Eka Robotics has raised $150 million in a Series A at a $1.45 billion post-money valuation, according to TechCrunch. The company says its robot can pick about 1,500 objects weighing up to one kilogram per hour with more than 95% success. It combines a force-sensitive gripper with a vision-force-action model, rather than relying on vision and language alone.
Thesis
The economically meaningful unit for physical-agent progress is shifting from model capability in isolation toward reliable completion of embodied tasks under contact, force, and throughput constraints.
Observed facts
TechCrunch reports that Eka Robotics raised a $150 million Series A at a $1.45 billion post-money valuation.
1The company says its robot can pick about 1,500 objects weighing up to one kilogram per hour with more than 95% success.
1Eka combines a force-sensitive gripper with a vision-force-action model and uses physics simulation plus reinforcement learning.
1
Mechanism
Contact sensing, a flexible force-sensitive gripper, physics simulation, and reinforcement learning provide repeated physical interaction experience that can adjust grasping when objects shift or slip.
Why now
Funding and reported demonstrations suggest investor and industry attention is moving toward reliable physical manipulation rather than vision-only planning.
Implications
General-purpose robot systems may separate high-level task planning from specialized physical controllers.
Throughput and verified task completion may become more important evaluation units for embodied AI.
Evidence boundary
The reported metrics and demonstrations come from the company and cited experts; industrial durability, cost, and cross-task generalization remain unverified.
Watch conditions
Eka must show stable performance on real production lines for extended periods across varied tasks, objects, and integrator environments.