https://www.live-feeds.com/feed/openai-claims-its-new-chips-can-outperform-nvidia-processors-in-testsContinuously updated, source-cited coverage. https://www.live-feeds.com/feed/openai-claims-its-new-chips-can-outperform-nvidia-processors-in-testshttps://www.live-feeds.com/feed/openai-claims-its-new-chips-can-outperform-nvidia-processors-in-tests#u51397Sat, 29 Aug 2026 06:40:54 +0000OpenAI published the first benchmarks for its custom Jalapeño AI inference chip, claiming it achieves up to 1.9 times the efficiency of Nvidia Blackwell systems. Chip chief Richard Ho previously stated the processor outperformed Nvidia's GB300 in response speed and work completed per unit of power. OpenAI intends to deploy these chips into its infrastructure later this year to decrease its dependency on outside hardware vendors. While these results target Blackwell, the company has not provided data regarding Nvidia's Rubin architecture.Why it mattersCustom silicon allows AI develope https://www.live-feeds.com/feed/openai-claims-its-new-chips-can-outperform-nvidia-processors-in-testshttps://www.live-feeds.com/feed/openai-claims-its-new-chips-can-outperform-nvidia-processors-in-tests#u49929Thu, 27 Aug 2026 10:50:23 +0000OpenAI has unveiled Jalapeño, its first custom-designed AI inference chip, claiming it outperforms Nvidia hardware in speed and power efficiency. According to OpenAI chip chief Richard Ho, the processor led Nvidia's GB300 on a public benchmarking system regarding response speed and the amount of work completed per unit of power. One benchmark indicates Jalapeño delivers up to 1.9 times more performance per watt than Nvidia Blackwell systems. OpenAI plans to integrate these chips into its AI model infrastructure later this year to reduce reliance on external hardware providers.Why it matte