Brief by Shorts91 Newsdesk / 05:56am on 03 Feb 2026,Tuesday Tech Today
OpenAI is reportedly dissatisfied with certain Nvidia AI chips, particularly for inference workloads — the process where models like ChatGPT generate responses — and has been exploring alternative hardware options since 2025, sources say. Nvidia still dominates chips for training large AI models, but inference performance has become a growing priority for OpenAI’s products, including coding tools. The company has discussed potential chip partnerships with startups like Cerebras and Groq to improve speed and efficiency, though Nvidia’s related licensing deals have complicated such efforts. OpenAI and Nvidia continue to publicly affirm their partnership even as the hardware landscape evolves. (PC: Reuters)
