OpenAI and Broadcom plan AI chips and challenge Nvidia
OpenAI and Broadcom want to develop their own AI chips together. The chips are expected to be ready next year. Initially, the chips will be available exclusively to OpenAI.
The details
-
OpenAI will develop its own AI chips to reduce its dependence on Nvidia. Broadcom is reportedly to act as a partner for chip manufacturing.
-
According to the Financial Times, Broadcom’s CEO told investors about a new customer with an order worth more than 10 billion US dollars. OpenAI has not been officially confirmed, but people familiar with the matter have confirmed it.
-
OpenAI increasingly needs more computing power for its AI models. The new chips are mainly for use within the company. Whether the chips will also be available on the open market is unclear.
-
Other tech companies, such as Google, Meta, and Amazon are also working on their own chips. Broadcom has already developed TPUs (Tensor Processing Unit) with Google and now seems to want to take a similar direction with OpenAI.
Our thoughts
OpenAI CEO Sam Altman has wanted to develop his own AI chips for several years. He knows very well that the demand for computing power will increase significantly in the coming years.
OpenAI’s own AI chips would have the advantage that OpenAI could tailor them to its own needs. This will enable OpenAI to get more performance out of the chips in the long term and even save money, because Nvidia chips are expensive.
More information: 🔗 Heise Online | Financial Times
Magic AI tools
Boost Your Productivity with AI
Explore the best AI tools to boost your efficiency and productivity.
AI Tutorials for Devs
- How to Build a Segmented Revenues API for Public Companies with Python
- Build a Local AI Agent with Memory Using Agno and gpt-oss
- Building a beautiful-looking Stock Dashboard with Reflex and OpenBB
😀 Do you enjoy our content? If so, why not support us with a small financial contribution? This helps us fund our work to ensure we can stick around long-term.
