Nvidia CEO Jensen Huang brushed off OpenAI's Jalapeno chip and said he remains confident in Nvidia's technology and its ability to support the artificial intelligence industry, even as customers develop their own processors. According to Sina Finance, Huang said many projects start and many are canceled, and that Nvidia's role is to support partners and build world-class technology.

In a Wednesday interview, finance host Jim Cramer asked Huang whether he was upset that a company receiving major Nvidia investment was also developing chips that could challenge Nvidia. Huang said he was calm about the competition and pointed to the volatility of the AI chip industry.

Huang said Nvidia's broader platform, supply chain, and technology scale are advantages. He added that Nvidia has the supply chain and technology scale to become its partners' biggest supplier.

OpenAI released benchmark results for Jalapeno the day before Huang's remarks. According to Sina Finance, the chip was jointly developed by OpenAI and Broadcom and is OpenAI's first custom AI inference chip.

OpenAI said the processor can complete more AI tasks per unit of power and responds faster than Nvidia's GB300 in testing. Jalapeno is designed for AI inference, meaning it runs trained models and generates outputs rather than training new ones. OpenAI said the chip is part of its broader push to build more of its own AI infrastructure.

The comparison has limits, however. Jalapeno was not tested against Nvidia's newer Vera Rubin platform, and it is not aimed at AI training, where Nvidia remains firmly established.