To The Moon

In late July, during a layover in Zhengzhou late at night, I wrote about waiting for Yushu's "robot bird" to take off. On the evening of August 6, the IPO price was finalized at 150.80 yuan per share, with subscriptions imminent. This is not just an IPO pricing; it establishes the first valuation anchor for the A-share embodied intelligence sector. The anchor's significance lies not in whether it is cheap or expensive, but in that, from now on, all similar companies will have a reference point for financing, mergers and acquisitions, and share reductions. However, this anchor also serves as a stress test. At a price-to-earnings ratio of 219 times, investors are not buying into the 2026 financial report, but rather the "endgame of 2030." Yushu Technology's non-recurring net profit in the first quarter of 2026 has already declined by 52.55% year-over-year, with research and sales expenses heavily eating into hardware-scale profits. The day the anchor is set marks the market's shift from "believing in the story" to "quarterly verification." Embodied intelligence never lacks imagination; what it lacks are those who can deliver solid results every quarter.

The pricing reveals a nuanced strategy. A close look at the offline inquiry data shows the institutions' psychological price levels clearly. The weighted average and median of all investor bids were 152.15 yuan and 152.39 yuan, respectively, with the "lower of the four numbers" set at 152.1538 yuan. This indicates that 152 yuan was the consensus in the primary market. However, the final issuance price was set at 150.80 yuan, not touching the upper limit. This 2.35 yuan difference reflects the restraint shown by the issuer and its underwriter, CITIC Securities, amidst the frenzy. Facing a 219 times issuance P/E ratio, this "leaving room" pricing strategy aims to cope with potential early volatility post-listing and to test the secondary market's capacity. The 75,400 yuan subscription threshold per lot itself acts as a filter for risk.

The list of strategic investors reveals deep industry capital involvement. Within the 20% strategic placement, the most notable is not merely state-owned capital, but the presence of Hangzhou DeepSeek. This large language model company was allocated 933,400 shares, amounting to approximately 141 million yuan, with a lock-up period of 36 months. In the current environment, such a deep binding between a large model company and a robotics hardware company is extremely rare. DeepSeek's entry is not solely for arbitraging the gap between primary and secondary markets but to ensure its large model technology is first deployed on Yushu's hardware platform. Combined with CITIC Securities' subsidiary following the investment with a 24-month lock-up and key employee asset management plans locked for up to 36 months, this list forms a close-knit interest community. However, this also raises new questions: when the lock-up period expires in three years, if performance falls short, will these "industry capitals" hold firm or rush to exit to recover funds?

Valuation is always the core issue. Yushu's issuance price-to-sales ratio stands at a high 35.89 times, while the average PS ratio for Hong Kong-listed peers like UBTECH and Yuejiang Technology is only 19.74 times. More striking is the P/E ratio of 219.23 times. Peers like UBTECH are still loss-making, with the market pricing in future prospects; Yushu, having already achieved a non-recurring net profit of 591 million yuan, is given an even higher premium. This creates a peculiar "profit paradox": typically, profitable companies should have lower valuations than loss-making ones, but Yushu breaks this rule. This implies extremely high expectations for Yushu's growth rate. However, Yushu's financial concerns cannot be ignored: the average price of humanoid robots has dropped from nearly 600,000 yuan in 2023 to 166,400 yuan in 2025. The "trading volume for price" strategy has secured the global top spot in shipments but has severely eroded profit margins. If high-margin scenarios like industrial inspection and home services cannot scale up quickly, relying solely on low-price competition in the scientific research and education market will be difficult to sustain a 35.89 times PS ratio.

Originally planning to raise 4.202 billion yuan, the final net proceeds amount to 5.917 billion yuan, an oversubscription of nearly 1.7 billion yuan. This huge capital, coupled with the 2.022 billion yuan allocated to embodied intelligence model R&D, forms Yushu's technological moat. But having too much money is not entirely a blessing. On one hand, mismanagement of large funds can lead to inefficient investments. On the other hand, capital market patience is limited. With 5.9 billion yuan in cash, the market's expectations for Yushu's performance growth have been instantly raised to the maximum. In the past, Yushu could focus on refining its technology; now, it must face quarterly financial report pressures, balancing high R&D investment while maintaining gross margin stability. The oversubscribed 1.7 billion yuan is both fuel for R&D and a sword of Damocles, forcing Yushu to transform its technological advantages into irreplaceable commercial barriers in the shortest possible time.

Setting aside the noise around the issuance pricing and future first-day gains, what matters more is whether, three or five years later, when DeepSeek's 36-month lock-up period expires, Yushu can deliver matching performance results. The oversubscribed 1.7 billion yuan should not become "sleeping assets" on the books but should be forged into a sharp blade to conquer core embodied intelligence algorithms. An IPO is just a coming-of-age ceremony; production and operations are the long runway. I hope Yushu, while enjoying the capital feast, remembers the plane that landed at the airport at 2 AM. The reason it could take off was not a runway paved by valuations, but the real fuel in its engine. This is for reference only, not investment advice. Entering the market carries risks.