The first mixture-of-experts model is trained end to end off Nvidia

Claim

Zyphra's ZAYA1-8B became the first mixture-of-experts model trained end to end on an AMD stack, while Datadog's open-weights time-series model kept improving with no saturation at 2.5 billion parameters.

Reporting date: 2026-05-15 (not necessarily the event date)

Sources

Attribution and reporting linked by this record; links do not imply independent verification.

• Welcome to May 15, 2026 (theinnermostloop.substack.com)