Eidon & Data is a Broken business

We shut Eidon down after two years. The thesis held the whole way: embodied AI really is bottlenecked on data. The value of that data just never justified a business around it. Logistics never went away, the buyer network is maybe a couple hundred people globally, and labs wanted 3,000 hours monthly when it had taken us five months to collect 3,000 total.

We chose the easy thing twice. The glove was worth more because dexterity is where robotics actually breaks, but it was harder to manufacture and support, so we shipped trackers instead. Same call, bigger stakes, back in 2024 when we started with a gamified app. Difficulty is the moat, and the window for data companies is maybe 1 to 3 years before labs close their own flywheel.