Just witnessed something that really drives home how the robotics sector is reshaping the investment landscape. Lei Jun publicly thanked Wang Xingxing at a Xiaomi event for letting them invest in Unitree five years ago—and the very next day, Unitree's IPO application got formally accepted by the Shanghai Stock Exchange. The timing was almost too perfect, but it tells you everything about how this deal has been anticipated.

What makes this moment significant isn't just the IPO itself. It's that Unitree is breaking through as the first humanoid robot company hitting the A-share market with actual profitability. Most robotics startups are still burning cash trying to get prototypes working. Unitree? They've already sold 5,500 humanoid robots with a gross margin sitting at 62.9%. Their 2025 revenue is projected to hit 1.708 billion yuan—that's a 335% jump year-over-year—and net profit is expected to exceed 600 million yuan. The contrast is stark when you look at competitors still operating at a loss or managing margins below 30%.

The product evolution here is actually fascinating to track. Five years ago, Unitree was basically known as the robot dog company. Quadruped robots made up 76% of revenue back in 2022. Then they launched the H1 humanoid in August 2023, sold just 5 units that year. By 2024, the G1 entered mass production at 99,000 yuan, and suddenly they hit profitability. But the real inflection point was 2025. Those 16 H1 robots performing on CCTV's Spring Festival Gala changed the conversation overnight. Humanoid robots went from niche tech to cultural moment. In the first nine months of 2025, humanoid sales hit 595 million yuan—51% of total revenue—surpassing their quadruped business for the first time. They went from 5 units to 3,500 units in two years. The average price dropped from 593,400 yuan in 2023 to 167,600 yuan by mid-2025. Classic playbook: trade margin for scale, scale for data, data for faster iteration.

Now look at the shareholder list and you see why this IPO matters so much. Wang Xingxing holds controlling stakes, but the institutional side reads like a who's who of hard tech investment. Meituan's invested companies hold about 9.6%, Sequoia China around 7%, Matrix Partners about 5.5%. Then you've got Tencent, Alibaba, and Ant Group all in the cap table. BYD, Geely, China Mobile funds, state-owned enterprises—it's rare to see this kind of convergence around a single robotics company. The message is clear: everyone from pure financial investors to industrial players to state capital agrees this sector matters strategically.

What does being first really mean here? First, there's the scarcity premium. Hong Kong has some listed robotics companies, but A-share valuations for hard tech operate on completely different logic. The STAR Market liquidity and "domestic substitution" narrative make Unitree uniquely positioned. Second, Unitree's pricing will set the benchmark for over 20 other robotics companies waiting to go public—companies like Leju Robotics, Cloud Depth, Standard Robots, and others. If Unitree gets premium valuation, the whole sector rises. Third, this opens exit channels for early investors. The robotics funding frenzy has been insane—207 financing deals in the sector as of late March, with 115 involving humanoid robots. Star companies are raising faster and bigger. Zhiyuan hit 15 billion yuan valuation, Galaxy General just closed a 2.5 billion yuan Series B+ at 22.5 billion yuan post-money. Multiple rounds in under a year have become the norm for top players.

But let's be real about the risks. Embodied AI is still in research and testing globally. Unitree admits they haven't deployed their self-developed embodied large models at scale yet. They've open-sourced WMA and VLA to position themselves in future tech directions, but there's still a gap. They describe robot capabilities as having a "cerebellum" (motor control) and a "brain" (understanding, interaction, decision-making). Their cerebellum is world-class—they're already promising to outrun Usain Bolt's 100m time by mid-2026. But the brain tech isn't there yet. That's the bottleneck that'll determine whether robots move from labs to factories and homes.

Still, the market is pricing in these risks because everyone's betting on embodied intelligence as the ultimate AI frontier. Wang Xingxing wrote in the prospectus about being on the eve of a global breakthrough. Whether that's hype or vision, early investors in Unitree are about to find out when the stock starts trading. Lei Jun's thanks? That's probably the most expensive thank you in recent venture history.