Eka Robotics Secures $150 Million Series A for AI Manipulation

Eka Robotics has raised $150 million in a Series A funding round. The investment, led by Andreessen Horowitz and F-Prime, values the Cambridge, Massachusetts-based startup at $1.45 billion. The capital will scale the company's core AI model for robotic dexterous manipulation, a critical capability for manufacturers facing labor constraints and seeking higher productivity. This funding surge aligns with a broader trend, as Crunchbase counted $47.4 billion of physical AI venture funding across 521 deals in the first half of 2026 alone.

Core technology breakthrough

Eka's system is built on a Vision-Force-Action model. It combines camera input with tactile feedback from a custom gripper. This model is trained through physics simulation and large-scale reinforcement learning, generating thousands of hours of robot experience in minutes. The approach aims for robots to master physics rather than simply mimic human demonstrations. Founder Pulkit Agrawal leads MIT's Improbable AI Lab and brings expertise in reinforcement learning and sim-to-real transfer.

The company claims this method reduces the sim-to-real barrier and enables generalization. A single robot can perform diverse tasks like screwing in a light bulb, inserting wires, and packing delicate produce. For object-picking, Eka states its robots can handle roughly 1,500 operations per hour. This compares to a human baseline of about 600.

The custom gripper costs less than $1,000 and uses a denser network of force sensors than typical robot claws. It can respond quickly enough to avoid crushing delicate items like berries. Raghu Raghuram, a general partner at Andreessen Horowitz who co-led the round, stated the technology's promise. "Eka’s Vision-Force-Action model is the first approach I’ve seen that truly solves manipulation with the generality and throughput needed for deployments," he said.

Strategic investor backing and expertise

The Series A round attracted significant venture capital. Andreessen Horowitz and F-Prime led the investment. They were joined by Obvious Ventures, E14 Fund, MFV Partners, Bison Ventures, Glasswing Ventures, Samsung Next, and unnamed Fortune 500 strategic investors. Existing backers Obvious Ventures and the E14 Fund, which led Eka's earlier $11 million seed round, continued their support.

The founding team combines deep academic and industry research. Pulkit Agrawal is an MIT professor. Co-founder Tuomas Haarnoja co-created the influential Soft Actor-Critic algorithm and previously trained humanoid behaviors at UC Berkeley and Google DeepMind. The broader team draws talent from institutions including MIT, Berkeley, DeepMind, Boston Dynamics, and Microsoft.

Investors highlighted the founders' unique approach. Shyam Kamadolli, a managing director at Fine Structure Ventures, emphasized the break from imitation. He said Eka's method achieves dexterous manipulation by building a deep physics understanding, ending dependence on human demonstrations. Habib Haddad of E14 Fund noted that robots need intuition for physical interaction, which drives Eka's strategy.

The new capital will expand the range of jobs the system can perform and accelerate commercial deployment with industrial customers. Eka is already testing its technology with unnamed manufacturers on tasks like object-picking and wire-handling. The broader market context is active. The International Federation of Robotics reported 603,000 new industrial robot installations in factories in 2025, an 11% increase, with the global operating stock surpassing 5 million.

While pilot results are promising, outside experts note the real test will be long-term, dependable operation on factory floors. Raghu Raghuram of Andreessen Horowitz expects rapid progress. He said the technology already exceeds superhuman performance on certain tasks and may achieve superhuman performance in many dexterous manipulation tasks within months.