There is a moment in every robotics IPO arc where the prospectus has to stop being a vision document and start being an income statement. Unitree’s moment arrived on May 26.
The Hangzhou-based humanoid maker disclosed unaudited Q1 2026 financials to Caixin and the South China Morning Post six days before its Shanghai Stock Exchange listing-committee hearing. Revenue: 422.8 million yuan, up 68.4% year on year. Adjusted net profit excluding non-recurring items: 40.3 million yuan, down 52% from 84.8 million yuan. R&D and sales expenses are the named pressure points. The Shanghai Stock Exchange’s STAR Market committee reviews the application on June 1, with the company seeking to raise 4.2 billion yuan (≈$618.9 million) at a target valuation of roughly $6.2 billion. Application to hearing: a 73-day sprint.
The two numbers the committee has to reconcile
The top line is what the cap table bought. Revenue up 68% in a quarter where the broader humanoid sector is, as Unitree itself characterised in the Caixin filing, “slowing momentum in the broader humanoid robotics hype cycle” — a remarkable thing to write in your own prospectus update if you are the company most associated with that hype cycle. The H1 guide is 1.05–1.13 billion yuan, up 35.6% to 45.4% YoY. That is the line the committee will treat as the floor of the equity story.
The bottom line is what the income statement is doing while the cap table is being sold. Adjusted net profit halved against a year-ago base of 84.8 million yuan. R&D and sales expenses are surging — that is the company’s own framing, not an analyst’s. The implication is that the unit economics that powered the 2025 surge (Unitree was one of the very few humanoid makers operating at scale and at a profit) are being deliberately re-leveraged into next-generation product and channel build-out. Whether that re-leveraging is a one-quarter dip or a structural margin reset is exactly the question the listing committee has to score on June 1.
The comp on the cap table is heavy. The backers on the strategic round include China Mobile, Tencent, Alibaba, Ant Group, and Geely Capital — a slate that reads less like a venture round and more like a state-blessed industrial consortium. Founder Wang Xingxing retains majority control. The political case for listing — keep a Chinese humanoid leader’s float on a Chinese board, fund the manufacturing build-out on Chinese capital — is the strongest argument the committee will hear on Monday. The financial case has to share the same room as the 52% number.
What the unit-economics question actually looks like
Unitree’s installed base story is the cleanest in the sector. Roughly 5,500 units shipped in 2025, with a 2026 guide of 10,000 to 20,000 units. The Q1 revenue print implies an average revenue per unit in the low five-digit dollar range, which lines up with the Unitree R1 AliExpress listing at $8,150 and the JAL Haneda trial at approximately $15,400 per unit for the Unitree-based platform GMO AI & Robotics deployed in May. The mass-market story — a sub-$20K humanoid that can ship in five-digit volumes — is the one the public-market pitch depends on.
The compressed profit print is the cost of holding that price line. Surging R&D buys the H2 generation and the next dexterity stack. Surging sales expense buys the global distribution layer — AliExpress is the consumer channel, JAL is the enterprise reference, and a half-dozen more enterprise pilots are in flight. None of those investments mature inside a single quarter. The S-1 has to argue that they mature within four to six quarters of listing.
If the committee accepts that argument, the prospectus reads as a deliberate margin reset to defend a category-leading position. If the committee does not, the prospectus reads as a top-line growth story already absorbing operating-leverage pressure ahead of public scrutiny. The valuation gap between the two readings is roughly the entire $6.2B sticker.
The China humanoid context
TrendForce’s April projection is the macro frame. China humanoid output is expected to grow 94% in 2026, with Unitree and AgiBot capturing roughly 80% of total shipments combined. That is the duopoly the STAR Market listing is meant to anchor. AgiBot is not yet public; if Unitree clears June 1, the next obvious move on the calendar is AgiBot’s own prospectus.
The competitive geography is also legible. Xpeng’s Iron has the consumer-tech buzz. UBTech and Leju have the older Chinese enterprise relationships. Boston Dynamics is locked into Hyundai’s 25,000-unit commitment and School of Football brand-ambassador campaign. Figure and 1X are the US plays, with 1X’s Hayward, California factory commencing production on April 30 at a 10,000-unit annual capacity. Unitree is the only one of them filing into a domestic public-market backstop right now.
What to watch
- June 1 committee outcome. Acceptance, deferral, or conditional approval. Deferral is the cleanest signal that the Q1 print outweighed the cap-table backing.
- The audited Q1 number. The 422.8M / 40.3M figures are unaudited. The audited print and the audit firm’s emphasis paragraphs will sit in the final prospectus. Discrepancy of more than a few percentage points is a re-pricing event.
- H2 product launch cadence. Unitree’s pitch for the surging R&D line is the next-generation product. Whatever follows the R1 and H1 has to land before year-end for the margin story to hold.
- AgiBot’s filing window. If Unitree clears June 1, watch the back half of 2026 for AgiBot’s STAR Market application. If Unitree gets deferred, AgiBot’s calendar slips with it.
- Pricing on the strategic comparable. Tencent, Alibaba, Ant, and China Mobile are on Unitree’s cap table; some of them are also on AgiBot’s. Cross-comp pricing on the secondary market after listing will tell the market what the consortium thinks the Chinese humanoid leader is actually worth.