Has it been replaced by the highest-valued robotics company?

After leading companies successively entered the “Hundred Billion Valuation Club,” embodied intelligence enterprises are beginning to pursue new goals.

On April 2nd, Xinghaitu announced it had secured 2 billion yuan in Series B+ funding, with its latest valuation surpassing 20 billion yuan, making it the highest-valued embodied intelligence company in the primary market.

The company was founded in September 2023 and positions itself as a foundational company for embodied intelligence. Its team has a strong academic background from Tsinghua University and experience in the autonomous driving industry.

In addition to Xinghaitu, recent announcements include companies such as Zhihui Square, Qianxun Intelligent, Variable, Star Epoch, and Pasini, all of which have valuations exceeding 10 billion yuan. Including earlier companies that broke the 10 billion mark like Galaxy General, Zhiyuan Robotics, and Yushu Technology, the “Hundred Billion Valuation Club” now has at least nine members.

Xinghaitu, which completed this round of financing, has a diverse shareholder lineup, including industry capital, long-term funds, and local state-owned assets. Notably, Xinghaitu is a key partner of lens technology Lansi Tech, an industrial shareholder, and the two will collaborate deeply in hardware supply chains and large-scale mass production, accelerating product penetration across all scenarios.

Looking back, in February this year, Xinghaitu’s Series B valuation first reached 10 billion yuan. When asked why the valuation could double in such a short period, CFO Luo Tianqi summarized three reasons in interviews with media such as Shanghai Securities News:

First, Xinghaitu has been “aggressively” increasing R&D investment since the end of 2025. He revealed that R&D expenses over the past six months are several times the company’s total since its founding, aiming to “shift spending from efficiency to effectiveness” by 2026. This clear strategic shift from “reserving” to “advancing” and the corresponding R&D progress have attracted the attention of the primary market.

Second, at the model level, Xinghaitu’s open-source G0 VLA large model in August 2025 surpassed the then state-of-the-art; in January 2026, it open-sourced the world’s first ready-to-use VLA model G0 Plus; in February 2026, it released the G0 VLA model tailored for clothing folding and the G0 Tiny lightweight model supporting edge deployment—demonstrating the company’s scale, industry position, and talent density to the market.

Third, the restructuring of the industry valuation system and the scarcity of “authentic brain” targets. Since the listing of Zhipu and MiniMax earlier this year, the secondary market has assigned very high valuations to AI large model companies. This valuation logic has quickly propagated to the primary market, prompting a reassessment of embodied large model targets. “Investors are viewing the industry with a longer-term perspective, believing that the next few years could see annual growth rates of several tens of percent or even 100%. They can quickly digest the current high valuations.”

Regarding the 2026 goals, Xinghaitu has two main plans. First, to develop foundational models, which are the starting point and a key variable in the embodied intelligence revolution. Xinghaitu revealed that its upcoming world model, Fast-WAM, has already shown significant progress in internal evaluations.

On the data side of the model, unlike some industry companies that use simulation or training ground modes, Xinghaitu has insisted on collecting data from real-world scenarios since its inception. The company states that this choice has saved many detours, as real-world data is more effective for training models that need to operate in actual environments.

Second, to “run through” a very important productivity scenario.

“‘Running through’ doesn’t mean having one or two units, making a video, and doing a POC (proof of concept), but rather that this product actually performs well in its designated role, and from a cost perspective, it can be clearly calculated, enabling a transition from 1 to 10, and from 10 to 100, in a replicable stage,” Luo Tianqi explained.

In addition to strengthening its own business, Xinghaitu also plans to participate in building the industry ecosystem through investments. Industry insiders believe that embodied intelligence technology will not converge within the next two or three years. This means various technical routes and algorithm models are still rapidly evolving and have not yet stabilized.

Therefore, Xinghaitu is leading the formation of an early-stage investment fund focused on downstream scenario application companies and frontier technology firms.

Author: Sun Xiaocheng

Massive information, precise analysis, all on Sina Finance APP