JPMorgan Sees Unitree Robotics Riding the Physical AI S-Curve, Sets RMB995 Target Despite Margin Pressure
J.P. Morgan's Asia Pacific equity research team published a detailed third-quarter review of Unitree Robotics on October 29, 2026, maintaining an Overweight rating and setting a December 2027 price target of RMB995 — implying roughly 43% upside from the stock's last close of RMB695.90. The note arrives at a pivotal moment: Unitree's shares have already surged approximately 55% since October, leaving investors to debate whether the physical AI adoption curve has genuinely inflected or whether the rally has simply priced in expectations that the market has yet to validate.
Revenue Scaling Broad-Based, But Costs Are Rising Fast
The headline number is hard to dismiss. Unitree posted revenue growth of approximately 70% year-over-year in the third quarter of 2026, with year-to-date sales reaching approximately RMB1.8 billion (US$249 million). Volume growth is no longer concentrated in a single product line — it is spreading across both bipedal and quadrupedal platforms, small-sized and large-sized robots alike. The company's G1 humanoid has become the world's best-selling model, with cumulative shipments now in the 10,000–20,000 unit range, while four-legged platforms continue to hold global market leadership. In education and research, smaller quadruped models are gaining traction; in industrial and public service environments, larger variants are expanding their footprint.
What makes this growth strategically significant, according to J.P. Morgan, is that it is being driven by real-world deployments rather than pilots — a distinction that matters enormously for the data acquisition and feedback loops that define the long-term economics of physical AI platforms.
The Ecosystem Flywheel Is Spinning — At a Price
Unitree is not simply shipping hardware. The company is actively constructing an ecosystem designed to accelerate scenario diversity, edge-case generation, and model improvement. A recent ecosystem event drew over 3,000 participants, spanning city governance, emergency response, education, entertainment, and industrial applications. The company is funding global competitions — with prize money in the millions of RMB — covering athletics, fighting, dexterous manipulation, and healthcare, open to teams and individuals worldwide.
These are not marketing exercises. As J.P. Morgan frames it, the events are engineered to compress the simulation-to-deployment cycle, turning Unitree into the default partner for model companies, universities, and system integrators seeking real-world data touchpoints.
The cost of that ambition, however, is visible in the financials. R&D expenditure jumped approximately 252% year-over-year in Q3 2026, reaching roughly RMB130 million for the quarter and approximately RMB260 million year-to-date. Management explicitly linked the step-up not just to product development but to physical AI infrastructure and computing power — a signal, J.P. Morgan notes, that Unitree may be turning more constructive on the pace of progress in embodied intelligence.
Margin Pressure: A Feature, Not a Bug — For Now
Gross margins are under pressure from multiple directions: bill-of-materials inflation, competitive average selling price erosion, and surging investment in platform infrastructure. Operating cash flow for the first nine months of 2026 came in at approximately RMB190 million, down roughly 55% year-over-year, reflecting working capital absorption and accelerating investment in inventory, headcount, and technology. Capital expenditure tripled year-over-year to approximately RMB100 million over the same period, though J.P. Morgan notes this remains self-fundable.
Critically, Unitree remains profitable — a rare distinction among robotics OEMs globally at this stage of scaling. J.P. Morgan views this as a key differentiator and a meaningful floor under investor confidence, even as near-term margin headwinds persist.
Valuation and the S-Curve Debate
J.P. Morgan's RMB995 price target is anchored to 30x FY2028 estimated price-to-sales, benchmarked against the highest-quality global comparables. The thesis rests on the view that the physical AI sales S-curve is being pulled forward by rapid advances in large-model planning AI — shifting the sector's bottleneck from algorithmic capability to real-world deployment infrastructure, precisely where Unitree is most entrenched.
The bank also highlights an unusual structural tailwind: regulatory delays and higher IPO hurdles in China are extending Unitree's scarcity premium in capital markets, creating a window for value capture that competitors cannot easily replicate.
The bank names several component plays as complementary positions alongside Unitree: Leader Harmonious Drive System, Jiangsu Hengli Hydraulic, Zhejiang Sanhua Intelligent Controls, and Orbbec.
Downside risks are not trivial. Slower enterprise adoption, failure to convert pilots into repeat orders, intensifying competition, and geopolitical or regulatory setbacks — particularly constraints on international expansion — could each erode the premium the market is currently awarding.
The core question J.P. Morgan poses is not whether Unitree can grow, but whether its next phase will be defined by the quality and durability of its earnings rather than raw volume. For now, the market appears content to fund the ambition and wait for the "ChatGPT moment" of physical AI to arrive.
Related Coverage:
Unitree Robotics Q3 Revenue Jumps 69.93%, but Margin Squeeze Signals a Critical Inflection Point
Unitree Robotics' 466x Valuation Demands a Model Breakthrough It Has Yet to Prove