Unitree IPO Signals Real Hardware, Humanoid Yet Early

Unitree Robotics’ plan to raise about $610 million on China’s STAR Market signals a rare win for real hardware in robotics, even if the coveted humanoid promise remains a work in progress.

The Shanghai filing is notable less for the hype around walking bots than for the concrete numbers behind a hardware business. Production data shows a company that can scale a robot platform while squeezing costs. In 2025, Unitree reportedly pulled in about $248 million in revenue, a credible top line for a hardware maker chasing global expansion. More telling is the price dynamic: the average selling price reportedly declined from roughly $85,000 in 2023 to about $25,000 in the first nine months of 2025. Yet companywide gross margin rose to 59.8% over roughly the same period, a combination that undercuts the usual story of margin compression when prices fall.

From a deployment perspective, those numbers matter. A rising gross margin at lower ASPs points to cost leverage—more efficient component sourcing, better automation in manufacturing, or a more favorable product mix that emphasizes scalable, high-volume platforms over bespoke systems. In the robotics community, that is the signal investors tend to trust: a hardware business that can grow throughput and reduce unit cost as volumes rise, rather than a high-price, low-output lab prototype.

Even with the price pressure, the humanoid angle remains the unknown. The IPO filing underscores that Unitree has built a credible, mass-market hardware business, but the bigger bet—humanoid robots solving broad industrial tasks—has yet to prove a sustainable, profitable model at scale. The company’s data suggests a healthy cash-generating core, but the path to widespread humanoid deployment hinges on reliability, long-term maintenance economics, and integration into mixed-task environments where real work happens in factories and warehouses.

For practitioners watching the IPO, several implications stand out. First, hardware leverage is real and visible in margins, but risk remains in the transition from prototype to production-grade systems. Second, the price decline implies customers are sensitive to total cost of ownership, including integration, maintenance, and uptime—factors not captured by a simple sticker price. Third, the market will likely reward companies that can pair affordable hardware with robust software, sensors, and service ecosystems that deliver measurable ROI, not just clever demos.

Two concrete takeaways for operators and engineers: expect a tighter focus on total cost of ownership, not just upfront price; and watch how vendors bundle training, spare parts, and uptime guarantees into lifecycle contracts, because those elements often determine payback in real deployments. The Unitree story suggests that hardware fundamentals—volume, cost discipline, and margin resilience—are finally getting the attention they deserve in a field historically driven by aspirational capabilities.

If the IPO proceeds as the filing suggests, investors will be looking for a durable, scalable hardware platform that can support both affordable robots and, eventually, more ambitious humanoid ambitions. For facilities evaluating automation investments, the lesson is clear: hardware with a credible path to lower unit cost and higher uptime can attract capital and scale—while the humanoid promise still requires careful scrutiny of integration, reliability, and ROI.