China's Humanoid Robot Output Set to Surge 94% in 2026 as Unitree Files $610M IPO

The global humanoid robot race has a clear frontrunner, and it is not a Silicon Valley startup. China’s humanoid robotics ecosystem is accelerating at a pace that is rewriting industry forecasts, with research firm TrendForce projecting a 94% surge in annual output this year — and two companies at the center of it: Unitree Robotics and AgiBot.

The most striking signal of that momentum came in March, when Hangzhou-based Unitree Robotics filed for a $610 million initial public offering on Shanghai’s STAR Market. If approved, it would be the first public listing of a dedicated humanoid robotics company anywhere in the world — a milestone that would create the industry’s first traded benchmark for valuation and investor scrutiny.

Unitree’s Breakout Year

The IPO filing, formally accepted by the Shanghai Stock Exchange on March 20, 2026, reads like a case study in hypergrowth. Unitree’s operating income reached ¥1.708 billion ($236 million) in 2025, representing a 335% year-on-year increase. Net profit surged 674%, making 2025 the company’s first ever profitable year. The turnaround was driven by a fundamental product mix shift: humanoid robots, once secondary to the company’s popular quadruped “robot dogs,” climbed from 27.6% of revenue in 2024 to 51.5% in the first three quarters of 2025.

In terms of units, Unitree shipped more than 5,500 humanoid robots in 2025, capturing an estimated 32.4% of the global market — making it the world’s top seller by volume. The company’s flagship G1 humanoid, priced at a striking $16,000 base price, has become the benchmark for accessible robotics: standing 127 centimeters tall, weighing 35 kilograms, equipped with 23 degrees of freedom, 3D LiDAR, and depth cameras, with a two-hour battery runtime. Advanced EDU configurations with up to 43 degrees of freedom and an onboard NVIDIA Jetson Orin processor serve researchers and developers pushing the capability envelope.

CEO Wang Xingxing has forecast deliveries of 10,000 to 20,000 units in 2026, backed by plans to expand annual manufacturing capacity to 75,000 humanoids and 115,000 quadrupeds. Nearly half of the planned IPO proceeds will fund the development of what Wang calls “video-based world models” — embodied AI architectures that enable robots to learn new tasks by watching video of humans performing them, rather than requiring explicit motion programming.

AgiBot: The Factory-Scale Challenger

Unitree’s closest rival, Shanghai-based AgiBot, is pursuing a different but equally aggressive path. Rather than targeting consumer, research, and developer markets, AgiBot has anchored its strategy firmly in industrial deployment — primarily automotive manufacturing, consumer electronics assembly, and logistics operations.

In late March 2026, AgiBot rolled out its 10,000th cumulative unit of the Expedition A3 general-purpose humanoid, a milestone that was reached just three months after hitting 5,000 units. The company’s production scale jumped from roughly 1,000 units in 2025 to 5,000 and then to 10,000 within a single quarter — a ramp rate that suggests AgiBot has resolved many of the supply chain bottlenecks that constrained earlier humanoid manufacturers. Growing orders from Tier 1 automotive suppliers and consumer electronics contract manufacturers signal that adoption has shifted from proof-of-concept pilots to recurring industrial contracts.

Together, TrendForce projects that Unitree and AgiBot will capture approximately 78 to 80% of total global humanoid robot shipments in 2026 — a remarkable market concentration for an industry that, just two years ago, featured dozens of roughly equal competitors globally.

China’s Structural Manufacturing Advantage

The Chinese government has identified humanoid robotics as a strategic priority alongside AI, semiconductors, and electric vehicles. State-level support flows through subsidized manufacturing land, preferential financing for approved suppliers, and direct procurement from state-owned enterprises seeking to automate. Several Chinese cities, including Shanghai and Shenzhen, have created dedicated humanoid robot industrial parks with preferential lease terms and infrastructure subsidies.

But policy support is only part of the story. China’s existing manufacturing ecosystem gives humanoid robotics startups an unmatched advantage in production speed and cost reduction. The supply chains for actuators, harmonic drive gears, torque sensors, and structural components that enable humanoid robots are already embedded in the industrial districts around Shenzhen, Dongguan, and Shanghai — the same clusters that enabled China’s dominance in consumer electronics and electric vehicles. Iterating from prototype to mass production — a process that might take three to five years for a well-funded US startup — can happen in 12 to 18 months for a well-connected Chinese company with access to this ecosystem.

The contrast with US competitors is stark. Tesla’s Optimus program remains in limited trial deployment at its own facilities, with its ambition of sub-$30,000 consumer pricing still contingent on advances in Dojo supercomputer training that have lagged public timelines. Boston Dynamics’ electric Atlas, unveiled at CES 2026, has its entire 2026 production committed to early customers Hyundai and Google DeepMind, with broader commercial availability not expected until 2027 at the earliest.

A Commercialization Inflection Point

TrendForce’s April 2026 report marks a linguistic shift worth noting. Earlier forecasts described China’s humanoid robot sector in terms of “development” and “demonstration.” The new report uses the language of “commercialization” and explicitly identifies the second half of 2026 as a “critical phase” for the transition from capability showcase to scalable business.

Visible signals support this framing. A McDonald’s franchise in Shanghai recently began testing AgiBot units in customer-facing service roles — a deliberate publicity move, but also a genuinely novel operational deployment in a high-foot-traffic commercial setting. Logistics companies are signing multi-quarter contracts rather than single-site pilots. And China’s new national humanoid robot standards, released in March 2026, establish interoperability and safety requirements that were previously absent — the kind of regulatory scaffolding that a market builds when it is preparing for serious scale, not continued experimentation.

Unitree, meanwhile, is preparing to take the G1’s commercial ambitions global. The company has announced plans to debut its more affordable R1 humanoid model on Alibaba’s AliExpress platform, targeting North America, Europe, Japan, and Singapore — a direct challenge to Western incumbents on their home turf, at price points that no Western manufacturer is currently positioned to match.

What the IPO Would Mean

Unitree’s IPO, if completed on schedule in mid-2026, would be a landmark event for the industry beyond the fundraising itself. A publicly traded humanoid robotics pure-play would create a reference valuation for everything from venture funding rounds to cross-border acquisitions. It would expose Unitree’s financials to quarterly scrutiny and investor pressure in ways that privately funded competitors — including many US startups — currently avoid. And it would validate the asset class in the eyes of institutional capital that has been watching the sector with interest but waiting for liquidity.

For observers of the broader technology competition between the US and China, Unitree’s trajectory also carries a familiar resonance. A pattern established across EVs, solar panels, drones, and consumer electronics — Chinese manufacturers scaling production rapidly, pushing prices below levels that Western competitors can sustain, and capturing global market share before domestic rivals can respond — appears to be repeating itself in humanoid robotics.

The race is not theoretical. It has a leader, and that leader is in Hangzhou.