Big news for A-shares! Unitree rushes toward IPO, the biggest mystery has arrived

Special Topic: A-shares’ Core Foundations Remain Unshaken; Investment Focus on Prosperity and Dividends

Stock trading insights with Golden Kylin analyst reports—authoritative, professional, timely, comprehensive—helping you uncover potential thematic opportunities!

The market’s focus is undergoing a fundamental shift: from “whether we can develop cool robots” to “whether we can sell profitable robots.”

Author: Zhao Xinliang

In March 2026, Yushu Technology submitted an IPO application to the STAR Market, planning to raise 4.2 billion yuan. As the world’s leading supplier of humanoid robots in terms of shipment volume, the company demonstrated its commercialization capability with impressive financial data: revenue of 1.708 billion yuan in 2025, net profit excluding non-recurring gains of 600 million yuan, and a gross margin of 60.27%.

The numbers alone are enough to excite the capital market. But what truly sparks debate is: how should a company, supported mainly by scientific research, education, and government orders, be valued?

As it attempts to shift from “selling tools” in the scientific research market to “selling productivity” in the industrial market, can the market continue to price it using the same logic?

High Gross Margins from Scientific Research and Education

Breaking down Yushu’s revenue structure in 2025 reveals an interesting phenomenon: revenue from humanoid robots surpassed that of quadruped robots for the first time, reaching 51.53%. However, 73.6% of this came from scientific research and education, with only 9% from industrial applications.

Source: Yushu Technology Prospectus

This indicates that Yushu’s current primary role is selling high-end research platforms to universities, research institutes, and tech companies, rather than selling productivity tools that replace labor in factories.

This positioning is reasonable. Scientific research clients are less sensitive to price, with their robot procurement budgets coming from research funds, and they value platform openness and motion performance. Yushu has reduced joint motor costs to below 50% of imported products, and harmonic reducers to one-third of similar Japanese products, through its full-stack self-developed core components—motors, reducers, controllers. In 2025, the overall cost of its core components decreased by 30-60% compared to overseas procurement, boosting gross margins above 60%.

This “self-reliant technology + supply chain collaboration” model creates a strong moat in the scientific research and education market. Yushu achieved a net operating cash flow of 672 million yuan in 2025, exceeding its net profit excluding non-recurring items, indicating high-quality cash inflows.

This explains why the capital market is willing to assign it a high valuation: in an uncertain innovation track, Yushu has established a clear and proven business model.

But the question is: where is the ceiling of this model? The capacity of the scientific research and education market is limited. If the company cannot successfully shift its focus to industrial and commercial scenarios with greater scale potential, its current high growth may soon plateau.

Three Valuation Transitions in New Tracks

The valuation logic for innovative tracks is never static. Looking back at the development of new energy vehicles, three stages are evident:

• Around 2010, the industry was in a technology validation phase, with market valuation driven by the narrative of “disrupting the auto industry.” Tesla’s valuation was detached from financial data, driven instead by imagination about its technological route and long-term potential.

• From 2015 to 2018, Tesla experienced “production hell,” with valuation anchored to sales growth and capacity ramp-up. The market no longer bought into the story but focused on quarterly delivery figures and gross margin changes.

• After 2020, Tesla achieved consistent profitability, and valuation shifted again: the market began evaluating its long-term cash flows using DCF models, and its automotive sales, FSD software subscriptions, and energy business were broken down to assess their independent values.

The smartphone industry experienced similar evolution. From 2007 to 2010, the iPhone was priced as a “disruptor”; from 2011 to 2015, Android manufacturers competed fiercely, with valuations reverting to manufacturing logic—PE ratios generally between 10 and 15. Post-2016, the industry entered a stock-competition phase, where only companies with ecosystem and brand premiums (like Apple) maintained high valuations, while pure hardware assemblers faced continued pressure.

These historical lessons point to a common pattern: the valuation of innovation tracks inevitably shifts from “expectation-driven” to “cash flow-driven.” The timing of this shift depends on three conditions—whether the technology is mature, costs are acceptable, and the business model is validated.

Currently, the humanoid robot industry is transitioning from the “technology validation” phase to the “scenario penetration” phase. This means the market’s focus is fundamentally changing: from “can we make cool robots” to “can we sell robots that make money.”

Source: Yushu Technology Official Website

“Selling productivity” is the real business

For Yushu to move from the scientific research market into the industrial market, it must overcome three hurdles.

First hurdle: technology. Yushu has become relatively mature in motion control and “small brain” capabilities, capable of fast walking, running, and even flips. But industrial clients need more than performance; they require “big brain” capabilities—autonomous recognition of parts, path planning, and assembly in noisy factory environments. Environmental perception, semantic understanding, and task planning remain technical bottlenecks. Yushu’s prospectus also admits that its self-developed UnifoLM large model is still in R&D testing.

Second hurdle: cost. Industrial clients’ procurement decisions are strictly based on ROI. A humanoid robot must recover its investment within 2-3 years by replacing human labor. Currently, single-unit costs often run into hundreds of thousands of yuan, making this difficult in most scenarios. Yushu has achieved the industry’s lowest unit costs through self-research, but still has a gap compared to what industrial clients find acceptable.

Third hurdle: scenario recognition. Even if technology and costs meet standards, a key question remains: in which scenarios are humanoid robots the “best solution”? Many tasks can be performed more cheaply and efficiently by non-humanoid robots (like robotic arms, AGVs) or humans. Therefore, Yushu needs to identify “killer scenarios” where only humanoids can do the job—this requires extensive field testing and customer collaboration, and cannot be achieved overnight.

These three hurdles determine that Yushu’s transition from “craftsman mode” to “worker mode” will be a “risky leap.”

Why is the market willing to pay?

Understanding Yushu’s current situation and challenges, let’s examine how the capital market is pricing it.

Currently, the robotics industry as a whole is in a “low-price, not low valuation” state. From 2024 to 2025, the humanoid robot sector index declined for two consecutive years, yet the sector’s rolling PE (TTM) remains high at 31-36 times. This divergence reflects the market’s optimism mixed with risk concerns.

The high valuation assigned to Yushu actually contains two parts: one is the certainty of its existing scientific research and education business, and the other is the “option value” of its future industrial market success.

The question is: is this “option value” priced reasonably? Historical experience shows that during the “production hell” phase of new energy vehicle companies, the market validated their options through sales and gross margins. Currently, Yushu’s challenge is not to produce more impressive flips but to secure non-affiliated, repeatable industrial orders with clear ROI.

In 2025, only 9% of Yushu’s revenue came from industrial applications. If in the next 2-3 years this rises above 20%, it would indicate initial success in its commercial transformation, and valuation logic could shift to growth stocks. Conversely, if industrial breakthroughs are delayed, the “option value” embedded in current valuation will gradually be stripped away.

Imminent Segregation and Reshuffling

In the next 2-3 years, the humanoid robot industry will enter a brutal “truth-seeking” phase. Market focus will shift from broad industry narratives to trackable, verifiable micro-indicators.

Order quality will become the primary screening criterion. Investors will distinguish strictly between framework agreements and firm contracts. Companies relying on “strategic cooperation” announcements without actual revenue realization will face valuation corrections.

Capacity utilization will determine survival. Current industry plans for capacity far exceed medium-term demand. Companies with persistently low utilization rates will be crushed by depreciation costs.

Technical reliability will become more important than flashy technology. The market will no longer pay for lab flips but will focus on whether robots can operate continuously and reliably for hundreds of hours in real industrial environments.

This means the industry will accelerate differentiation. Companies with full-stack capabilities, validated benchmark scenarios, and healthy cash flow will gain more trial opportunities and order resources. Those relying solely on financing, lacking breakthroughs, and unable to secure substantial orders will face valuation crashes and survival risks.

Valuation Requires a “Risky Leap”

Returning to the initial question: how should Yushu Technology be valued?

The answer depends on which stage investors believe it is in. If viewed as a scientific research equipment company, valuation should follow logic similar to high-end instruments or military information companies. If believed to successfully transform into an industrial productivity tool provider, future industrial orders’ expectations can be used for valuation. If further believed to become a consumer platform, more distant imagination can be incorporated.

But one thing is certain: the valuation of innovation tracks will inevitably shift from “expectation-driven” to “cash flow-driven.” The timing of this shift depends on actual progress in technology, costs, and business models, not just visions.

Yushu Technology has demonstrated its value in the scientific research and education market. The real test is whether it can make the risky leap from “selling tools” to “selling productivity.” This is not only Yushu’s story but also a question that the humanoid robot industry must answer in the next 2-3 years.

For investors, the greatest opportunity does not lie with the companies that tell the best stories but with those that can continuously convert technology into customer value and prove their business models with financial data. Between rationality and vision, a cautious balance may be the best approach.

Editor: Du Yan Proofreader: Qiao Yi Review: Xu Wen