Reuters reported via Yahoo Tech on June 24, 2026 that OpenAI showed off the company's first custom artificial intelligence chip designed in conjunction with Broadcom, as it seeks to speed development of its infrastructure.
Why OpenAI is building custom silicon
- AI labs such as OpenAI and Anthropic are struggling to obtain enough computing horsepower to run the latest, most powerful chatbots and coding apps.
- OpenAI has turned to developing in-house chips to reduce cost and create an alternative to Nvidia graphics processing units commonly used for AI.
What Jalapeño is designed to do
- OpenAI engineers designed the chip, called Jalapeño, together with Broadcom to perform inference—crunching data to answer a user's query to a chatbot like ChatGPT.
- Broadcom CEO Hock Tan told Reuters the chip is as good as Nvidia Blackwell chips or Google tensor processing units.
- OpenAI hardware chief Richard Ho said Jalapeño is designed to work speedily and efficiently with the large language models that power many AI applications.
- "It will be performant on, we think, all kind of future iterations of LLMs," Ho told Reuters.
Deployment and manufacturing partners
- OpenAI plans to deploy Jalapeño by the end of 2026 as the first step in a multi-generation chip development plan.
- Canadian electronics manufacturer Celestica will build the server systems, which, like the chips, will be used only by OpenAI.
- OpenAI said it has lab samples running at target power and performance with GPT-5.3-Codex-Spark.
- Engineers completed the design in roughly nine months before sending it to TSMC for manufacturing, partly using AI to speed specific aspects of the process.
Industry context Reuters highlights
- Reuters first reported OpenAI was exploring making its own chip in 2023.
- Meta, Amazon, and Google have turned to Broadcom and Marvell for design services and intellectual property when building in-house chips.
- Anthropic is weighing building an AI chip of its own, sources told Reuters in April.
Economics and supply chain
- Tan said Broadcom's profit margin on custom AI chips is not as high as some other products, such as networking switches, because of the AI-related surge in memory demand.
- AI chips require large amounts of high-bandwidth memory, which challenges Broadcom's margins on custom AI chip products, Tan said.
- SK Hynix and Samsung Electronics supply Broadcom with memory chips.
Primary source: Yahoo Tech (Reuters) — OpenAI unveils custom chip it designed with Broadcom to boost its AI infrastructure (June 24, 2026).