OpenAI and Broadcom Introduce Faster and More Affordable Chip for Model Execution
chips openai
| Source: Bloomberg on MSN | Original article
OpenAI and Broadcom unveil a custom AI chip to boost model performance. The chip aims to run models faster and cheaper.
OpenAI and Broadcom have unveiled a custom artificial intelligence chip, dubbed Jalapeno, designed to run models faster and cheaper. As we reported on June 25, this development is part of OpenAI's bid to reduce its dependence on Nvidia. The Jalapeno chip is the result of a partnership between OpenAI and Broadcom, with OpenAI designing the chip and Broadcom co-developing and manufacturing it.
This unveiling matters because it marks a significant step in OpenAI's efforts to optimize its AI systems for better performance, efficiency, and scale. By embedding its learnings from building models directly into the hardware, OpenAI aims to improve the overall performance of its AI workloads. The chip's architecture is designed to reduce data movement and balance compute, memory, and networking resources, allowing workloads to run closer to peak performance.
What to watch next is how the Jalapeno chip will be deployed and its impact on the AI industry. OpenAI has received the first samples of the chip and is currently testing its capabilities. With plans to deploy an additional 10 gigawatts of power, the company is poised to make a significant leap in its AI computing capabilities. As the first processor in a multi-generation computing platform, the Jalapeno chip is likely to be followed by further developments in this space.
Sources
Back to AIPULSEN