Eka Robotics Secures $150 Million to Advance Dexterous Automation

CAMBRIDGE, Mass. -- Eka Robotics has raised $150 million in Series A funding to develop artificial intelligence that helps robots grasp, move and manipulate objects in industrial environments, valuing the startup at a reported $1.45 billion.

The company announced the funding Oct. 20, identifying Andreessen Horowitz and F-Prime as lead investors. Existing backers Obvious Ventures and E14 Fund also participated, alongside Samsung Next and several other investors.

Founded by Massachusetts Institute of Technology professor Pulkit Agrawal and former Google DeepMind researcher Tuomas Haarnoja, Eka is developing a vision-force-action model that combines camera information with force feedback from robotic hands.

Unlike systems that rely primarily on vision to guide movements, Eka's approach allows robots to adjust their grip and motion in response to physical contact, helping them handle objects with different shapes, textures and resistance.

The company trains its control models using reinforcement learning in physics simulations, where virtual robots practice handling objects under different conditions before transferring those skills to physical hardware.

Eka says its robots can handle 1,500 objects per hour, roughly 2.5 times the human baseline of 600 operations per hour used in its comparison. The company also reports that its system can manipulate objects outside its training data, screw in light bulbs and insert wires.

In a separate investment announcement, Andreessen Horowitz described the technology as a potential way to automate tasks that require robots to adapt to changing conditions without extensive reprogramming.

The funding will support further development of Eka's AI model, expansion into additional manipulation tasks and faster customer deployments, according to the company.

Eka has not disclosed customer names, deployment volumes or independently verified results showing that its reported laboratory performance can be maintained during extended industrial operations.