👩🍳 How we use AI at Tech in Asia, thoughtfully and responsibly.
🧔♂️ A friendly human may check it before it goes live. More news here
🧔♂️ A friendly human may check it before it goes live. More news here
Nvidia to unveil AI processor with Groq chip for OpenAI
The Wall Street Journal reported Nvidia planned a new processor to speed AI inference and will unveil it at its GTC developer conference in San Jose next month.
The platform would focus on inference computing and include a chip designed by startup Groq, according to people familiar with the matter.
The processor was designed to help OpenAI and other customers build faster, more efficient AI systems, the WSJ added.
One source said Nvidia struck a US$20 billion licensing deal with Groq.
🔗 Source: Reuters
🧠 Food for thought
Implications, context, and why it matters.
The $20 billion Groq deal followed OpenAI’s search for faster inference chips
- Nvidia began work on a new inference processor after customer frustration surfaced, including from OpenAI.
- Reuters previously reported that OpenAI was unhappy with Nvidia’s inference speed on some tasks such as code generation.
- Reuters previously reported that OpenAI explored options to cover about 10% of future inference compute, with talks that included Groq and Cerebras.
- One source told Reuters that Nvidia signed a $20 billion licensing deal with Groq, which ended OpenAI’s discussions with the startup.
How tech giants can buy assets and talent without acquiring the company
- The $20 billion Groq agreement was structured as a licensing deal, and the reporting did not say Nvidia would acquire Groq.
- Such arrangements can lower regulatory friction compared with a full acquisition, though the reporting provided includes no market-share figures.
- Under a license, Nvidia can use Groq’s inference technology without a formal merger.
Recent Nvidia developments
Stay updated on the go with our mobile app.
Get latest insights with smoother, more personalized experience through TIA mobile app.