OpenAI unveils first ai chip, marking a major step in the company’s effort to strengthen its AI infrastructure and reduce its reliance on external hardware suppliers. The chip, called Jalapeño, was developed in partnership with Broadcom and is expected to be deployed by the end of 2026.
As AI models become larger and more demanding, access to computing power has become one of the biggest challenges facing companies like OpenAI and Anthropic. Building custom hardware could help reduce costs, improve efficiency, and give AI firms greater control over the technology powering their products.
OpenAI’s First AI Chip: What Is Jalapeño?
Jalapeño is OpenAI’s first internally designed AI chip and was built specifically for a process known as inference. Inference is the stage where an AI model generates responses to user requests, such as answering questions in ChatGPT or assisting with coding tasks.
According to OpenAI, the processor was designed to deliver fast and efficient performance when running large language models. The company says the chip is intended not only for today’s AI systems but also for future generations of increasingly powerful models.
Designed With Broadcom
The chip was developed by OpenAI engineers in collaboration with Broadcom, one of the world’s leading semiconductor companies. Broadcom provided the design expertise and intellectual property needed to help bring the project to life.
Broadcom CEO Hock Tan told Reuters that Jalapeño performs at a level comparable to Nvidia’s Blackwell chips and Google’s Tensor Processing Units (TPUs), two of the most advanced AI accelerators currently available.
Why OpenAI Is Building Its Own Chip
For years, Nvidia has dominated the AI hardware market, supplying the GPUs used to train and run many of the world’s most advanced AI models. However, demand for AI computing power has exploded, making access to high-performance chips increasingly competitive and expensive.
By developing its own hardware, OpenAI hopes to gain more control over costs, optimize performance for its specific models, and reduce dependence on third-party suppliers. The move reflects a broader trend across the technology industry, with companies such as Google, Amazon, and Meta also investing heavily in custom AI chips.
Already Running GPT Models
OpenAI says early samples of the Jalapeño chip are already operating inside its laboratories. According to the company, the hardware has successfully reached its target power and performance levels while running GPT-5.3-Codex-Spark, one of OpenAI’s latest AI models.
The company also revealed that it took approximately nine months to complete the chip’s design before sending it to Taiwan Semiconductor Manufacturing Company (TSMC) for production. OpenAI noted that AI tools were used to help accelerate parts of the design process.
The Beginning of a Long-Term Strategy
OpenAI describes Jalapeño as the first step in a multi-generation chip roadmap. Rather than a one-off project, the company appears to be building a long-term hardware strategy aimed at supporting future AI systems.
Server systems built around the new chip will be manufactured by Canadian electronics company Celestica and will be used exclusively by OpenAI.
What This Means for the AI Industry
The launch of Jalapeño highlights a growing shift in the AI industry. Companies are no longer focused solely on developing powerful models—they are increasingly investing in the hardware required to run them.
As AI becomes more central to business, software, and everyday technology, control over computing infrastructure may become just as important as control over the models themselves. OpenAI’s move into custom chip development suggests the next phase of the AI race could be fought not only in software, but also in silicon.
OpenAI’s hardware ambitions come at a time when competition for top AI talent is becoming increasingly intense.
Conclusion
OpenAI’s first custom AI chip represents more than just a new piece of hardware. It signals the company’s ambition to build a larger share of the infrastructure behind its AI products and reduce its dependence on external suppliers. While Jalapeño is only the beginning, it could play an important role in how OpenAI develops and scales future generations of AI systems.
