NVIDIA's VFlash Enhances Robot Trajectory Decoding Speed by 12 Times
- Published
- Source
- InterestingEngineering.com
NVIDIA has introduced VFlash, a new visual planning method that accelerates robot trajectory decoding by up to 12 times while reducing 3D trajectory prediction errors by 61%. This innovative approach allows vision-language models to generate complete robot trajectories in a single pass, significantly improving efficiency in robot manipulation tasks.
The importance of VFlash lies in its ability to address limitations in traditional vision-language models, which typically generate spatial predictions sequentially. By enabling robots to predict longer trajectories and process detailed spatial information more quickly, VFlash enhances their responsiveness to dynamic environments, which is crucial for applications in robotics and autonomous driving.
NVIDIA reported that VFlash reduced inference latency on its Jetson Thor platform from 3,552.7 milliseconds to 703.3 milliseconds, marking an 80% decrease. Future developments may focus on further optimizing VFlash for broader applications in robotics and autonomous systems, although no specific timeline was disclosed at the time of publication.
Original report
NVIDIA’s VFlash speeds robot trajectory decoding 12x, cuts prediction errors by 61%
InterestingEngineering.com · Neetika WalterThis briefing is an independently written summary based on publicly available reporting and is provided for industry information and news discovery. The original report and source publication are credited and linked where applicable. RobotToday does not claim ownership of third-party source material.
Rights concerns? Contact [email protected] with the relevant URL and details. We will review the matter and take appropriate action where warranted.
More related news
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
- 11
- 12
Related Suppliers
NVIDIA Robotics
United States
NVIDIA robotics division delivering Isaac simulation and training tools, Jetson edge compute modules, and Omniverse digital twin frameworks for physical AI.
Deci AI (Acquired by NVIDIA)
Israel
Israeli deep learning optimization startup acquired by NVIDIA in May 2024 for ~$300M; known for AutoNAC neural architecture search technology.
Connect Tech Inc.
Canada
Guelph, Ontario embedded-hardware maker of NVIDIA Jetson Thor and Orin carrier boards and ROS-ready edge systems for robots; majority-owned by HEICO.
Vecow Co., Ltd. 維真資訊股份有限公司
Taiwan
Taiwan-based embedded computing provider specializing in fanless AI edge systems, GPU computing platforms, and robotics compute modules for industrial and.
Enactic, Inc.
Japan
Tokyo startup behind OpenArm, a fully open-source 7-DOF bimanual humanoid arm for physical AI research, and Ena, a humanoid for care facilities.
Nscale
United Kingdom
London-based Nscale runs a full-stack AI/GPU cloud platform and on Aug 1, 2026 agreed to acquire Anyscale for $1.65 billion.
Allen Institute for Artificial Intelligence
United States
Seattle non-profit AI institute founded by Paul Allen; releases fully open robot models (MolmoAct, MolmoBot) and simulators (AI2-THOR, MolmoSpaces).
Cortex Robot, Inc.
United States
YC F25 startup collecting egocentric workplace video and mobile bimanual robot trajectories, and building Cortex Harness for human-guided robot rollouts.
I Hub Research and Robotics Private Limited (iHub Robotics)
India
Kochi startup (relaunched 2022) making the TARA Gen1 semi-humanoid service robot and Daksha industrial humanoid, exported to the UAE and Saudi Arabia.
Mobile Vision Technology GmbH
Germany
German startup offering no-code mobile computer vision and AI inspection systems for manufacturing defect detection.
Decart
Israel
Israel-based AI lab founded 2023 by Dean Leitersdorf and Moshe Shalev; builds the Oasis world model and Lucy real-time video generation model for physical AI.
ADAXIS USA Inc.
France
French-founded software firm whose AdaOne platform turns industrial robot arms into large-scale multi-axis 3D printers and hybrid manufacturing cells.
Robotics needs a service framework.
RSF defines a common language for robot service capability, lifecycle operations, certification pathways, and service-provider networks.
Join the RobotToday community on LinkedIn
Daily robotics news, in-depth analysis, conference highlights, and discussions with professionals worldwide.
