grow the infrastructure capacity of OpenAI
The new agreement, before MI300x And MI350x The AMD-OPENAI, which started with the series, is a natural continuation of cooperation. These GPUs enable artificial intelligence applications with high data density to work more efficiently. On the other hand, MI450 The architecture optimized according to the server shelves of the models facilitates the installation of scalable systems. Thus, the OpenAI will have a infrastructure that can work with lower delay times in large data centers. These new generation accelerators developed by AMD aim to reduce operating costs in the long run by increasing energy efficiency.
According to the evaluations in the sector, this GPU capacity, which is the size of 6 Gigawatt, means a significant increase in a global scale in artificial intelligence services. Analysts say that OpenAI can make new breakthroughs in areas such as images, sound and video processing, as well as large language models with this capacity. In addition, AMD’s data center revenues are expected to increase with this agreement. In the period when the demand for artificial intelligence infrastructure increases rapidly, such long -term collaborations are critical for both producers and developers.
OpenAI and AMD officials emphasize that the infrastructure to be established is planned to meet the increasing demand for artificial intelligence around the world. GPU systems to be established within the scope of the agreement will be used in the training of models that require high performance processing power. The first systems will be activated in the second half of 2026 and the remaining years are expected to gradually increase the capacity. When this process is completed, OpenAI will have one of the world’s largest GPU infrastructure.