Gasgoo Munich- M&A activity in the humanoid robot sector has surged since the second half of 2025.
At home, AgiBot has taken control of Shangwei New Material, while UBTECH acquired a controlling stake in Fenglong Co. for 1.665 billion yuan, and Hangzhou Colin moved to control Kepler Robotics. Abroad, giants like Meta, Amazon, and Mobileye have been similarly snapping up top-tier teams.
Driving these deals is a pressing need for industry integration as the commercialization of humanoid robots accelerates.
As the sector's core challenge shifts from "can it move?" to "can it be built at scale?", players across the supply chain are viewing M&A as a fast track to filling gaps in their capabilities.
From Domestic to Global, M&A Intensifies
Hangzhou Colin (688611.SH) recently announced plans to spend up to 300 million yuan of self-raised funds to acquire a 41.57% stake in Kepler Robotics from founder Yang Hua and other shareholders.
Prior to this, Hangzhou Colin had already acquired a 9.43% stake in Kepler for 100 million yuan on December 31, 2025, signing a share transfer agreement with Yang to become a shareholder.
That means Hangzhou Colin will have spent a total of 400 million yuan to secure a 51% stake. Once the latest transaction closes, Kepler will become a controlled subsidiary.
Image Source: Kepler
Based on the latest transaction, Kepler’s overall valuation sits at roughly 720 million yuan. In the previous deal, it was valued at the 1 billion yuan mark. Even so, that figure barely registers in a humanoid robot sector populated by 10 billion-yuan "unicorns."
That explains why Kepler, founded less than three years ago, has opted for an early exit. Rather than exhausting itself chasing funding while surrounded by giants, it chose to plug into industrial capital early, buying time and space to bring its technology to market.
Unlike Kepler’s passive move, top-tier players like UBTECH and Zhiyuan are on the offensive, using M&A to drive deeper industry integration.
UBTECH has set its sights on Fenglong Co.
In December 2025, UBTECH (9880.HK) announced plans to acquire a 43% stake in A-share listed Fenglong Co. (002931.SZ) for a total of 1.665 billion yuan, using a combination of private share transfers and a tender offer.
The acquisition follows a "two-step" strategy designed to secure absolute control of Fenglong Co.
AgiBot, meanwhile, has moved into Shangwei New Material.
The deal began in July 2025, and by November of that year, AgiBot held a cumulative 63.62% stake in Shangwei New Material, making AgiBot Chairman and CEO Deng Taihua the company’s de facto controller.
Recently, AgiBot co-founder and CEO Peng Zhihui officially took the chairmanship of Shangwei New Material, marking the conclusion of this consolidation of control.
Beyond the frequent domestic deals, tech giants like Meta and Amazon are also making aggressive moves in humanoid robotics abroad. While their tactics are similar, their M&A logic differs.
Meta is focused on platform-level intelligence.
In early May, reports surfaced that Meta had completed a full acquisition of robotics startup ARI. The roughly 20-person team will join Meta’s Super Intelligence Lab, aiming to bolster Meta’s technical footprint in humanoid robotics and embodied intelligence. ARI specializes in developing AI models that allow robots to understand and predict human behavior in complex, dynamic environments.
Meta CTO Andrew Bosworth has previously stated that the company wants to play a role in robotics similar to Google’s Android. With that goal in mind, Meta’s current strategic priority leans toward building a general-purpose software and algorithm platform.
Amazon, on the other hand, prioritizes closing the loop on scenarios.
To that end, Amazon acquired logistics robotics firm Rivr and humanoid robot company Fauna Robotics in the first quarter of this year.
Fauna Robotics, with about 50 employees, will be integrated into Amazon’s robotics division as a subsidiary. Its core product is a humanoid robot named Sprout, released in late January and positioned as a "ready-to-use" autonomous machine capable of indoor navigation, human recognition, and detection.
Combined with its in-house warehouse robot Blue Jay, Amazon is completing a closed loop spanning warehousing, delivery, and home services.
Image Source: Mobileye
Mobileye, for its part, is committed to accelerating the deployment of physical AI through the synergistic development of robotics and autonomous driving.
Anchored by that goal, Mobileye announced in January the acquisition of Israeli humanoid robotics company Mentee Robotics. The move aims to fully merge the AI technology and mass-production experience amassed in autonomous driving with Mentee’s robotics platform and talent, creating a physical AI powerhouse spanning two transformative sectors.
Taken together, this M&A wave sends a clear signal: humanoid robotics is rapidly moving past the initial stage of technological exploration into a capital-driven phase of consolidation and elimination.
Whether domestic startups are selling to survive, leading players are taking control of listed companies to strengthen supply chains, or overseas giants are buying teams, technology, or scenarios, the essence is the same race: to secure a strategic foothold before the industry shakeout accelerates.
Filling Gaps, Pivoting for the Future
Across this wave of humanoid robot M&A, every deal points to a single imperative: filling capability gaps.
That encompasses not just technology, but also supply chains, mass-production capabilities, deployment scenarios, and even capital access and pivots for transformation.
Image Source: Fenglong Co.
UBTECH’s rationale for controlling Fenglong Co. is stated plainly in the filing: integrating UBTECH’s advanced humanoid robotics technology with Fenglong’s mature manufacturing capabilities, supply chain relationships, and broad customer base will enhance overall product competitiveness, cost structures, market coverage, and the ability to achieve mass production.
Zhiyuan’s takeover of Shangwei New Material follows similar logic. Shangwei’s technical reserves in eco-friendly resins and lightweight composite materials can be repurposed for robotics. For instance, the joint module of the Lingxi X2-N robot uses Shangwei’s vinyl ester resin, achieving dual optimization in weight reduction and battery life.
Similarly, Hangzhou Colin explicitly noted that Kepler’s in-house development system for core embodied intelligence components effectively fills its own technical gaps in humanoid robotics. This complements its existing six-axis force sensors and dexterous hands, accelerating the construction of a full-chain technology system spanning "perception, decision, and execution."
Crucially, having an autonomous and controllable supply chain becomes a lifeline once the stage of mass production is reached.
By comparison, overseas tech giants like Meta and Amazon are largely using acquisitions to quickly plug gaps in core technical talent, accelerating their own R&D and the formation of application ecosystems.
The market waits for no one. Compared to internal development, buying an entire team is the fastest catch-up strategy. This is especially true as humanoid robots shift rapidly from technical exploration to large-scale deployment, placing higher demands on players' systematic capabilities and production efficiency—amplifying the value of M&A.
Of course, because both UBTECH and Zhiyuan acquired A-share listed companies—and given the valuation premiums and high barriers to IPO on the A-share market—some suspect the deals are aimed at "backdoor listings." This view is bolstered by Zhiyuan’s previous endorsement of Shangwei New Material’s long-term value and the A-share market’s capacity to serve technological and industrial innovation.
Both companies, however, have denied this.
UBTECH pledged that it has no plans to change Fenglong Co.’s core business within the next 12 months and will not push for a backdoor listing within 36 months, aiming to maintain stable operations of its existing business.
Zhiyuan Robotics also stated that it has no plans to pursue a backdoor listing through Shangwei New Material (688585) within the next three years.
Previously, rumors circulated that Zhiyuan planned a Hong Kong IPO in 2026 with a target valuation of 40 billion to 50 billion Hong Kong dollars, but the company denied this. However, with top players like Unitree, Leju, and DeepRobotics actively pushing for IPOs, it remains to be seen whether Zhiyuan will accelerate its own listing timeline.
Another driving force is the anxiety of traditional enterprises facing transformation.
Hangzhou Colin stated plainly that controlling Kepler is expected to drive the upgrade and extension of its traditional business into high-value-added intelligent equipment. From Shangwei New Material’s perspective, Zhiyuan’s takeover is also aimed at transformation.
Image Source: PrimeBOT
After Zhiyuan Robotics took the helm, Shangwei New Material clarified a "dual-drive" development strategy focusing on "green new materials and embodied intelligence." On December 31, 2025, it swiftly launched its robot brand, "PrimeBOT," along with its first product: the Q1, a small-sized humanoid robot with whole-body force control, targeting scenarios such as research, education, and home companionship.
However, despite backing from Zhiyuan, Shangwei New Material emphasized that while Zhiyuan Innovation is an indirect shareholder, the two parties remain independent in production, operations, and core technology development. As of now, there are no other related transactions requiring disclosure.
Furthermore, Peng Zhihui, while serving as both Zhiyuan’s CEO and Shangwei New Material’s chairman, holds no senior management or administrative positions at Shangwei and does not participate in specific R&D work.
This implies that this cross-sector acquisition is not a simple business integration, but rather a strategic alliance of mutual convenience. The goal is to seize the initiative to "survive" before the industry shakeout gains speed.
The Other Side of the Double-Edged Sword
The flurry of M&A deals in the humanoid robot sector sends a clear signal: the industry's core contradiction is shifting from "technical verification" to "volume delivery."
The consensus is that global humanoid robot shipments will double this year from a base of nearly 20,000 units delivered in 2025. IDC predicts global shipments will exceed 50,000 units in 2026. More optimistic estimates even suggest the figure could hit 100,000.
Following this, top players like Unitree, Zhiyuan, UBTECH, Leju, and DeepRobotics are aggressively pushing scenario deployment and capacity expansion.
Image Source: Zhenjiang Robotics
Against this backdrop, the industry's main challenge is no longer just developing a viable prototype. It is more critical to achieve cost reduction and mass production while securing continuous orders—all without compromising performance reliability.
This pivotal shift places higher demands on players across the board—whether in terms of capital, supply chain integration, or mass production management.
That explains why financing in the humanoid robot sector has intensified this year, particularly with frequent 1 billion-yuan-plus rounds flowing primarily to top players. Behind this flurry of fundraising lies a consistent goal: accelerating commercial mass production.
But as top players absorb the lion’s share of market capital and resources, the window narrows for everyone else. In that sense, every massive funding round by a leader squeezes the survival space for other players.
Viewed through this lens, Hangzhou Colin’s takeover of Kepler is easy to understand.
In fact, from Kepler’s perspective, this "sale" is not a total retreat. The company has built up expertise in humanoid robot bodies and dexterous hands. However, capital in the primary market can no longer sustain the dual burn of R&D and mass production, and the company cannot close the commercial loop on its own in the short term.
With the path of independent financing narrowing, merging into a listed company is a rational choice.
Yet this transaction also signals that the humanoid robot elimination race has begun first at the capital level, accelerating a divergence. Top players like Zhiyuan and UBTECH are pursuing independent operations and full supply-chain integration, while companies like Kepler—which have a technological foundation but lack capital-raising prowess—are becoming targets for acquisition.
Following current trends, analysts predict that as the deployment of humanoid robots continues to accelerate, similar strategic M&A or controlling acquisitions will increase across the sector and its supply chain from the second half of this year through next.
Image Source: Kepler
However, the other side of this M&A double-edged sword is equally sharp.
On one hand, while the math for industrial synergy looks good on paper, execution can be a different story.
For instance, while UBTECH’s acquisition of Fenglong Co. and Hangzhou Colin’s control of Kepler are both aimed at fixing supply chain gaps, the actual overlap in their businesses remains limited. Moreover, post-merger technical retrofits, production line integration, organizational fusion, and talent retention are all challenges that could significantly dampen synergy expectations.
Consider talent: top AI teams are typically driven by a handful of core scientists. Once research freedom is squeezed by corporate bureaucracy, the loss of core members becomes highly probable. Over the past period, several leading companies have seen executive turnover, partly attributed to capital intervention.
On the other hand, the hidden costs of cross-sector integration cannot be ignored. When Hangzhou Colin acquired Kepler, the latter’s annual revenue was just 4.33 million yuan with a net loss of 66.94 million yuan, while Hangzhou Colin itself was in the red due to the energy storage price war. Merging two loss-making companies will only make the financial statements look uglier in the short term.
Furthermore, the path from R&D to mass production in humanoid robots involves a long cycle of testing, certification, and scenario validation—every step burns cash. If Kepler cannot quickly establish a positive commercial cycle, Hangzhou Colin’s already strained cash flow risks being dragged into a deeper financial quagmire, turning its ambitious diversification blueprint into a "burden" that weighs on the core business.
Of course, these risks have not gone unnoticed.
Buffer zones have been built into the deal structures of several acquisitions. For example, UBTECH’s promise not to pursue a backdoor listing or inject assets within 36 months effectively gives itself a three-year observation window. Similarly, the independent operation of Shangwei New Material’s robotics business can be seen as a form of risk isolation.
This "advance or retreat" design suggests that all parties are maintaining pragmatic expectations regarding the prospects of integration.
Conclusion
In the minds of many, M&A is often equated with "selling out" or "defeat."
But in reality, being acquired is itself a form of value—it proves you hold scarce assets worth buying.
The "Wild West" era of humanoid robotics, where warlords rose together, is ending. A new phase of capital-driven alliances and maneuvering has begun. In this game, how you survive matters less than ensuring you still have a "chair" at the table.
