← @bgSponge

OpenAI Jalapeño: Better Than Nvidia Blackwell

Summary

SemiAnalysis visited OpenAI's lab and verified InferenceX runs of Jalapeño, a Broadcom-partnered inference ASIC designed in about 16 months.

  1. It is a general inference chip, not an OpenAI-model specializer.

    SemiAnalysis ran InferenceX on DeepSeek R1, Kimi K2.5, and GPT-OSS in the lab; media overread comments about model-specific tuning.

    Everyone says that OpenAI’s chip is specialized for OpenAI models, but that’s wrong

  2. Compare it to Rubin, not Blackwell.

    Both use HBM4, Vera Rubin is already shipping, and Jalapeño is still engineering samples, so beating Blackwell is the expected bar.

    Jalapeño beats Blackwell on perf/W across almost all scenarios without being tuned for any specific point in the curve.

    we believe that comparison to Blackwell is somewhat incomplete and unfair.

  3. Power is the scoreboard.

    OpenAI is datacenter-power limited, so tokens per megawatt matter more than floorspace or budget; Jalapeño STP already beats Rubin's published MTP tok/MW.

  4. Keep the fleet homogeneous.

    OpenAI skips prefill/decode disaggregation so the mix of input, cache, and output tokens can change across model eras without stranding a specialized pool.

  5. Software codesign is the lead.

    A clean-sheet Gluon/Teacup stack plus Codex kernel search produced fast bring-up; the authors treat that pace as pressure on the CUDA software moat.

    The CUDA moat is potentially dead given how fast OpenAI can bring up new models on their silicon.

Read the original ↗Save to your Sponge

Summarized with Sponge. Quotes are the author’s own words.

More from @bg

All readings →