NVIDIA unveils VFlash: parallel decoding of visual plans for physical AI
ZhidingYu · x · 2026-11-06
NVIDIA Research (LPR Team, Jim Fan's GEAR Lab and collaborators) introduces VFlash, a System-1 VLM decoding framework for physical AI. Embodied models can think in visual plans — predicting grasp points, motions, and scene evolution — but autoregressive coordinate-token generation is slow. VFlash exploits the known geometric structure of visual primitives (boxes, points, depth, 2D/3D trajectories) to adapt speculative decoding: a lightweight drafter trained on annotated geometry generates complete primitives in parallel, yielding faster inference and more accurate robot trajectories than its autoregressive target.
Related event: NVIDIA Releases VFlash, Speeding Up Physical AI Inference 5.1x(4 posts)→
More from Embodied
- RL post-training can improve generalist robot policies on novel unseen tasks — arkrause · 2026-11-06
- Nemotron 3.5 Bolt E4B hits 27.7 tok/s on Jetson Orin Nano Super, 1.8x faster than baseline — Any_Frame9721 · 2026-11-06
- Agent builds a working satellite TV receiver on FPGA with $500 in tokens — appenz · 2026-11-06
- OpenTac brings HDR tactile sensing to solve the range-sensitivity tradeoff — _krishna_murthy · 2026-11-06
- Zipline raises $1.5B at a $21.5B valuation for autonomous drone delivery — lukas_m_ziegler · 2026-11-06
- Developer ships Drivr Glass: voice-first Grok cabin bot now displays cards in-car — bfrench · 2026-11-06