We didn't see this coming. A quadruped robot company, best known for mimicking a Boston Dynamics dog at a third of the price, just secured approval for a $619 million IPO on Shanghai's high-tech board. Unitree Robotics isn't a blockchain project, but its funding journey tells us everything about how centralization works in the real world—and why we need to build decentralized alternatives.
The approval, confirmed by the China Securities Regulatory Commission late last month, allows Unitree to raise up to $619 million through an initial public offering. The funds will be used to "expand AI robotics," a phrase so generic it could mean anything from hiring more engineers to building a new factory in Hangzhou. But for those of us who live in the crypto ecosystem, this event is a stark reminder of how capital flows when you don't have a token.
Context: The Robot That Walks Like a Dog, Costs Like a Car
Unitree is a Chinese robotics company founded in 2016 by Wang Chen, a former DJI engineer. Its flagship products include the Go1, a consumer-grade quadruped robot priced around $2,300, and the B2, an industrial version that competes directly with Boston Dynamics' Spot—at roughly one-third the cost. Last year, it unveiled the H1, a full-size humanoid robot intended for warehouse and manufacturing tasks, priced at $90,000.
The company has sold robots to universities, security firms, and even a few Hollywood studios for film shoots. But its revenue remains modest by IPO standards. Industry estimates suggest Unitree's annual revenue is below $100 million, making the $619 million raise a massive bet on future growth. For context, Boston Dynamics was acquired by Hyundai in 2020 for just $1.1 billion—only about double what Unitree is now raising in equity.
The approval itself was fast. In China, IPO reviews typically take six to twelve months. Unitree reportedly cleared the process in under six, a signal that regulators view robotics as a strategic industry worth fast-tracking. That speed also hints at government support, possibly through the "special new enterprises" (专精特新) program, which offers subsidies and tax breaks to companies deemed critical for technological self-sufficiency.
Core: The Centralization of Robotics—and What Crypto Can Learn
Here's where the story gets interesting for blockchain natives. Unitree's IPO is a textbook case of capital centralization. The company will issue shares, lock them in an exchange, and let institutional investors determine price discovery. Retail investors can participate, but only after the underwriters (likely a consortium of Chinese banks) have taken their cut. There is no token, no governance vote, no community treasury. The robots will remain the property of the corporation, and their intelligence will be controlled by a board of directors.
From my perspective—having spent nearly three decades watching both traditional finance and open-source ecosystems—this is exactly the kind of system we're trying to decentralize. Unitree's machines could one day replace human workers in dangerous environments like mines and power plants. But who decides how those robots are deployed? Who owns the data they collect? Who profits when a robot replaces a salary?
In a decentralized model, those robots could be governed by token holders. Their compute power could be rented out on-chain. Their training data could be contributed by users and rewarded with tokens. Instead, Unitree's IPO concentrates all that value into a single entity, subject to the whims of the stock market and geopolitical tensions.
The real insight here is not about Unitree itself—it's about the missed opportunity. We have the technology to create decentralized autonomous robot networks, where machines are owned by DAOs and operated by smart contracts. But capital markets still reward centralization. Until we build bridges between robotics and DeFi—perhaps through tokenized robotic compute or GPU-sharing protocols—we'll keep seeing these equity raises, each one pulling more value into silos.
Contrarian: Maybe the IPO Is Actually a Good Thing for Decentralization?
Before we write off Unitree's IPO as just another centralized capital event, let's play the contrarian. The $619 million will fund research that could make robots cheaper and more durable. Open-source software has already played a huge role in quadruped robotics—Unitree itself builds on the MIT Cheetah project and uses ROS (Robot Operating System). The company contributes code back to the community. A stronger Unitree means more resources for open-source robotics development.
Moreover, the IPO could accelerate the adoption of robotic labor in sectors where blockchain projects need physical infrastructure. Think about decentralized data centers that need automated maintenance robots, or autonomous vehicles for last-mile delivery in a tokenized economy. Unitree's industrial robots, if priced affordably enough, could become the commodity hardware that powers the physical side of Web3.
But there's a catch. Unitree's chips come from Nvidia's Jetson series, which is subject to U.S. export controls. The company has started using domestic alternatives from Horizon Robotics and Huawei, but those are still less powerful. Any disruption in the chip supply chain could halt production. And here, the decentralization community could step in—by funding open-source hardware designs that are chip-agnostic, or by creating decentralized compute networks that let robots run AI models on edge nodes rather than relying on proprietary hardware.
Takeaway: The Robot You Own vs. the Robot That Owns You
Unitree's IPO is a wake-up call for anyone who believes in the decentralization of technology. We are building decentralized finance, decentralized storage, and decentralized identity. But we have not yet begun to build decentralized robotics. The next wave of automation will be physical, and if we don't start now, the robots will be controlled by the same centralized entities that control the stock market.
The question I leave you with is not whether Unitree will succeed—it probably will. The question is: will we build an alternative where the robots serve the community, not the shareholder?
We didn't start the fire of industrial automation, but we can channel its heat into a decentralized furnace. The code is open. The opportunity is waiting. Let's not let it be captured by another round of equity.