Before the robot is even released, Alibaba and Tencent have bet on a Huawei-affiliated company with a valuation of 7 billion.

A new player with a valuation of 7 billion yuan has emerged in the embodied intelligence track.

According to reports from 'Waves', Mochi Intelligence has completed a Series-A angel financing of over 1 billion yuan within six months, reaching a post-investment valuation exceeding 7 billion yuan. This is one of the largest disclosed first-round financing deals in China's domestic embodied intelligence sector to date.

Investors include more than ten institutions such as Alibaba, Tencent, BlueRun Ventures, Legend Capital, Gaorong Capital, and Source Code Capital.

Interestingly, as of the financing announcement, Mochi Intelligence's first product has not yet made its official debut.

Even more notably, the company has not publicly disclosed any customers, orders, delivery volumes, or revenue figures. Founder Huang Qingqiu has outlined a roadmap that plans to launch its first service-oriented robot in July.

Yet how has this company, which has kept nearly all critical details under wraps, managed to secure a 7-billion-yuan valuation in the first place?

A 7 Billion Valuation, Riding on Two Core Executives

It's no exaggeration to say that a significant portion of Mochi Intelligence's 7-billion-yuan valuation is built on the reputation and track record of its two core executives.

Huang Qingqiu, CTO of Mochi Intelligence, previously served as the head of Huawei's Autonomous Driving AI division, leading algorithm development and mass production for Huawei ADS from version 1.0 to 4.0. In interviews with 'Waves', he also spearheaded the development of the one-stage end-to-end WEWA architecture, achieving mass production at the million-unit scale.

CEO Gao Wenli previously managed overseas markets in Huawei's Carrier Business Line, before co-founding cross-border logistics firm iMile, where he oversaw the expansion of overseas operational networks and physical business operations.

One executive brings experience from advancing autonomous driving from algorithm R&D to large-scale mass production, while the other has deep expertise in overseas markets, operational networks, and commercial delivery.

Capital is betting on whether these two distinct skill sets can be successfully recombined within the robotics industry.

The high level of investor confidence in this combination is closely tied to the development trajectory of embodied intelligence over the past two years.

Robots that can fold clothes, pour water, and sort items are no longer a novelty. Demonstrations at product launches have grown increasingly seamless, while large models, VLA systems, and world models are undergoing rapid iterations.

However, there remains a substantial gap between polished demonstrations and commercially viable products.

Laboratory robots are only required to complete tasks for a few minutes in pre-configured environments. When deployed in real-world settings like shopping malls, hotels, or homes, they must contend with fluctuating lighting conditions, object occlusions, moving people, hardware wear and tear, and various unforeseen scenarios — all while operating continuously for hours, often repeating tasks on a daily basis.

What robots currently lack is the ability to perceive what is happening around them in complex environments and respond reliably and consistently.

This is exactly the problem Huang Qingqiu has been solving over the past several years.

From ADS 1.0 to ADS 4.0, Huawei's intelligent driving system has been designed to handle increasingly complex road environments. Vehicles must detect jaywalking pedestrians, temporary obstacles, intricate intersections, and irregular objects, while also adapting to diverse conditions like rain, fog, backlighting, and nighttime driving.

The system must continuously identify new edge cases, collect corresponding road data, and through algorithm training and version iterations, enable vehicles to comprehend an ever-wider range of complex scenarios.

Huang Qingqiu's role in advancing the one-stage end-to-end WEWA architecture to achieve million-unit mass production demonstrates that his work extends far beyond developing a demo-ready algorithm — he has overseen the entire lifecycle of an intelligent driving system, from continuous learning of real-world conditions to long-term deployment across millions of vehicles.

In other words, Huang Qingqiu's proven ability to refine autonomous driving algorithms over time, enabling vehicles to adapt to complex road environments, is precisely what the embodied intelligence field urgently needs right now.

Vehicles need to understand roads, but robots need to comprehend the entire physical world they operate within.

Gao Wenli's background provides investors with another layer of strategic imagination.

For robots to transition from laboratories to mass markets, they must navigate supply chains, cost optimization, global operations, and commercial delivery. Gao Wenli's accumulated experience at Huawei and iMile has allowed Mochi Intelligence, from its very inception, to build a team that extends far beyond just algorithm and prototype development.

Thus, the 7-billion-yuan valuation assigned by investors is a bet on Huang Qingqiu's ability to bring the iterative, training-driven methodology that allows autonomous vehicles to understand complex physical environments into the robotics sector — and on Gao Wenli's ability to translate that technology into deliverable, market-ready products.

Whether these two capabilities can truly be re-deployed successfully within the embodied intelligence industry will ultimately be validated through Mochi's first robot.

The First Robot: Circumventing Household Deployment Initially

When discussing Mochi Intelligence's first robot, although its physical design, technical specifications, and precise use cases have not been disclosed, Huang Qingqiu has confirmed that the company will not immediately target the massive household robotics market.

This strategic choice alone reveals a great deal about the company's approach.

Mochi Intelligence's long-term vision remains focused on developing general-purpose household robots.

According to Huang Qingqiu's vision, future robots should be capable of dynamically adapting to changes in home environments, independently completing tasks like cleaning, organizing, and delivering items — without requiring users to rearrange their furniture or adjust their lifestyles to accommodate the machine.

However, there are virtually no two identical household environments anywhere in the world.

The same cup that sits on a dining table in one home might be found on a bedside table, bathroom counter, or cluttered desk in another.

Robots must not only recognize these objects but also understand their contextual relationships, determine which items are movable, where they should be placed, and avoid colliding with people or damaging other objects during operation.

Compared to autonomous driving systems that primarily address safe point-to-point navigation on roads, household robots operate in a more confined space but must perform complex 3D tasks like grasping, physical interaction, manipulation, and placement — requiring a far deeper level of understanding of the physical world.

In other words, even with Huang Qingqiu's extensive autonomous driving expertise, attempting to immediately conquer the household robotics space — widely regarded as one of the most challenging deployment scenarios in embodied intelligence — would be far from straightforward.

As a result, Mochi's decision to launch with commercial service scenarios represents a much more pragmatic and realistic path forward.

Compared to residential settings, commercial spaces typically feature more static layouts, and the scope of tasks robots are required to complete is far more easily defined.

The team can deploy robots to repeatedly perform tasks in highly consistent environments, systematically identifying instances of misrecognition, navigation failures, or operational errors. They can then collect targeted data to continue training and refining their algorithms.

This type of environment is ideally suited for Mochi to leverage the closed-loop data iteration principles proven in the autonomous driving industry.

Similar to how autonomous driving systems accumulate data over time, Mochi aims to deploy its robots in real commercial environments, incrementally improving their understanding of the physical world through repeated cycles of failure, learning, and updates.

But herein lies the challenge.

The strategy of starting in relatively standardized commercial scenarios, accumulating data through real-world deployments, and then gradually expanding into households has become nearly ubiquitous across the embodied intelligence industry.

Numerous robotics companies targeting the home market are pursuing comparable approaches: honing their technology in industrial, commercial, or other semi-structured environments before tackling the far more unconstrained residential market.

Therefore, simply choosing to focus on commercial service scenarios does not yet demonstrate how much of a competitive advantage Huang Qingqiu's Huawei autonomous driving background can deliver to Mochi Intelligence.

Whether their data iteration cycle can outpace competitors, whether their robots can adapt to new environments faster, and whether their products can maintain reliable performance during long-term operations will only be determined once their first robot is officially launched and deployed in real scenarios.

What we can confirm at this stage is that Mochi has selected a far more pragmatic starting point for its journey.

Whether this team of Huawei autonomous driving veterans can execute this widely adopted industry strategy in a uniquely differentiated way, ultimately justifying their 7-billion-yuan valuation, remains an open question for now.

References:

Waves, 'Exclusive: Former Huawei Autonomous Driving Large Model Lead Huang Qingqiu Enters Embodied Intelligence, Valuation Exceeds 7 Billion Yuan in Half a Year'

Ifeng, 'Valuation Hits 7 Billion Yuan Within Six Months of Establishment, Mochi Intelligence Rewrites Record for Largest First-Round Financing in China's Embodied Intelligence Sector'

OFweek, 'Huawei Autonomous Driving Veterans Launch Startup, Embodied Intelligence Firm Secures Over 1 Billion Yuan Angel Round from Alibaba/Tencent'

Yiou, '1 Billion Yuan Raised in Angel Round! Huawei Autonomous Driving Team Launches New Venture, Alibaba and Tencent Heavily Back Mochi Intelligence'

This article originates from the WeChat public account 'Blue Book Project', authored by Huang Xiaobin, and published with authorization from 36Kr.