OpenAI, the company behind ChatGPT, unveiled its first custom chip on Wednesday in partnership with Broadcom, the two companies announced in a joint statement, a move that pushes the AI leader into designing the silicon that runs its own models and chips away at its heavy dependence on Nvidia. The processor, named Jalapeño, was delivered to OpenAI CEO Sam Altman and President Greg Brockman by Broadcom President and CEO Hock Tan, marking what the companies called an important step in OpenAI’s strategy to “build the full stack” behind its models and products.
The chip is built for a specific job. Jalapeño was designed for inference, the process of running pre-built AI models in response to user commands, rather than the more intensive work of training them. It is an ASIC (application-specific integrated circuit), a type of chip that industry experts say is less flexible than Nvidia’s GPUs but is also cheaper and can be tailored to specific AI tasks. In practice, that means it is purpose-built to serve products like ChatGPT and the company’s coding tools at lower cost.
The development speed was unusual, and AI itself helped. OpenAI said the chip was designed end to end in just nine months with help from its own AI models. President Greg Brockman said, “The degree to which our models have been able to accelerate it was very surprising to us.” The company described the effort as what may be the fastest ASIC development cycle ever achieved in high-performance semiconductors.
Early results carry a clear message to the market leader. The companies said Jalapeño provides better performance per watt than current state-of-the-art chips in early testing, a direct challenge to Nvidia’s dominance. Performance per watt has become one of the industry’s most important measurements because electricity is among the largest and fastest-growing costs of operating AI systems at scale. A more efficient chip can dramatically reduce the cost of delivering AI services.
The strategic logic extends beyond technology. OpenAI is one of the world’s largest buyers of Nvidia processors but competes with nearly every major AI company for access to those chips. Designing its own processors gives OpenAI greater control over its computing infrastructure while reducing reliance on outside suppliers. Even modest reductions in inference costs could significantly improve the economics of operating products used by hundreds of millions of people.
The partnership is also another major victory for Broadcom, which has quietly become one of the biggest beneficiaries of the AI boom by helping hyperscalers and frontier AI labs design custom silicon. Broadcom shares have climbed roughly 10% so far in 2026 and have increased nearly sevenfold since the end of 2022. CEO Hock Tan said the collaboration will enable “gigawatt-scale data centers” with Microsoft and other partners beginning in 2026, describing it as the beginning of a multi-generation roadmap.
The announcement is part of a broader shift taking place across the AI industry. Earlier this year, OpenAI reached agreements to use Amazon Web Services’ Trainium chips while also expanding partnerships with Advanced Micro Devices (AMD) and Cerebras, which completed its initial public offering in May. Together, those deals reflect a growing determination among leading AI companies to diversify beyond Nvidia for the most critical—and expensive—component of AI infrastructure.
The rollout will happen gradually. Initial deployments of Jalapeño are expected by the end of 2026, beginning with limited prototypes before expanding in the years ahead. The platform combines OpenAI-designed AI accelerators with Broadcom’s networking technology and Celestica’s board and rack systems to build complete AI computing platforms.
For businesses and consumers, the importance of Jalapeño is not that it will appear on store shelves, but that it could lower the cost of artificial intelligence itself. OpenAI argues that AI-assisted chip design can accelerate innovation while reducing computing expenses across the industry. Lower inference costs make advanced AI services more affordable and scalable as billions of users rely on them daily. By designing its own hardware, OpenAI is betting that controlling both the models and the chips powering them will be essential to driving down the cost of intelligence—and reducing dependence on the company that has dominated the AI hardware market for years.
JBizNews Desk
© JBizNews.com All Rights Reserved. Reproduction or distribution without written permission is prohibited.
