Analysis
Eka closed a $150 million Series A at a $1.45 billion post-money valuation, led by a16z and F-Prime, with a16z general partner Raghu Raguram co-leading the fundraising, according to TechCrunch. The company was founded by MIT professor Pulkit Agrawal and former Google DeepMind researcher Tuomas Haarnoja, who met as graduate students at UC Berkeley.
Eka's pitch is a "vision-force-action" model -- a deep-learning system paired with a force-sensitive gripper that the company says differs from the vision-language-action models most humanoid and warehouse-robotics startups are building. Where those models lean on language and vision to plan actions, Eka's architecture is built around force feedback, trained through reinforcement learning inside a physics simulator that the company says can generate thousands of hours of simulated robot experience in minutes rather than months of real-world data collection.
The numbers behind the valuation are claims, not audited results
The company says its robot can pick up roughly 1,500 objects weighing up to a kilogram per hour, compared with about 600 for a human doing the same task, and that it succeeds more than 95% of the time. Those are the company's own figures, reported by TechCrunch but not independently verified or benchmarked against a named competitor's equivalent numbers -- a gap any investor or customer evaluating the claim would need to close with their own testing. The gripper itself is described as costing under $1,000 and rated for 5 million duty cycles, with a stated goal of reaching 20 million, the standard for industrial robotic arms; getting from 5 million to 20 million is an unresolved engineering milestone, not a solved one.
Eka is running pilots with unnamed industrial customers on pick-and-place and wire-manipulation tasks, but the company hasn't disclosed customer names, revenue or a timeline to broader commercial deployment. University of Illinois Urbana-Champaign professor Kris Hauser offered positive commentary on the approach, while robotics veteran Rodney Brooks is cited more broadly as a skeptic of current generative-AI robotics models' ability to generalize -- a tension at the center of the entire physical-AI funding wave, not specific to Eka alone.
At $1.45 billion with no disclosed revenue, Eka's Series A valuation is a bet that force-based manipulation generalizes faster and cheaper than the language-model-driven robotics approaches other well-funded labs are pursuing. That bet pays off if Eka's claimed pick rates and reliability numbers hold up in real factory deployments beyond pilots; it looks expensive fast if the company needs several more funding rounds before a named customer is paying for production volume.
