The Groq 3 LPU chip (~500 MB on-chip SRAM, ~150 TB/s bandwidth) — deployed as the LPX rack of 256 LPUs (~128 GB aggregate SRAM) — fills the disaggregated decode/FFN slot in Vera Rubin serving (no HBM); Rubin GPUs retain prefill + attention. Rubin CPX (announced 2025-09-09) was reportedly removed from the roadmap at GTC 2026 with no official NVIDIA cancellation; the decode role was reassigned to Groq under the ~$20B NVIDIA–Groq deal.
GTC 2026 (reported)observed
| Value kind | observed — Reported measurements, counts, and specifications keep the precision and scope stated by their source; an exact specification is not treated as a range. |
|---|---|
| Scope | Rubin CPX reportedly pulled (GTC 2026, no official NVIDIA cancellation); decode reassigned to Groq LPU chip / LPX rack under the ~$20B deal |
| Caveat | Status per credible secondary reporting; no official NVIDIA cancellation or roadmap reaffirmation as of 2026-07-21. |
| As of | 2026-07 |
| Source | The Next Platform |
| Review | checking…review by 2026-08-06 · standard cadence |
| Recorded changes | last 2026-08-24 · 3 revisions tracked |
| Claim id | groq-3-lpx-prefill-sram-lpu-fills-the |
Where the guide uses it
← Full numbers register — every date-stamped figure in the guide, with revision history.