宇树4000亿之后 具身智能该如何定价？

This article examines the valuation and commercial development of the embodied intelligence industry following the public listing of Unitree Robotics, a leading humanoid robot developer. It sorts out the current industry status, key pain points, and emerging development directions, with core takeaways as follows:

1. Overall industry momentum: Unitree Robotics debuted on the STAR Market with an opening market capitalization of approximately 450 billion yuan, making it the "first humanoid robot stock" in A-share markets. In the first half of 2024, total financing for China's embodied intelligence track exceeded 90 billion yuan, a 5x year-over-year surge, reflecting steadily growing interest from both capital and the broader market.

2. Existing industry pain points: Most companies remain stuck at the demo demonstration phase, and numerous flaws are exposed when deploying solutions to real-world scenarios. Currently, most projects are highly customized, and the marginal cost of expanding into new scenarios does not decrease as scale grows. This often leads to a situation where larger operation volumes translate directly into greater losses.

3. New development approach: Kairoves Robotics has proposed Physical Token Economics, which centers on measuring physical intelligence output. Its technical framework lowers the marginal cost of adding new capabilities. To date, the approach has been deployed in over 50 application scenarios, with more than 1,500 mature embodied models accumulated.

This article sorts out the valuation logic and development trends of the embodied intelligence industry after Unitree's listing, with key takeaways for brands laying out positions in the humanoid robot track:

1. Branding and pricing reference: Valuation logic varies significantly across different types of embodied intelligence firms. For brands focused on robot hardware, valuation centers on shipment volumes, supply chain capabilities, brand influence, and sales channels. Brand strength and channel access directly determine a company's market value.

2. Consumer and market trends: Currently, both end consumers and B-side enterprise clients show extremely high interest in embodied intelligence products, as reflected by the strong audience response at the World Robot Conference. With massive capital inflows, the family, manufacturing, and commercial service sectors will all be core potential markets for embodied intelligence, with large room for consumption demand growth.

3. Product R&D direction: Brands should not only chase flashy demo results. They need to prioritize deployment costs and capability reuse early in development. Only by lowering the marginal cost of entering new scenarios and tackling new tasks can brands achieve sustainable profitability and avoid the trap of diseconomies of scale.

Unitree's listing has opened a new stage of capitalization for the embodied intelligence industry, with these key insights for sellers entering the track:

1. Market opportunities: The embodied intelligence track is currently experiencing extremely high momentum. Total domestic financing exceeded 90 billion yuan in the first half of the year, a 5-fold year-over-year increase. More than 20 embodied intelligence firms have already announced clear IPO plans, reflecting strong capital recognition. There is massive unmet demand across industrial manufacturing, commercial retail, logistics services, and home scenarios, leaving broad room for growth.

2. Risk warning: The industry currently has clear froth. Companies that only rely on demo demos and concept storytelling will face growing survival pressure. Many firms are trapped in a project-based cycle, where every new project requires new investment in debugging and model training, keeping marginal costs high. This often leads to the problem of greater losses as scale increases.

3. Replicable development strategy: Sellers can reference Kairoves Robotics' business model. By building data infrastructure and simulation systems, the company enables experience accumulation and reuse, continuously driving down the marginal cost of new capabilities. This approach allows firms to capture scale compounding and win higher valuations from capital markets.

Embodied intelligence is accelerating into industrial production, with these core insights for factories looking to advance intelligent transformation:

1. Business opportunities: Embodied intelligence has already been deployed across multiple industrial links including sorting, assembly, and quality inspection. It can replace human workers for repetitive, standardized tasks to significantly boost production efficiency. Factories are a core application scenario for embodied intelligence, and intelligent transformation has now entered the real-world deployment phase.

2. Solution selection criteria: When choosing an embodied intelligence solution, factories should not only judge by demo performance at trade shows. They need to prioritize deployment costs and capability reusability. Solutions that can adapt to multiple production tasks within a single technical framework have lower marginal costs for new tasks in the long run, delivering higher overall cost-effectiveness.

3. Insights for digital transformation: Factories should prioritize solutions that can accumulate production data and continuously iterate models during transformation. Experience data generated from real factory production can flow back to improve model performance, further reducing deployment costs for future new tasks, forming a positive cycle and lowering long-term intelligent transformation investment costs.

The embodied intelligence industry is currently shifting from the technology demonstration phase to the commercial deployment phase, with these key insights for service providers serving the track:

1. Industry development trends: Unitree's listing has accelerated the overall capitalization process of the industry. More than 20 embodied intelligence companies have already announced IPO plans, and industry consolidation is set to begin. Players that can control costs and achieve capability reuse will ultimately win. The track has broad overall growth space, and demand for related supporting services will continue to grow.

2. Core client pain points: The biggest pain point for clients deploying embodied intelligence today is the high cost of customized deployment. Entering every new scenario requires new data collection, model training, and on-site debugging, keeping marginal costs persistently high. Many projects fail to meet return on investment thresholds and never go live.

3. Reference solution direction: The Physical Token economics and technical path proposed by Kairoves Robotics has already been validated through real-world deployment. Through unified data representation, generative simulation to compress environment setup time, pre-training to accumulate general capabilities, and real-machine experience feedback for iteration, the framework continuously lowers the cost of deploying to new scenarios. This approach can serve as a valuable reference for serving clients.

The rapid growth of the embodied intelligence industry offers these key takeaways for platform players布局 the track:

1. Core industry demand for platforms: Business models vary widely among embodied intelligence players, and the profit logic of four categories — hardware manufacturing, scenario delivery, embodied brain development, and robot labor platforms — are completely different. The existing one-size-fits-all valuation framework is mismatched. Platforms need to build a layered, categorized valuation and service system to match the needs of different players.

2. Investment and operation direction: When sourcing and screening projects, platforms should not only prioritize demo performance and technical concepts. They need to focus on evaluating companies' cost control capabilities and level of capability reuse. Companies that can continuously drive down the marginal cost of new scenarios have greater long-term growth potential and stronger risk resistance.

3. Risk mitigation: Platforms should be wary of concept-driven companies that only rely on storytelling. These firms deliver impressive demos but lack commercial deployment capabilities, cannot lower marginal costs, and struggle to achieve sustainable profitability. They will be eliminated during industry consolidation, so platforms should screen them out early to avoid associated risks.

This article sorts out the latest developments in the embodied intelligence industry after Unitree's listing, and puts forward new problems and a new framework for industry development, with core insights for industry researchers:

1. Latest industry trends: Unitree's STAR Market listing made it the "first humanoid robot stock" in A-shares, with an opening market capitalization of approximately 450 billion yuan. In the first half of the year, total financing for China's embodied intelligence track exceeded 90 billion yuan, a 5-fold year-over-year surge. More than 20 embodied intelligence firms have already announced clear IPO plans, significantly accelerating industry capitalization while also triggering an impending industry shakeout.

2. Core new industry problems: Widespread valuation misallocation exists across the industry today. Companies with different business models share a single valuation framework, which cannot accurately reflect firms' true value. At the same time, a large number of companies are stuck at the commercial deployment stage. The project-based model prevents marginal costs from falling as scale expands, making it impossible to achieve economies of scale.

3. New theory and business model: This article introduces Physical Token Economics, which establishes a unit of measurement for physical intelligence output and defines the core path to achieving economies of scale for general embodied intelligence: continuous reduction of the marginal cost of new capabilities. This provides a new analytical framework for the next phase of development for the embodied intelligence industry.