The gold rush cliché says you should sell shovels, not dig for gold. In the robot boom, the shovel is a chip the size of a postage stamp — and one company just got paid to keep making it.

D-Robotics closed a $400 million Series C on September 17 (PR Newswire), the biggest single startup round of the day and, by one outlet's count, the largest Chinese robotics raise in four years. The company doesn't build robots. It builds the compute that goes inside them — and it has already shipped more than 8 million of those chips.

Here's the part that should make you sit up: the newest chip in that line, the Sunrise S600, landed in more than 20 robot companies within six months of launch. Many of them compete directly with one another. All of them are buying from the same supplier.

When the buyers can't agree on anything except who makes their brains, the brain-maker wins.

🧠 Why This Matters

Every headline about humanoid robots is about the robot — the walking, the backflips, the coffee it can't quite pour. Almost none of them are about the silicon underneath. That's the layer D-Robotics quietly moved into.

The company was spun out of autonomous-driving chip firm Horizon Robotics in 2024 to do one thing: sell the computing platform, operating system, and developer tools that a robot startup would otherwise have to build from scratch. Instead of betting on a single robot design, it sells picks and shovels to everyone digging.

That's a very different risk profile from the robot startups it supplies. A humanoid company lives or dies on one product. A chip-and-software layer underneath the whole category collects regardless of which specific robot wins — as long as some of them ship.

An integrated hardware-software infrastructure for AI robots, supporting every development scenario, from mature robot categories to general-purpose humanoid robots.— D-Robotics, describing its own platform in the Series C announcement

📊 Deep Dive

The $400 million wasn't a first date. It's the third round in eighteen months, and the ladder tells the story better than any pitch deck:

That's roughly $770 million raised across three rounds, with each round bigger than the last. No valuation was disclosed for the Series C — more on that below.

The traction underneath it is concrete. Cumulative shipments of the Sunrise chip series passed 8 million units. The flagship S600, launched in November 2025, reached 20-plus embodied-AI customers in half a year — a roster that includes UBTECH, Fourier, Astribot, X Square Robot, Spirit AI, TARS, PaXini Tech, and Booster Robotics. Around that hardware sits a developer ecosystem the company puts at more than 500 universities and over 100,000 developers across 20-plus countries, plus a "Gravity" program supporting 500-plus robotics teams. Revenue, the company says, grew "several times year over year" in the first half of 2026.

The money goes two places: expanding the Sunrise chip lineup across every compute tier, and building a software platform that spans the entire robot-development chain — train in simulation, test virtually, deploy to hardware without rewriting the code. Sell the chip, then sell the toolchain that only runs well on the chip. It's the classic platform playbook, pointed at robots.

⚠️ The Catch

The customers are startups too. Twenty buyers sounds like a moat until you remember most of them are pre-revenue humanoid companies burning venture money on robots that don't have a mass market yet. D-Robotics' order book is only as durable as their runways. If the humanoid hype cools before the robots find paying work, the shovel-seller feels it fast.

Then there's what wasn't said. No valuation, no named lead investor. "Several times year over year" is a growth rate without a starting number — impressive-sounding, impossible to check. A round this size with this little disclosed is worth noting, not celebrating.

And as a China-based chipmaker, D-Robotics operates in a sector shadowed by export-control uncertainty. Its ecosystem spans 20-plus countries on paper; how much of that stays reachable depends on trade rules well outside the company's control.

🎯 What Happens Next

Watch two numbers. First, whether that "20-plus customers" figure keeps climbing — a chip platform's value compounds with each maker that standardizes on it, because switching costs rise for everyone already locked in. Second, whether any of those 20 customers ship a robot that actually sells in volume. D-Robotics needs one breakout hit in its customer base to prove the whole thesis.

Expect the software half of the pitch to get louder. Chips are a margin business under constant price pressure; the development platform is where a supplier turns a one-time sale into a recurring relationship. The $400 million is, in large part, a bet that D-Robotics can become the operating system layer of robotics before a rival does.

🧩 Bigger Picture

Robotics funding has shifted this year from the robots themselves toward the layers beneath them — chips, simulation, tooling. It's the same pattern that played out in AI, where the companies selling compute and infrastructure captured value long before most model startups turned a profit. The interesting money has started moving one level down the stack.

D-Robotics is a clean example: a company that decided the safest seat in a hype cycle isn't the product everyone's watching, but the component nobody can build without. Twenty robot makers are now, in a small but real way, downstream of one supplier's roadmap.

In a boom, the loudest bets are on who builds the best robot. The quieter, harder-to-dislodge bet is on who every robot has to buy from.

Sources