When Unitree and Sequoia China jointly announced the “Shanyu·Create Initiative” on October 28, 2026, the news was easy to file under “another corporate VC launch.” That would be a mistake. Unitree and Sequoia China are not strangers experimenting with a side project—they are the founder and the earliest institutional backer of one of China’s most closely watched robotics companies, now teaming up to fund the next generation of embodied intelligence startups.
The timing is what makes the Unitree and Sequoia China partnership worth pondering. Unitree had just gone public on the STAR Market in August, and its first quarterly report since listing showed revenue climbing while profit fell—a company spending heavily on R&D and now reaching outward for innovation. Sequoia China, which first invested in Unitree’s Pre-A round in 2019, is doubling down on a relationship that began when Wang Xingxing was still carrying a robot dog onto a green-skinned train to make it to an investment meeting in Beijing.
Among the five principles the Unitree and Sequoia China initiative officially stated, one stands out: “non-interference.” For Wang Xingxing, who spent seven years going from carrying a robot dog onto a green-skinned train to raise money, to now standing alongside his investor as “the one giving the money,” the posture he has chosen is to be a co-pilot, not to grab the steering wheel. The move itself is more worth pondering than the financing.
A “Identity Transformation” Seven Years in the Making
The story begins in 2019. That summer, Li Yannan, a managing director at Sequoia China, found Wang Xingxing’s QQ email through a Zhejiang University alumnus and traveled to Hangzhou for a visit. At the time, Unitree had been founded for just three years, and quadruped robots were still a niche track. Wang Xingxing demonstrated an AlienGo, but within Sequoia there was no consensus on the investment what exactly are robot dogs for? Do factories really need them? No one could give a definitive answer. To make it to the investment decision meeting in Beijing, Wang Xingxing carried the robot dog on a green-skinned train for more than ten hours, because the battery capacity exceeded the limits for high-speed rail and air travel.
Seven years later, Unitree went public on the STAR Market on August 19, 2026, as the “first humanoid robot stock,” with its market value briefly exceeding 440 billion yuan on the first day. Sequoia became one of its earliest and most important backers, holding a 7.11% stake before the IPO, making it the largest independent private equity fund shareholder in Unitree. The birth of the Shanyu·Create Initiative, in Wang Xingxing’s words, is “a deep cooperation and natural extension based on a high degree of mutual trust.” But the deeper signal lies in this: industrial players are beginning to use capital to build ecosystems, rather than fighting alone.
“Engineer + Investor”: A New Startup Acceleration Model
The core methodology of the Shanyu·Create Initiative is summarized as the dual perspective of “engineer + investor.” Unitree provides engineering capabilities and industry insight officially described as “understanding technology and industry” to help startups avoid the trap of “feasible in the lab but not viable in mass production and the market.” Sequoia provides capital and post-investment capabilities, with fast decisions and early discovery. The commitments at the attitude level are “valuing people” and “non-interference”: focusing on young entrepreneurs, maintaining high-trust, low-friction communication, “giving methods, resources, and advice, and being a good co-pilot.”
The underlying logic of this combination is not complicated. The barriers to entry for embodied intelligence startups are rising sharply pure financial investment can hardly solve the “dirty and hard work” of hardware mass production, supply chain management, and scenario validation. The engineering experience Unitree has accumulated in quadruped and humanoid robots over the past few years is precisely the resource that early-stage teams lack most. Unitree is not alone: this year, peers such as Agibot, Galbot, and Galaxea have all launched external investment platforms, with Galaxea partnering with Cathay Capital on a 1 billion yuan, five-year early-stage investment plan.
The “Other Side” in the Earnings Report: Why Do Investment Now
On the same day the Shanyu plan was announced, Unitree released its first quarterly report since going public. The data presents a contradictory picture: revenue for the first three quarters was 1.817 billion yuan, up 55.71% year-on-year; net profit attributable to shareholders was 387 million yuan, up 270.54%. But looking only at the third quarter, revenue was 665 million yuan, up 69.93% year-on-year, while net profit attributable to shareholders was 113 million yuan, down 17.37% year-on-year.
Behind the revenue growth without corresponding profit growth is a surge in R&D investment. Third-quarter R&D expenses were approximately 127 million yuan, a year-on-year increase of about 249.58%; the R&D expense ratio reached 19.10%, up 9.82 percentage points from approximately 9.28% in the same period last year. The company explained that this was mainly due to increased investment in embodied intelligence large models, especially model R&D and computing power.
This set of data explains another motivation for Unitree’s investment: when its own R&D investment is already large enough, investing in early-stage teams to leverage external innovation is a more efficient way to expand the ecosystem. At the World Physical AI Conference on October 22, Wang Xingxing revealed that the “Yuchuang Plan” had received more than 200 projects in just over a month, and he hopes to incubate more than 200 ecosystem companies over the next five years.
What It Means for Entrepreneurs
For early-stage teams in embodied intelligence and hard tech, the Shanyu plan sends several signals worth watching. First, industrial capital is replacing pure financial capital as the first choice for early-stage financing. When a leading company like Unitree is willing to open up its engineering capabilities and mass production resources, what early teams gain is not just money, but a shortcut that shortens the distance from “prototype to product.” Second, the promise of “non-interference” is itself a scarce commodity. Hard tech startups have long cycles and uncertain technology paths, and what founders fear most is investors interfering in decisions without understanding the technology. Third, the era of ecosystem competition has arrived. When leading companies begin to use investment to build moats, the choices for independent entrepreneurs become clearer: either integrate into an ecosystem, or go deep and fast enough in a segmented scenario.
It is worth noting that the fund size, contribution method, investment entity, and specific decision-making mechanism of the Shanyu Initiative have not yet been disclosed; the current framework is greater than the actual scale. In addition, early teams accepting industrial capital from Unitree also need to consider whether they may be viewed as “taking sides,” which could affect cooperation with other robot manufacturers. The promise of “non-interference” sounds beautiful, but the boundary between ecosystem synergy and independence needs to be navigated by the founders themselves.
When Wang Xingxing took the green-skinned train to raise money seven years ago, he probably never imagined that one day he would stand on the investor’s side. But the logic of this path is coherent: he understands better than most investors what a robotics startup team truly needs on the road from zero to one, and what it truly does not need. The three words “non-interference” may be harder to achieve than any investment term.
