That's the story. Everything else follows from it.
The round
Raised $152M
Series A · $1.35B post-money
Total raised to date $270M
Led by Prime Movers Lab
With Schaeffler, Bosch, Fubon Financial Holding Venture Capital, Aglaé Ventures
Founded ~2 years ago, London
Status Europe's first pure-play humanoid robotics unicorn
One correction worth making: Nvidia and Siemens are partners, not investors in this round. Nvidia provides the compute the robot runs on; Siemens ran a factory trial. Both matter enormously — but a technology partner and a customer are different from a backer on the cap table, and mixing them up is the kind of thing a robotics investor will notice.
Why wheels win
Factories have flat floors.
That sentence contains most of the argument. Bipedal locomotion consumes an enormous share of a humanoid's compute budget, battery, mechanical complexity, failure modes and — critically — safety-certification burden. It buys you the ability to climb stairs and cross rubble. A brownfield automotive plant has neither.
What a factory actually needs is a machine that can move to a station, reach, grip, orient a part, and place it precisely. That's manipulation, not locomotion. Humanoid built humanoid where the value is — from the waist up — and put wheels where legs would only add cost.
So the "humanoid" form factor turns out to be doing something specific: it exists so the robot can use tooling, fixtures and workspaces designed for human bodies. That's the actual economic argument for the whole category, and it has nothing to do with walking.
The cap table is the supply chain
Here's the structure that makes this more than a funding story.
→ Schaeffler invested, supplies components for the robots, and placed a 1,000-robot order → Bosch invested and will manufacture HMND 01 for the European market, with production capacity reported at 100,000 units → Siemens trialled the robot at its Erlangen factory — 60 tote moves per hour across eight-hour shifts, working alongside humans → Nvidiasupplies the compute
This isn't a startup that found customers. It's a consortium. German industrial giants are simultaneously funding it, building it, supplying it and buying it.
And that pattern should look familiar by now. Boeing took equity in Archer rather than fund Wisk itself. Airbus co-led a round in a startup whose founder says he'll displace the primes. Now Bosch and Schaeffler are financing and manufacturing a robot rather than developing one internally. Three times in three months, an industrial incumbent has concluded that owning a slice of the disruptor beats building the thing.
Where I'd be careful
The deal is partly intra-consortium. When your investor is also your component supplier and your largest customer, "largest commercial deal in history" is a slightly different claim than an arm's-length buyer choosing you over alternatives. That's not a scandal — it's how industrial scale-up has always worked, and Schaeffler putting cash and an order behind the same bet is a real signal. But it isn't independent market validation, and I'd want to see the second customer before calling it a category win.
"Thousands" is 1,000. The company describes deployment of thousands of robots in manufacturing environments; the specific figure in reporting is a 1,000-unit order. Big either way. Worth being precise about.
And the reliability gap is real. The Siemens trial ran at roughly 90% accuracy. Compare Figure's eleven-month BMW Spartanburg pilot: 1,250+ operating hours, 90,000 sheet-metal parts placed to five-millimetre precision, above 99% placement accuracy per shift, on an 84-second cycle time.
Those aren't measuring identical tasks, so it isn't apples to apples. But 90% means one intervention in ten. Across 1,000 robots working shifts, that is a meaningful volume of human rescues — which is fine for a co-working deployment and not yet good enough for lights-out autonomy.
My read
Your framing — that this is where humanoid robotics stops being a headline and becomes a factory floor — is right, and the wheels are why. This category has been slowed by an aesthetic commitment to walking that customers never asked for. Humanoid dropped it and immediately got a purchase order.
The open questions are whether that 90% climbs toward Figure's 99%, and whether an order from your own investor converts into orders from people who aren't on your cap table.
But the strategic lesson is already clear, and it generalises well beyond robots: the winning product often isn't the most impressive one. It's the one that removed the part of the problem the customer wasn't paying for.
Question for you: how much of the humanoid category's valuation is attached to bipedal walking — a capability that, on this evidence, factories don't want and won't pay for? I'd argue quite a lot of it, and that's an uncomfortable thought for a sector priced on the demo video.
