Radar / AI robotics / Dream-1
Dream-1: robot foundation model that controls unseen bimanual robots zero-shot, at $3/hour
Dream Labs came out of stealth on 19 October with Dream-1, a robot foundation model offered through a public Dream API that turns a robot's camera feed and a plain-English instruction into motor commands for bimanual robots it has never seen, alongside a $90M seed round.
Why it matters
A video-prediction 'World Action Model' trained on human egocentric video, outputting gripper poses in a shared action space so any bimanual robot with parallel grippers can use it via its own inverse-kinematics solver; 20+ hardware partners on board, API priced from $3/hour ($2/hour for Dream-1-flash).
What you could build with it
A warehouse-automation startup could offer pick-and-pack as a service on cheap two-arm robots by renting Dream-1 by the hour, skipping months of per-robot training and charging customers per completed task.
Does it hold up?
Research beta, not validated for safety-critical use per its own materials; an independent Robocurve evaluation exists, but real-world deployment evidence beyond beta partners is thin, so controlled pilots are the realistic use today, not production floors.
Built with Dream-1
- Dream Labs Sells Robot Control by the Hour, and Its Own Data Shows the LimitsDroidGadgets · Key-facts breakdown plus the fine print: bimanual parallel-gripper scope, 77-98 minute adaptation times for non-standard grippers, and research-beta status not validated for safety-critical use.
- Dream Labs raises $90M to sell robot control through a $3-an-hour APIRuntimeWire · Launch writeup noting the founders' bet that robot makers buy a general control layer instead of building a separate policy per machine.
- Dream Labs raises $90 million seed to put one brain in any robotGoldenEra · Critical read of the independent Robocurve evaluation on RoboDojo-RC Tier 1 and the narrower technical scope behind the 'any robot' claim.
Learn more
First spotted on article: source.
More AI robotics
FLUX 3 ActionBlack Forest Labs launched FLUX 3 Action on Sept 24: a 7B open-weight world-action robotics model that takes camera…robotics · JEV 0.72RoboParty RP1 (ROBOTO 01)Full-stack open-source humanoid robot unveiled at IROS 2026 with Romomo actuators, PartyOS software, 160 N·m peak…robotics · JEV 0.67Boston Dynamics fits Atlas humanoid with a factory-ready four-fingered handAtlas's new hand has 13 degrees of freedom (nearly double the prior three-fingered generation) with pressure sensors…robotics · JEV 0.66WorldOracle — action-conditioned world model benchmark vs real-world executionsNew benchmark from Princeton, UT Austin, Toronto, NVIDIA and Zhiyuan that tests whether a world model, given the same…robotics · JEV 0.6Bear Robotics KINO-2: demonstration-to-skill robot-learning platformBear Robotics launched KINO-2 on Oct 21 at GTC Berlin: a robot-learning platform that turns human task demonstrations…robotics · JEV 0.59NVIDIA Isaac ROS 5.0: agentic AI workflows for roboticsIsaac ROS 5.0 moves to ROS 2 Lyrical/Ubuntu 24.04 and adds agent-ready skills - reusable workflows like…robotics · JEV 0.59
Get the week's best AI launches, plus 3 ideas worth building
One email every Saturday. Ranked by traction, not hype. Free.