OpenAI has unveiled its first in-house AI chip, and the early read from Wall Street is that it won't be the last.
On Wednesday, OpenAI announced the chip, named Jalapeño, according to ExtremeTech. The company is positioning it as the first stage in a plan to build its own "first-party" infrastructure for running large language models—the technology behind ChatGPT. ExtremeTech reports the chip is branded an "Intelligence Processor" and is designed to run ChatGPT faster and more cheaply.
OpenAI did not go it alone. According to Forbes, Jalapeño was built with Broadcom, the semiconductor giant that helps companies design custom silicon tailored to their own workloads.
The move drew quick attention from analysts. According to Seeking Alpha, the investment firm Wedbush said OpenAI's custom chip is likely to be the first of many—suggesting the company sees in-house silicon as a long-term strategy rather than a one-off experiment.
Why build your own chip? Companies that run AI at massive scale currently lean heavily on outside suppliers for the specialized processors that power their models. Designing custom hardware can lower costs, improve performance for specific tasks, and reduce dependence on a single vendor.
Why it matters: If Wedbush is right that Jalapeño is just the beginning, OpenAI is signaling a shift from being purely a software and model company toward controlling its own hardware stack—a change that could reshape how the economics and supply chains of the AI industry play out.