Analysis
Eka Robotics raised a $150 million Series A at a $1.45 billion post-money valuation, co-led by Andreessen Horowitz and F-Prime Capital, with seed investors Obvious Ventures and E14 Fund returning for the round, according to Tech Funding News. The Cambridge, Massachusetts startup builds dexterous robot hands using what it calls a Vision-Force-Action model that combines vision, force sensing and reinforcement-learned actions, training skills in simulation before transferring them to physical hardware.
Eka was founded by Pulkit Agrawal, an MIT associate professor and co-founder of SafelyYou, and Tuomas Haarnoja, a former Google DeepMind researcher who helped develop the Soft Actor-Critic algorithm; the two met as graduate students at UC Berkeley. The company says its system performs about 1,500 operations per hour on unfamiliar objects, versus roughly 600 for a human worker -- a claim built on company-reported simulation and hardware testing rather than independent benchmarking.
A crowded, unevenly priced field
Dexterous-hand robotics has become one of 2026's most competitive physical-AI subcategories, and Eka's $1.45 billion mark sits well below several rivals:
- Physical Intelligence -- reported valuation near $11 billion, the category's current leader.
- Skild AI -- raised $1.4 billion in January at a valuation above $14 billion.
- Generalist -- raised close to $200 million in August at a reported $3 billion valuation, roughly double Eka's mark.
- THEKER (Barcelona) -- raised an $85 million Series A in June led by CRV.
- Dream Labs -- raised a $90 million seed a day before Eka's announcement, betting on teaching robots from video feeds instead of force sensing.
Eka's price tag works out to roughly half of Generalist's and about 13% of Physical Intelligence's, showing how wide valuations have spread across startups barely a year or two old chasing similar pitches.
a16z partner Raghu Raghuram said in backing the round that Eka's model is "the first approach I've seen that truly solves manipulation" at this level of generality, a claim the company itself tempers by framing generality without performance as "a lab demo" rather than a product. That tension -- between broad, flexible manipulation and reliable, repeatable performance on real factory floors -- is the central bet every dexterous-robotics investor is making right now, and it's why valuations across the category span an order of magnitude on similar-sounding pitches.
The Series A arrives as the International Federation of Robotics reports 603,000 industrial robots were installed globally in 2025, up 11% year over year -- a growing base of factory automation that dexterous-hand startups are racing to retrofit or replace with more flexible grippers.
What the valuation gap misses
A valuation at 13% of the category leader isn't necessarily a red flag -- early dexterous-robotics pricing has swung wildly on unproven claims, and both of Eka's seed backers returning suggests insiders see something the broader market hasn't caught up to yet. But the risk cuts the other way too: none of Eka's 1,500-objects-per-hour figures have been verified by an independent benchmark, and the gap to better-funded rivals could just as easily reflect a real capability difference the a16z-F-Prime money hasn't closed yet. Whether Eka's valuation catches up, or Physical Intelligence's and Skild AI's turn out to be overextended, remains to be seen.
