On June 24, 2026, OpenAI unveiled Jalapeño, its first custom inference chip co-designed with Broadcom, and said large-scale deployment will begin at gigawatt scale in late 2026. The design-to-tape-out cycle took nine months, and OpenAI's own models helped accelerate parts of the design.

What Was Announced

On June 24, 2026, OpenAI announced Jalapeño, its first custom chip co-designed with Broadcom. Jalapeño is a chip optimized for large language model (LLM) inference: OpenAI designed it, Broadcom handles manufacturing and networking, and Celestica handles boards, racks, and system integration. OpenAI described the chip as its first "Intelligence Processor."

Performance and Timeline

OpenAI self-reported that Jalapeño's performance per watt is "substantially" better than current hardware. It did not release specific benchmark figures, however, and OpenAI promised a separate technical report. Large-scale deployment is set to begin at gigawatt scale in late 2026, and according to press reports, Microsoft is expected to purchase roughly 40% of the initial volume.

OpenAI's Own Models Helped Design It

OpenAI said the path from design to tape-out for Jalapeño took nine months. OpenAI explained that its own models accelerated parts of the design process, and engineering samples are already running real workloads including the GPT-5.3-Codex-Spark model. The first-wafer handover included OpenAI's Sam Altman and Greg Brockman, along with Broadcom's Hock Tan and Charlie Kawwas.

Why It Matters

Jalapeño matters because it signals OpenAI's move to directly control its inference cost structure. Unlike training, inference is a recurring cost incurred throughout a service's operation, so gains in performance per watt translate directly into operating margins at scale. With gigawatt-scale deployment slated for late 2026, the question is how much the custom chip can reduce reliance on outside GPUs and how far it can pull down OpenAI's serving costs.

Summary

On June 24, 2026, OpenAI unveiled Jalapeño, its first custom inference chip built with Broadcom, and flagged gigawatt-scale deployment in late 2026 after a nine-month design effort. The performance-per-watt edge is OpenAI's own claim, with a separate technical report pending, so verified figures await later confirmation.

Source: OpenAI