Analysis
Enigma, a San Francisco-based physical AI startup, emerged from stealth on Monday with a $71 million seed round led by Index Ventures and Ribbit Capital, with Conviction Partners' Sarah Guo also participating. The company's pitch is disarmingly simple: build AI foundation models that make any robot, on any hardware, intelligent -- and build the interfaces on top of them so that controlling a robot feels as easy as adjusting the volume on a phone.
The founding story is doing a lot of work in generating buzz. Co-founder Jonathan Jacobi was Microsoft's youngest-ever employee, reportedly recruited into the physical AI space by Wiz founder Assaf Rappaport, and co-founded Enigma with longtime friend Gal Niv, whom he met as a teenager in hacking competitions before the two later served together in Israel's elite Unit 8200 signals intelligence unit. That pedigree -- cybersecurity-grade technical rigor applied to robotics -- is the implicit thesis behind the company's approach to building intuitive robot control layers.
Rather than launching quietly, Enigma opened what it calls the world's first live public AI robotics experience: anyone, anywhere, can go online and interact in real time with actual AI-powered robots housed in hangars in Israel and California. It's an unusually consumer-facing move for an infrastructure-layer robotics company, and one clearly designed to generate the kind of viral proof-of-concept that seed-stage physical AI startups rarely get before their first product ships.
“Enigma is entering a physical AI field that has gotten dramatically more crowded and better capitalized in 2026.”
Enigma is entering a physical AI field that has gotten dramatically more crowded and better capitalized in 2026. Figure AI, Skild AI and Physical Intelligence have all raised large rounds around humanoid and general-purpose robot foundation models, and Travis Kalanick's robotics company Atoms closed a $1.7 billion round led by Andreessen Horowitz just days before Enigma's announcement. Enigma's bet is that the interface and control layer -- not the physical hardware or even the underlying model -- is the more defensible and more urgently needed piece of the stack, since most robotics failures in the field come from clunky, brittle human-robot interaction rather than raw model capability.
The company says the new capital will go toward growing its research and engineering teams, scaling compute, and expanding real-world deployments, with early partners already lined up in entertainment, retail and health -- three verticals where an easy-to-use robot interface has an obvious near-term commercial case well before general-purpose humanoid robots are viable at scale.
For early-stage investors, Enigma is a useful data point on where physical AI capital is rotating: less toward yet another humanoid hardware play, more toward the software and interface layer that determines whether any of those robots are actually usable by a non-technical operator. A $71 million seed on a company barely a year removed from founding is a large check for an unproven interface thesis, and the real test will be whether Enigma's public robot demo converts into paying enterprise deployments faster than better-funded rivals building similar interface layers internally.
What to watch: how quickly Enigma converts its entertainment, retail and health partnerships into disclosed commercial deployments, whether larger players like Figure AI or Physical Intelligence respond by building comparable interface layers of their own, and whether the public robot demo stays online and popular past its initial launch buzz.
