NVIDIA developing new chips to address OpenAI's hardware concerns
What's the story
NVIDIA is working on a new processor to help OpenAI and other clients build faster, more efficient artificial intelligence (AI) systems, according to The Wall Street Journal. The tech giant is said to be developing a new system for "inference" computing, which helps AI models respond to queries quickly and accurately.
Upcoming launch
New inference computing platform expected at next month's GTC
The new inference computing platform is expected to be unveiled at NVIDIA's GTC developer conference in San Jose next month. The system will feature a chip developed by start-up Groq, the report said. This development comes after OpenAI expressed dissatisfaction with the speed at which NVIDIA's hardware can produce answers for specific types of problems such as software development and AI communicating with other software.
Strategic moves
OpenAI exploring chip sourcing from other start-ups
OpenAI is said to need new hardware for its future inference computing needs, which could account for about 10% of its requirements. The company has also been in talks with start-ups like Cerebras and Groq to source chips for faster inference. However, OpenAI's discussions with Groq were halted after NVIDIA signed a $20 billion licensing deal with Groq, according to Reuters.