OpenAI and Broadcom Unveil ‘Jalapeño’ AI Chip OpenAI’s decision to design its own “Jalapeño” AI chip marks both a technological milestone and an escalation in the battle to control the computing power behind advanced language models, intensifying pressure on longtime supplier Nvidia.

On June 22, OpenAI and Broadcom formally unveiled Jalapeño, describing it as the company’s first “Intelligence Processor,” an accelerator “architected around OpenAI’s vision for the future of LLM inference” and the first in a multi-generation compute platform. Engineering samples are already running machine-learning workloads, including GPT‑5.3‑Codex‑Spark, at target frequency and power, with early tests indicating “performance per watt substantially better than current state-of-the-art.”

That same day, OpenAI co-founder Greg Brockman highlighted the speed of the effort, calling Jalapeño “designed from scratch for LLM inference over nine months, accelerated by our models” and saying performance per watt looks “incredible.” CEO Sam Altman offered a characteristically terse celebration: “team cooked, spicily,” attached to OpenAI’s announcement tweet about its first chip “designed from the ground up” for the LLM workloads behind ChatGPT, Codex, the API, and “future agentic products.”

By June 24, external coverage framed Jalapeño as part of a broader strategic shift. Axios reported that OpenAI has begun testing the first in a “family of homegrown chips,” aimed at handling customer queries later this year and ultimately helping power 10 gigawatts of compute by 2029. Broadcom CEO Hock Tan argued that leading AI companies “cannot, should not rely on some other third-party GPU” for such a critical capability.

The Verge emphasized that Jalapeño is an inference-focused ASIC designed to reduce dependence on scarce Nvidia GPUs, with Tan claiming performance on par with Nvidia’s Blackwell chips and Google’s TPUs. Ars Technica, meanwhile, situated the chip in a longer-term plan to “own the full stack” and squeeze more capacity from data centers amid a global compute crunch.

By June 26, TechCrunch was grouping Jalapeño with a wider industry move toward custom silicon by firms like Google, Apple, and SpaceX, casting it as a hedge against single-supplier risk and a bid for Apple-style performance gains from tight hardware–software integration. In parallel, open-source advocates like Hugging Face CEO Clément Delangue signaled renewed confidence that “Local and Opensource AI are going to win,” suggesting that cheaper, more abundant inference hardware could also empower alternatives to proprietary giants.

Continue reading https://foxvector.com/stories/019f0640-b82d-19fb-70d8-120529544230