Enigma emerges from stealth with $70M to make human-robot interaction feel like turning a volume knob

Enigma's $70M Bet: The Robotics Bottleneck Is the Knob, Not the Brain

The robotics industry is racing to build the smartest brain. Physical Intelligence, Skild AI, Figure, and a crowd of well-funded labs are all pushing on the same frontier: foundation models that let machines perform tasks they were never trained for. Enigma, a research lab emerging from stealth today, is putting $70M behind a different thesis. The bottleneck is not intelligence. It is the interface.

Co-founder Jonathan Jacobi describes the gap with an analogy. Imagine adjusting your car stereo by typing in a volume percentage instead of turning a knob. You would have no feel for whether 47% is too loud or too quiet until you tried it. That is what commanding a robot feels like today. The models can already do surprising things, but the friction of telling them what to do is so high that most people would rather do the task themselves 1.

Enigma's answer is a public experiment at a scale unusual for a company less than a year old. More than 100 robotic arms, which Enigma says it designed and built from the ground up alongside its own underlying models, sit in hangars in Israel and California, connected to the internet. Anyone can log on and direct a robot to paint with a brush, sword-fight with another machine, or mix flasks of liquid. The purpose is not to show off what the robots can do. It is to harvest data on how humans instinctively try to communicate with machines 1.

The questions are deliberately open-ended. Should you type instructions, speak them, show a video, or drag and drop? Jacobi says Enigma does not have answers yet 1.

$70M is a large seed round by any standard. The round was led by Index Ventures and Ribbit Capital, with participation from Sarah Guo of Conviction Partners 1. What makes the check size meaningful is not just the dollar figure. It is that investors who can presumably tell a capability problem from an interface problem chose to fund the interface.

The founders arrived at this thesis from the outside. Jacobi, recruited to Microsoft by Wiz founder Assaf Rappaport as the company's youngest-ever employee, and Gal Niv met as teenagers in hacking competitions and bonded during their service in Israel's Unit 8200, where they conducted cybersecurity research 1. Neither is a roboticist. When they decided to start a company last year, they chose AI for robotics precisely because they lacked direct experience and believed that distance gave them a fresher angle on problems insiders had stopped questioning 1.

Their lead investor shares that read. Index Ventures partner Shardul Shah argues that insiders in the current embodied-intelligence wave tend to start from capability, asking how to improve teleoperation or dexterity. Enigma starts from the experience and asks what the human actually wants 1.

The obvious counterargument is that model capability is still the binding constraint. If today's robots cannot reliably fold a shirt or clear a table, perfecting the interface is like designing a steering wheel for an engine that does not run yet. Enigma's bet assumes that capability is already further along than most people realize and that the frustration users feel is not "the robot cannot do this" but "I cannot tell the robot what I want without a paragraph of instructions."

Enigma's robots can already paint, sword-fight, and mix chemicals, which suggests a baseline of physical competence 1. The question is whether the gap between "the robot can do it" and "I can make the robot do it" is where the next wave of value accrues.

Jacobi says Enigma is already working with companies in healthcare, logistics, and entertainment, though he declined to specify how 1. The business model, for now, is as unresolved as the interface question.

For builders and investors, the signal is straightforward. Every major robotics lab is optimizing the model. Almost no one is optimizing the moment of contact between human intent and machine action. If Enigma is right that interaction, not intelligence, is the constraint, the next dominant robotics company will look less like a model lab and more like a design studio, and the moat will be in experience, not parameters.