The news
NVIDIA's 31 May 2026 GTC Taipei newsroom post says Vera Rubin "is ramping into full production to power agentic AI factories worldwide."
Who is bound
Dell, HPE, Lenovo, and Supermicro are named as building at scale; Lambda is named only on Confidential Computing, with CoreWeave, Microsoft Azure, IBM Cloud, Nebius, and Nscale.
What's new
Against CES-era "full production" chip language, this post adds those OEM names, a five-rack POD, and Spectrum-X Photonics "now in production."
What it does not settle
The same page says production shipments "are set to begin starting this fall" and that products remain when-and-if-available.
MLPerf entries followed in September. NVIDIA says Vera Rubin NVL72 reached up to 3.7x the GB300 NVL72 result on Qwen3-VL in the 16 September MLPerf Inference v6.1 round. See MLPerf Inference v6.1 preview table for Vera Rubin NVL72 vs GB300 NVL72.
What to do now
Buyers with 2026 Rubin reservations should get a calendar month from Dell, HPE, Lenovo, or Supermicro this month.