OpenAI has announced the launch of its first custom-built artificial intelligence chip, Jalapeño, marking a significant milestone in the company’s efforts to develop more of the infrastructure that powers its rapidly expanding AI ecosystem.
The chip was developed in partnership with semiconductor giant Broadcom and is specifically designed for large language model (LLM) inference — the process that allows AI systems such as ChatGPT to generate responses in real time.
According to OpenAI, Jalapeño is the first accelerator in a planned multi-generation computing platform designed to make AI services faster, more reliable, and more cost-effective.
Early testing by the company suggests the chip delivers substantial efficiency improvements over current industry-leading hardware.
“While OpenAI is still measuring final performance, early testing shows that Jalapeño will deliver performance per watt substantially better than current state-of-the-art,” the company said.
OpenAI added that a comprehensive technical performance report will be released in the coming months.
Speaking on the development, OpenAI President and Co-Founder Greg Brockman described the chip as a key component of the company’s long-term infrastructure strategy.
“The world is moving to a compute-powered economy. Jalapeño is part of our long-term full-stack infrastructure strategy to make compute more abundant, resulting in AI which is faster, more reliable, more affordable for people and businesses, and can be used to solve more important problems,” Brockman said.
Richard Ho, who leads OpenAI’s hardware program, revealed that the chip was designed entirely around the specific requirements of advanced AI models.
“We optimized the architecture around the kernels, memory movement, networking, and serving patterns that matter most for frontier AI models. Based on early testing, Jalapeño will efficiently execute our most important workloads close to the hardware’s theoretical limits,” he explained.
Broadcom CEO Hock Tan also highlighted the significance of the partnership, describing it as a long-term commitment to building the infrastructure needed for the next era of artificial intelligence.
“Our collaboration with OpenAI represents a fundamental commitment to scaling the physical infrastructure required for the next decade of AI. This is just the beginning of a multi-generation roadmap,” Tan said.
He added that the technology could support gigawatt-scale data centre deployments with Microsoft and other strategic partners beginning in 2026.
Unlike many AI chips adapted from existing hardware platforms, Jalapeño was designed from the ground up as a dedicated inference accelerator tailored specifically for OpenAI’s workloads.
The company said the design was informed by real-world demands from products such as ChatGPT, Codex, its developer APIs, and future AI agent systems currently under development.
Engineering samples of the chip are already running machine-learning workloads inside OpenAI’s laboratories, including the company’s GPT-5.3-Codex-Spark model.
OpenAI also revealed that Jalapeño progressed from initial design to manufacturing tape-out in just nine months, which it described as the fastest ASIC development cycle ever achieved in advanced semiconductor development.
The company credited part of that achievement to its own AI systems, which helped accelerate parts of the chip design and optimization process.
Jalapeño is expected to become the foundation of a broader multi-generation compute platform that OpenAI and its partners plan to begin deploying by the end of 2026.
The announcement comes as OpenAI continues expanding beyond chatbots into a wider AI ecosystem. Earlier this month, reports indicated that the company is preparing a major redesign of ChatGPT to integrate coding tools, AI agents, and third-party services into a single platform, transforming the chatbot into a more comprehensive AI-powered “super app.”
With Jalapeño, OpenAI is taking a major step toward controlling more of the technology stack that powers its products, reducing dependence on external hardware providers while positioning itself for the next phase of AI development.
