Every funding round, launch and paper, as it's reported.
A vision-language-action (VLA) model is one neural network that reads camera images and a plain-language instruction and outputs a robot's next motor commands, built by training a vision-language model on robot demonstrations. Google DeepMind, Physical Intelligence, NVIDIA, Figure AI, Skild AI and AgiBot lead, and the commercial race now turns on the cost of action data rather than model design.
At GTC Washington on October 28, 2025, NVIDIA expanded its Mega Omniverse Blueprint for factory-scale digital twins, with Siemens, FANUC and Foxconn among the first partners, and named Figure, Agility Robotics, Amazon Robotics and Skild AI as robot developers on its stack. It also unveiled IGX Thor, a 5,581 FP4 teraflop industrial and medical edge computer due in December 2025.
Figure AI said on 15 March 2025 it has set up BotQ, a vertically integrated factory whose first-generation line can build up to 12,000 humanoids a year. The company redesigned its next robot around molded, cast and stamped parts, built its own factory software, and says its supply chain can scale to 100,000 robots in four years.