ChatGPT to Get Faster? OpenAI and Broadcom Unveil First AI Chip 'Jalapeño'

OpenAI has officially entered the chip-making business. On Wednesday, the ChatGPT maker and semiconductor giant Broadcom unveiled Jalapeño, OpenAI's first custom-built AI accelerator, marking a major step toward making the company's models faster, cheaper to run, and less dependent on Nvidia.

What Jalapeño Actually Is

OpenAI and Broadcom unveiled Jalapeño, OpenAI's first Intelligence Processor: an accelerator architected around OpenAI's vision for the future of LLM inference, and the first AI accelerator in a multi-generation compute platform the companies are building together to make advanced AI faster, more reliable, and more accessible to more people. Jalapeño was delivered to OpenAI CEO Sam Altman and President Greg Brockman by Broadcom President and CEO Hock Tan and President Charlie Kawwas, marking an important step in OpenAI's strategy to build the full stack behind its models and products.

Unlike a general-purpose chip, Jalapeño was built with a very specific job in mind. OpenAI stresses that Jalapeño is a purpose-built inference ASIC and not a repurposed training accelerator or a general-purpose AI processor. Jalapeño is specifically designed for inference — the process of running pre-built AI models in response to user commands — rather than the separate, more compute-intensive process of training new models from scratch.

Will It Actually Make ChatGPT Faster?

The early signs point to meaningful efficiency gains, though the companies are still finalizing benchmarks. Engineering samples of the Jalapeño chip are running ML workloads in the lab at production target frequency and power, including GPT-5.3-Codex-Spark, and 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. Greg Brockman noted in a CNBC interview that "this is a real performance improvement... on performance per watt and performance per dollar," and separately posted on X that "Perf[ormance] per watt looking incredible".

The cost savings could be substantial. The chip reportedly cuts inference costs by about 50%, according to Bloomberg — a figure that, if it holds up at scale, would meaningfully change the economics of running ChatGPT and Codex for millions of users.

A Remarkably Fast Development Timeline

What's drawing particular attention in the chip industry is how quickly Jalapeño went from concept to physical hardware. Jalapeño was co-developed from initial design to manufacturing tape-out in just nine months, and the custom AI accelerator program represents what the companies believe to be the fastest ASIC development cycle ever achieved in high-performance advanced semiconductors. Typically, it takes 1.5 to 2 years to design an ASIC from scratch, making this turnaround genuinely unusual for the industry.

Part of that speed came from an unusual source: OpenAI's own AI models helped build the chip designed to run them. That speed reflects deep software-hardware co-development with OpenAI's engineering teams, Broadcom's silicon implementation expertise, and the use of OpenAI models to accelerate parts of the design and optimization process — the same models served to users are helping improve the infrastructure used to run future models.

Why OpenAI Is Building Its Own Silicon

The move reflects a broader industry shift away from relying entirely on Nvidia's GPUs. OpenAI's chip plans have long been rumored as a way to reduce the company's dependence on Nvidia's GPUs, much like Google and Amazon have both built custom chips to serve a similar purpose, often called "AI accelerators". "Nobody wants to be beholden to Nvidia," Ben Barringer, global head of technology research at investment firm Quilter Cheviot, told CNN, noting that major AI players are trying to diversify their chip footprint.

Just how badly OpenAI needs more compute became clear from Brockman's own comments. Brockman told CNBC that OpenAI "cannot get compute fast enough," and Broadcom CEO Hock Tan backed up that take, saying compute demand from the company's six customers is "simply insatiable," adding, "It's just much more than we can address, and this is not just '26, not '27, we're seeing that same and even elevated demand in '28 as well".

The Bigger Financial Picture

Beyond raw performance, Jalapeño appears designed to help OpenAI's broader financial sustainability as it eyes a potential public listing. OpenAI's pivot into proprietary hardware is not just a quest for technical supremacy — it may also make its core unit economics far more sustainable, especially as the company faces scrutiny over its spending relative to revenue. The announcement comes as OpenAI is gearing up for an initial public offering that could potentially value it at a trillion dollars, putting more pressure on the company to prove it can generate revenue to match that valuation.

Broadcom, for its part, has been one of the biggest beneficiaries of custom AI silicon deals industry-wide. Shares of the chipmaker are up 10% so far in 2026 and have multiplied by almost sevenfold since the end of 2022, and the chipmaker's shares climbed further on Wednesday following the Jalapeño announcement.

What's Next for the Jalapeño Platform

This is just the opening chapter of a much longer roadmap. Jalapeño is the first step in a multi-generation compute platform designed for initial deployment by the end of 2026, expanding in the years ahead, combining OpenAI-designed accelerators with Broadcom silicon implementation, networking, and connectivity technologies. Hock Tan told CNBC there will be "small prototype development" in late 2026, with the platform scaling up from there, and future generations may carry equally spicy names like Serrano, Cayenne, Habanero, and Moruga, following the chili pepper theme set by Jalapeño.

For ongoing, authoritative coverage of how custom AI silicon is reshaping the broader chip industry, the CNBC Technology desk remains a leading source for tracking developments from OpenAI, Broadcom, Nvidia, and other major players in the space.

What This Means for ChatGPT Users

While Jalapeño won't change how ChatGPT looks or feels overnight, its real impact will likely show up gradually — in faster response times, lower operating costs that could eventually translate into pricing changes, and OpenAI's ability to serve more users without straining its compute budget. With full deployment still slated for late 2026 and ramping through 2027 and 2028, the true test of whether this custom silicon delivers on its promised efficiency gains is still to come.