Industry giants are causing disruptions in embodied intelligence. What can we expect at the World Robot Conference?

On April 19th this year, the half - marathon for humanoid robots was held simultaneously in Beijing Economic - Technological Development Area. The Tianguang robot took the lead in crossing the finish line and won the championship. From the 14th to the 17th of this month, the world's first large - scale comprehensive event with humanoid robots as the main participants, the World Humanoid Robot Games, will be held at the National Speed Skating Oval, also known as the "Ice Ribbon". More than 100 Chinese and foreign teams will showcase their multi - dimensional capabilities from hip - hop dancing and martial arts to factory handling and hotel services in 26 scenario battles.

Between these two events, a global conference in the field of embodied intelligence, the World Robot Conference (WRC), will open in Yizhuang, Beijing.

Compared with the emergence of "embodied intelligence" last year, its popularity has soared this year. In just the first seven months of 2025, there have been 141 domestic financing events related to embodied intelligence, with the total financing amount exceeding 23 billion yuan, more than the whole of 2024. Four companies, Galaxy Universal, Xinghai Map, DeepCloud AI, and Starry Era, account for more than 80% of the disclosed financing amount in Q3. Industrial capitals such as Meituan, CATL, and the Beijing Robot Fund are replacing pure financial VCs as the main investors. As the track attracts more attention, capital's consideration of "delivery ability" and "scenario implementation" is being advanced.

However, more intense than the capital influx is the early outbreak of the battle for market share - the industrial trend is quietly shifting.

Automobile manufacturers, AI companies, and system platform giants are fully entering the market. With their vehicle - level manufacturing capabilities, large - model - level computing resources, and full - link ecological integration capabilities, they are directly targeting the technological moats of existing players.

On the eve of the World Robot Conference, the issues have become extremely clear:

Will this be the real watershed for the robot industry?

01

The Entry of Giants Becomes a New Core Variable

Based on the highly compatible technological foundation and industrial cooperation path, the entry of automobile manufacturers is particularly worthy of attention - standing on the shoulders of giants, many of the wheels for building robots have already been made.

Humanoid robots and smart cars have a high degree of overlap in modules such as perception, decision - making, and execution. Automobile manufacturers' profound accumulation in sensors, chips, batteries, and autonomous driving algorithms can almost be "transferred" to the robot field, bringing great acceleration to its development.

More importantly, automobile manufacturers have already established a mature supply chain and manufacturing system. From product design, component integration, to production line debugging and quality control, automobile manufacturers have a natural mass - production advantage and can quickly push robots from the prototype stage to large - scale delivery. Take Tesla as an example. When planning the Optimus humanoid robot, it based on its factory manufacturing capabilities, planning to produce a small number in 2025 and expand production at a ten - fold rate starting from 2026. In China, XPeng's Iron humanoid robot is driven by self - developed chips and end - to - end large models. It plans to start with the production line and then enter households, outlining a full - closed - loop logic from the workshop to the living room. GAC has launched its third - generation humanoid robot with a wheel - legged fusion structure, and Great Wall has strategically invested in Unitree Robotics to layout the quadruped field. Even Xiaomi is advancing on two fronts - testing CyberOne on its own production line while leveraging its advantages in the automotive chain and AIoT fields to expand service robot scenarios.

It can be said that the trend of automobile manufacturers building robots has emerged. According to statistics, nearly 20 automobile manufacturers around the world have announced their entry into the humanoid robot field. They come with capital, technology, and the bonus of engineers, injecting a real "industrial - level acceleration" into the robot industry with a combination of "technology reuse + supply chain collaboration + scenario self - operation".

The entry of AI large - model giants is expected to reshape the threshold of the robot track from the "intelligence" level.

Currently, most robots are still trapped in the shackles of "low intelligence" and "weak generalization", relying on preset tasks and having difficulty in flexibly responding in complex environments. This is exactly the area where general artificial intelligence (AGI) excels. AI companies with large - model training capabilities are accelerating the migration of their algorithm systems to robots, trying to connect the full - link intelligence of perception, language, and motion control.

Large technology companies such as OpenAI, Google, Microsoft, and domestic ones like Baidu and Tencent, which have a large amount of computing power and algorithm talents, have begun to migrate general AI models to the robot field. For example, Figure, a company invested in by OpenAI, has developed the Helix large model, trying to uniformly process robot perception, language, and motion control to overcome the adaptability problem in complex environments. Domestic startups such as Zhipu Robot have also launched the general embodied large model GO - 1 and open - sourced its humanoid robot base platform, aiming to "make the large model the brain of the robot". The humanoid robot "Mornine" jointly developed by Chery Automobile and AI company Aimoga is embedded with a large language model and has the ability to conduct natural language conversations, communicating with customers as a sales guide in 4S stores.

It can be predicted that the perception generalization, autonomous decision - making, and human - machine dialogue abilities brought by large models will transform robots from "mechanical executors" to "semantic understanders" and "intelligent agents", especially achieving a leap - forward upgrade in adaptability in unstructured environments.

In addition, the data and cloud service capabilities of AI companies also provide another layer of "remote intelligence" for robot systems: through mechanisms such as cloud - based brain support, online model updates, and multi - robot data joint training, continuous evolution and unified scheduling of intelligence can be achieved. This ecological - level ability is exactly what traditional robot companies lack.

Once these AI giants enter the market in a platform - based manner, whether by launching integrated hardware - software products or building intelligent platforms at the operating system level, they may quickly establish new industry standards. Technology companies such as Microsoft, Google, and NVIDIA, known as the "large - model enablers in the robot track", have deeply laid out the embodied intelligence ecosystem through investment, open - sourcing, cooperation, and other means.

Another chain effect is industrial collaboration and application scenario expansion.

Compared with previous startups focusing on single - point breakthroughs, automobile manufacturers and technology giants naturally have rich industrial resources and in - depth scenarios. They not only "build robots" but also "use robots" - deploying products in their own factories, warehouses, and store systems first, verifying and refining them in real - world operations, and then gradually promoting them to partners and the upstream and downstream of the industrial chain, promoting the formation of a positive - cycle implementation flywheel. For example, BMW, NIO, etc. have introduced humanoid robots on their production lines for handling, quality inspection, and other operation tests. They are no longer just "show - case prototypes" but real "production - line employees".

Technology companies are also accelerating the embedding of robots into their original ecological layouts. Baidu, with its Apollo autonomous driving and PaddlePaddle deep - learning platform, already has the full - stack foundation for building a robot operating system. Alibaba has the resource aggregation ability of T-head Semiconductor chips, large - model computing power, and Cainiao logistics scenarios, which is expected to incubate integrated hardware - software robot solutions. In this model, robots are regarded as nodes of "ecological extension" rather than independent product lines, thus leveraging ecological collaboration to quickly achieve scale.

This is exactly where many startups are hard to match - they usually have to struggle from scratch in technology and manufacturing, while giants can start from an established foundation and break through multiple value chains with systematic resource integration capabilities.

02

The Elimination Round Has Begun: The Delivery Closed - Loop Determines the Outcome

This also brings a key question: Will the entry of giants completely rewrite the existing pattern?

The answer is: It may be inevitable.

On the one hand, the giants do bring a dimensionality - reduction strike. They have capital, technology, scenarios, and talents. Once they enter the market, they may quickly raise the industry threshold and squeeze out a number of small and medium - sized enterprises without moats from the track; on the other hand, they are also "lighting up the track". The high - profile entry of giants re - confirms the commercial value of robots and opens up capital and cooperation channels for startups. There are many cases where giants actively invest in excellent startup teams, pulling the entire ecosystem to leap forward through a combination of "capital + scenarios".

In fact, more and more automobile manufacturers are choosing a collaborative path of "investment instead of self - research". For example, SAIC, BAIC, etc. have strategically invested in new teams such as Zhipu Robot and Zhujidongli to quickly enter the embodied intelligence layout. This "big company + startup" combination model may become an industry consensus: giants are responsible for providing capital, and small companies are responsible for developing the core technology, jointly improving the ability to connect the product delivery closed - loop.

Ultimately, what determines the survival of an enterprise will no longer be concepts, traffic, or exposure, but - delivery ability.

Whether a robot can be truly delivered to users on a large - scale and stably and continuously generate value will be a life - and - death question for all players. This is not only about R & D and cost control but also a competition of systematic capabilities involving manufacturing, supply chain, project management, service system, and commercial implementation.

"You can tell stories in the prototype era, but you can only show strength in the delivery era." Entering 2025, capital no longer pays for prototypes, and the market no longer votes for the future. Companies that only show off their strength and pile up parameters but cannot deliver product value are being eliminated. As giants promote the upgrading of delivery standards, the ability to "deliver faster, at lower cost, and with higher reliability" will become the entry ticket for the new round of competition.

The reshuffle has begun, and the real industrial pattern will emerge in this round of major tests.

This article is from the WeChat official account "Jixin", author: Wang Xian, published by 36Kr with permission.