JPMorgan: Humanoid Robot Hardware Chain Priced In First, Highlighting Leader Harmonic Drive and Estun Automation
0xBroomberg
J.P. Morgan's latest report says the AI 'brain' for humanoid robots needs another 2–3 years, but the hardware chain is already entering its pricing window — the market is buying companies that can ship, cut costs, and scale, not waiting for AI to mature.
How far are humanoid robots from a breakout?
J.P. Morgan sees Physical AI's "ChatGPT moment" as still 2 to 3 years away, but the robot body — torso, joints, actuators — has made generational progress.
This means → the market won't wait for the AI brain to mature. It will price the hardware chain first: whoever can ship, cut costs, and scale gets valued now.
Global humanoid robot shipments are projected to reach roughly 1.75 million units by 2030, implying a 2025–2030 CAGR of about 150%. China is expected to contribute around 50% of global volume, and over 80% in 2025 alone.
Where are Chinese manufacturers on the lab-to-mass-production curve?
Chinese OEMs are moving from lab prototypes to scaled deployment in logistics, manufacturing, and data-collection scenarios. Industry shipments are forecast to jump from roughly 20,000 units in 2025 to 50,000–80,000+ in 2026.
Government-backed data-collection centers are a key demand source. In plain terms = the government is helping robots "stockpile training data," and that alone drives hardware procurement.
J.P. Morgan also raised its China industrial robot shipment forecast: 20% growth in 2026, above consensus. Output in the first five months of 2026 already rose 28% year-on-year. The report notes industrial robots are less "exciting" than humanoids but closer to real orders and profit statements — new demand comes from AI server assembly, optical and thermal management systems, PCB manufacturing, and overseas capacity expansion.
Unitree's IPO — does it help or hurt existing names?
Unitree (宇树科技) completed registration on June 2 and cleared the SSE STAR Market listing committee. CSRC approval typically takes 20 business days; the earliest listing date is mid-July. Consensus valuation is roughly RMB 100 billion, or about 35× FY26E price-to-sales.
J.P. Morgan sees two paths for the secondary market: short-term fund diversion — capital may rotate out of existing names like UBTech and Greenharmony; re-anchoring — a high IPO valuation could set a new reference point for the entire China humanoid and component chain.
This reflects the industry shifting from "thematic diffusion" to rising concentration: capital is flowing toward the few platforms that can scale, while smaller OEMs face mounting capex pressure from VLA training, data collection, and manufacturing ramp-up — fundraising is getting harder.
Greenharmony — what risk hides behind the high growth rate?
J.P. Morgan projects Greenharmony — China's leading harmonic reducer maker — to deliver a FY26–FY28 sales CAGR of 50%: traditional industrial robot applications grow at roughly 20% annually, while humanoid robot applications, driven by clients like Aubo and UBTech, are expected to see volume CAGR of about 140%.
The catch: the company may adopt aggressive pricing this year to grab share, with prices falling 8 percentage points year-on-year before normalizing to 3 percentage points of annual discount in 2027–2028. Put simply = cut prices now to win volume, betting that scale economics will protect margins once volume ramps.
Target price is RMB 481, implying a FY27E P/E of 222×, with DCF assumptions of 8.7% WACC and 6.5% terminal growth. This means → the valuation already prices in high growth; the margin of safety for investors is narrow.
Estun — where is the expectations gap?
J.P. Morgan defines Estun as China's leading servo system and industrial robot maker. Its edge is vertical integration: CNC systems, motion controllers, servos, robot vision, and small reducers are all produced in-house. The company also enters humanoid robots through its subsidiary Estun Codroid.
The earnings recovery path is concrete: a 2026 target of RMB 6 billion in revenue and 5% net margin, with plans to lift gross margin to 36% and net margin to 12% by 2030, improving 1–1.5 percentage points per year. Supporting factors include a one-time receivables and inventory write-down completed in 2024, and a No. 1 market share in China (10%) as of Q1 2026 — with share still rising in higher-end, better-margin segments.
This means → unlike Greenharmony, whose story rests on a "volume explosion," Estun's thesis is about margin recovery — the expectations gap is whether the market believes this recovery path will deliver.
How does J.P. Morgan's full framework fit together?
The core logic: humanoid robots raise the valuation ceiling; industrial robots reduce earnings-delivery uncertainty. Only when both curves are revised upward together does the sector offer a complete, tradeable framework.
In plain terms = humanoid upside alone isn't enough — you also need real industrial-robot orders as a floor. Both legs must move together.
Verification window: Greenharmony's quarterly sales data and Estun's margin-delivery cadence are the two most critical data points to watch next.
Content is for reference only, not financial advice.