According to Reuters, Nvidia Corp. is planning a new computer chip designed to accelerate artificial intelligence processing, the Wall Street Journal reported on Saturday. The chip is intended to improve the speed and cost-efficiency of AI inference, the phase in which trained models generate answers or predictions, according to people familiar with the matter.
The new processor is expected to be unveiled at Nvidia’s GTC developer conference in San Jose, California, next month, the Journal said. It would complement Nvidia’s existing products, which are widely used in both training and inference tasks for large AI models.
Part of the technology in the planned chip reportedly includes components developed by AI-chip startup Groq Inc. under a licensing agreement. Neither Nvidia nor Groq responded to requests for comment, and OpenAI, a major customer of Nvidia’s chips, also declined to comment.
Reuters was unable to independently verify the details in the Journal’s report. The development comes as demand continues to grow for more advanced AI-focused semiconductors amid rapid expansion of generative AI applications and competition in the AI hardware market.