OpenAI and Broadcom unveiled Jalapeño, their first jointly developed AI chip, on Wednesday, with initial deployment targeted for the end of 2026 as OpenAI moves to diversify beyond Nvidia's GPUs and build out its own silicon stack.
OpenAI and Broadcom on Wednesday unveiled Jalapeño, their debut custom AI chip, marking the ChatGPT maker's first entry into artificial intelligence silicon eight months after the two companies publicly announced their chip partnership.
The companies are designating Jalapeño an "Intelligence Processor" and describe it as the first "AI accelerator" in a platform being built "to make advanced AI faster, more reliable, and more accessible to more people," according to a joint statement. A physical sample of the chip was delivered to OpenAI on Wednesday.
The chip is an application-specific integrated circuit, or ASIC — a design that industry experts say is less flexible than a traditional graphics processing unit but less expensive and optimized for specific AI tasks. OpenAI said it designed Jalapeño in nine months and also crafted large portions of the broader computer system in which it will run.
Broadcom will manufacture the chips, which OpenAI intends to use for inference — the compute-intensive process of serving AI models to users in ChatGPT and other applications. The companies are targeting initial deployment by the end of 2026, "expanding in the years ahead," they said.
"By designing more of the stack ourselves, we can serve more intelligence with greater efficiency and keep pushing advanced AI toward broader access," Greg Brockman, OpenAI's president, said in the statement.
The announcement reflects OpenAI's broader ambition to "build the full stack behind its models and products." Since the company helped ignite the generative AI boom in 2022, it has been among the largest buyers of Nvidia's graphics processing units. Surging demand has driven OpenAI to diversify its silicon supply across multiple partners.
Earlier this year, OpenAI forged an agreement with Amazon Web Services that includes access to the company's Trainium AI chips. The company has also signed deals with Nvidia rival Advanced Micro Devices and with AI chipmaker Cerebras, which held its initial public offering in May.
OpenAI and Broadcom had been collaborating for 18 months before going public in October with plans to develop and deploy racks of OpenAI-designed chips starting late this year, ultimately aiming to build enough infrastructure to require 10 gigawatts of power.
Broadcom has emerged as one of the primary beneficiaries of the generative AI buildout by helping hyperscalers and frontier AI labs develop custom silicon. Shares of the chipmaker are up 10% so far in 2026 and have multiplied by nearly sevenfold since the end of 2022.
Wednesday's reveal arrives amid a broader reshaping of the AI chip landscape. UBS on Wednesday raised its 12-month price targets on Advanced Micro Devices — to $670 from $470 — and on Arm Holdings — to $455 from $260 — citing accelerating demand for central processing units in agentic AI deployments. AMD shares have climbed 153% over the past three months, while Arm has risen 171% in the same period, according to market data.
"We raise estimates for AMD and ARM as key beneficiaries of standalone CPU and head node deployments," UBS analyst Timothy Arcuri said in a note to clients Wednesday.
With Jalapeño now heading toward production, OpenAI's push to control more of its own hardware stack signals an intensifying race among frontier AI labs to reduce dependence on any single chip supplier — a dynamic that stands to reshape procurement relationships across the industry well into the next decade.
Marc Sabatini is a staff writer at TechEchelon covering enterprise software, cybersecurity, and the regulatory beats that shape both. He focuses on the deal flow and policy decisions that move markets.
More from Marc →Cerebras Systems reported $193.4 million in first-quarter revenue, up 92% year over year, in the AI chipmaker's first earnings release since its Nasdaq IPO in May 2026. Shares fell 7% in after-hours trading despite the strong top-line results.
AI memory startup Engram raised $98 million from General Catalyst, Kleiner Perkins, and Sequoia, claiming its models can match frontier AI labs using up to 100 times fewer tokens, as enterprises look to cut soaring AI costs.
Oracle cut 21,000 jobs — nearly 13% of its workforce — over the past year, citing AI deployment, while spending $1.8 billion on restructuring costs as it races to build out AI infrastructure.
Written for readers who already know the basics — markets, AI, and the policy decisions that shape both.