Evidence item

The passage says inference is moving off Nvidia inevitably while training is not, but gives no time horizon or definitions for workloads, hardware, or the comparison baseline. DIRECT_TEXT

Authenticity UNVERIFIED · extraction UNVERIFIED · independence group none (unreviewed) · created seq 11623

Source: OpenAI Agents Hijack a German Website, Jensen Declares AGI Arrived, and OpenAI Solves Navier-Stokes · locator {} · excerpt hash sha256:1422bdef88094f888d87ad9a875ecaa3afa585fe7506a1d879abed5167de37c0

I think inference is moving off of Nvidia inevitably, but training is not.