Eka Robotics Raises $150M to Build Superhuman Intelligence for the Physical World

Today, we’re announcing our $150M Series A, led by Andreessen Horowitz (a16z) and F-Prime, a venture capital fund affiliated with FMR LLC, the parent company of Fidelity Investments, with continued support from our seed lead investors Obvious Ventures and E14 Fund. Other investors include MFV Capital, Bison Ventures, and several Fortune-500 strategics. This new funding will help us scale our technology and customer deployments. 

We are starting with one of the most important problems in robotics: dexterous manipulation. Trillions of dollars of value flow through human hands, from packing delicate strawberries to assembling GPU server racks. Bringing robots into this work requires both the versatility to handle changing conditions and the performance to deliver useful work every day.

Bridging the Generality-Performance Gap

There’s a tension at the frontier of modern robotics between the goals of performance and generality. Performance means achieving speed and high accuracy. Generality is getting a robot to new tasks in new environments, with minimal retraining. Commercial deployments demand performance, i.e. speed and accuracy. To date, commercially deployed robots have sacrificed generality to achieve this performance. Much of modern robotics is rightfully focused on generality, however, generality without performance isn’t a product. It’s a lab demo. Unlocking the transformative potential of robotics requires both.

The pathway to making general-purpose robots performant is not by mimicking human data, but by mastering physics. Physics is how we and robots can reason about the world. How firmly should the robot grasp a soft raspberry? How quickly should the hand move to lift it quickly without causing damage? When has a cable been fully inserted and how should the robot respond if it isn't? Seeing an object tells a robot where it is and what it might be. Feeling the interaction tells it what is happening as the robot acts. A robot can build intuition for these tasks the same way a human can by jointly reasoning about vision and forces.

Simulation and Reinforcement Learning are the path to Superhuman Robots

Eka is building a vision-force-action (VFA) model to develop an intuition of physics through reinforcement learning at scale. Our approach has historical precedence. AlphaZero showed the pathway to achieving superhuman performance in Chess and Go, through reinforcement learning and self-play. AlphaZero was groundbreaking because it used zero human data. Large-scale reinforcement learning is also responsible for producing the incredibly powerful coding agents we've seen over the past year.

Large scale reinforcement learning is impractical using real world data. Simulation offers the scale needed, but the sim-to-real gap has historically limited its use for dexterous, contact-rich manipulation. Eka’s VFA is demonstrating that this sim-to-real barrier is falling.

In simulation, we can generate thousands of hours of robot experience in minutes, varying objects, environments, and physical conditions. Transferring that learning to the real world unlocks what we believe is the most scalable, performant, and capital-efficient path to a foundation model for manipulation.

Reinforcement learning at scale buys performance, and breadth of experience buys generality—letting Eka achieve both at once.

The results of Eka bridging the performance-generality gap are here. Our robots can pick a wide variety of objects outside the training distribution at a superhuman speed of 1,500 operations per hour, roughly two and a half times the human baseline of about 600 operations per hour. The same robot can screw in a light bulb and insert wires. Our VFA model has been showing signs of generalization across environments, objects, and tasks. Seeing the results of an earlier version of our model, WIRED wrote that “Eka’s robotic claw feels like we're approaching a ChatGPT moment for the physical world.”

What our investors see in Eka Robotics

“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. This team has the ideal DNA to push the frontier of robotics research while scaling commercial operations. We are absolutely thrilled to support Pulkit, Tuomas, and the exceptional team at Eka.”

Raghu Raghuram,
General Partner, a16z

“Beyond Pulkit’s technical brilliance, what strikes me is his ability to think ten steps ahead. He and Tuomas understood that simply teaching robots to imitate humans and expect them to work reliably at higher speed simply doesn’t work. Robots need to develop intuition for interacting with the physical world, and that requires a fundamentally different approach. That insight drives Eka’s focus on building robots that are useful, reliable, and safe.”

Habib Haddad,
Founding Managing Partner, E14 Fund

“Three words: Team and Dream. Generalized manipulation will unlock one of the largest markets in history, but solving it is one of the most difficult problems in robotics. From the time I’ve spent with Pulkit and Tuomas, and the numerous references we’ve conducted, it is clear they are among the best roboticists in the world and are uniquely suited to solve this. I could not be more excited to be on this journey with them.”

Kahini Shah,
Partner, Obvious Ventures

“From the factory floor, a robot’s value comes down to two things: its success rate and its speed over thousands of repetitions. Early on, I recognized that current approaches to generalization simply won’t get us there. Achieving the next level of performance requires a deep understanding of physics and, crucially, breaking free from our dependence on human demonstrations. Eka’s VFA model, trained through simulation and reinforcement learning, achieves this for the hardest problem in robotics, dexterous manipulation, in a way no other company has. I truly believe Eka’s approach is the only one that will hold up in real-world deployments of complex tasks.”

Shyam Kamadolli,
Managing Director, Fine Structure Ventures

What’s Next

The best products in human history, from the car to the PC to coding agents, have been truly remarkable feats of human ingenuity. They don’t just mimic what humans do. They help reimagine our lives. Robots powered by Eka’s technology will again help reimagine the future.

High-performance, general-purpose robots can transform how physical work gets done globally. They change where work happens, how facilities are designed, shorten transportation routes, make supply chains more responsive and ultimately reimagine which products are economical to make. That is the future Eka is building.

This funding will help us scale our VFA model to more tasks, expanding the capabilities of our robotic systems, and accelerate customer deployments. Our focus is on transforming these capabilities into sustained, reliable performance in real operations.

Our team brings experience from leading research institutions and companies, including MIT, Berkeley, DeepMind, Boston Dynamics, Microsoft, CMU, UPenn, Harvard, Northeastern, Caltech and Oxford. We bring diverse intellectual and cultural backgrounds, united by the belief that breakthroughs happen when different perspectives meet and people have the trust, freedom, and support to do their best work. That is Eka’s culture. 

There is much more to build. If you want to push the frontier of robotics, come build with us.

Onward!