OpenAI diversifies AI hardware: partnership with Cerebras

OpenAI has announced the deployment of the GPT-5.3-Codex-Spark model on Cerebras hardware, in what represents the first real production deployment outside the Nvidia ecosystem. This strategic decision signals a potential diversification in OpenAI's approach to large model inference.

The collaboration with Cerebras, known for its large chips designed specifically for AI workloads, may offer performance and scalability advantages over traditional GPUs. It remains to be seen whether this partnership will lead to further developments and greater use of alternative hardware for AI model deliveries.

For those evaluating on-premise deployments, there are trade-offs to consider carefully. AI-RADAR offers analytical frameworks on /llm-onpremise to evaluate these aspects.