Viewing New Competitive Variables in U.S. Manufacturing through the Asia-Pacific Physical AI Trillion-Dollar Market

New Variables in U.S. Manufacturing Competition: Insights from Asia-Pacific's Trillion-Dollar Physical AI Market

Abstract

A recent report by Precedence Research indicates that the Asia-Pacific physical AI market will grow from $21.98 billion in 2025 to $1.032 trillion by 2035, representing a compound annual growth rate of 46.95%. China's annual installation of industrial robots is nearly nine times that of the United States, South Korea ranks first globally in robot density, and India is emerging rapidly with a growth rate of 50.7%. Physical AI is transforming manufacturing from simple automation to autonomous systems integrating perception, decision-making, and action—yet the United States is clearly lagging in this critical arena. This article examines the logic behind the explosive growth of physical AI in Asia-Pacific and explores its deep implications for U.S. manufacturing reshoring, supply chain restructuring, and future industrial competitiveness.

Key Observations

1. Asia-Pacific physical AI is growing at an astonishing pace, expanding nearly 47-fold in a decade: The market was valued at $21.98 billion in 2025 and is projected to reach $1.03 trillion by 2035, with a CAGR of 46.95%. This means market value doubles every two to three years, and 85% of the value growth will occur after 2030. 2. Autonomy is becoming the dominant trend: The share of semi-autonomous systems will decline from 72% to 47%, while fully autonomous systems will rise from 28% to 53%. Humanoid robots are the fastest-growing segment, with a CAGR of 58.4%, and their market share will increase from 5% to 15%. 3. China's "state capital + manufacturing scale" dual-engine model is highly effective: China deployed approximately 295,000 industrial robots in 2024, accounting for 54% of global new installations—roughly nine times that of the United States. The RMB 100 billion national venture capital fund established by the Chinese government far exceeds the $12.4 billion in private investment that year, highlighting the dominance of policy capital. 4. India is becoming the next growth pole: India's market CAGR is as high as 50.7%, making it the only major economy expected to gain share, rising from 10% in 2025 to 14% by 2035. 5. Physical AI is reshaping global manufacturing cost and location logic: The Asia-Pacific region accounts for 74% of global industrial robot installations, meaning the depth of automation in the region is widening the manufacturing cost gap with other regions.

Main Text

I. What is Physical AI? Why is it Experiencing Explosive Growth Now?

Physical AI refers to intelligent systems that integrate computer vision, sensor fusion, reinforcement learning, and generative models into robots, vehicles, and industrial machinery. It enables machines to no longer be tools that execute fixed instructions, but rather intelligent agents capable of perceiving their environment, making autonomous decisions, and taking physical action. The growth of this market in the Asia-Pacific region from $22 billion in 2025 to $1.03 trillion by 2035 is driven by the convergence of three forces:

• China's manufacturing scale and concentrated investment of state capital.- The scale of China’s manufacturing sector and the concentrated injection of government capital. China has the world’s largest factory system and supply-chain network. In recent years, under policy guidance, large amounts of capital have poured into AI and robotics. The role played by the 100-billion-yuan national fund far exceeds that of market-driven private capital.
• Economies such as Japan and South Korea are facing severe population aging. For these countries, automation is no longer an efficiency option but an inevitable choice for sustaining their manufacturing base. South Korea has 1,220 robots per 10,000 manufacturing workers—2.7 times the Asia-Pacific average (131)—reflecting the rigid demand created by labor shortages.
• Humanoid robots are moving from the laboratory to commercial delivery. In August 2026, Unitree Robotics went public in Shanghai with a market capitalization of $9 billion. It sold about 5,500 humanoid robots in 2025 and posted a gross margin of around 60%. The listing provided the humanoid-robot economy with its first public market transparency, proving that humanoid robots are not a conceptual story but a real, profitable business.

2. Market Structure Upheaval: From Semi-Autonomous to Fully Autonomous, From Fixed Equipment to Humanoid Robots

The report reveals a key structural turning point: semi-autonomous systems are losing their dominant position, and fully autonomous systems will account for more than half of the market by 2035. In parallel, the market share of humanoid robots will jump from 5% to 15%, making them the fastest-growing segment.

This means that manufacturing tasks requiring human flexibility are being taken over by robots that can move, operate, and collaborate autonomously. Traditional industrial robotic arms can only perform repetitive actions along fixed paths, while robots driven by physical AI—especially humanoid robots—can adapt to complex, unstructured factory environments. The sales figures (5,500 units) and gross margins (60%) disclosed after Unitree’s listing show that humanoid robot commercialization is far outpacing expectations.

Meanwhile, generative AI is also penetrating physical AI at a rapid pace, becoming the fastest-growing segment of the technology layer with a 53.8% CAGR. It enables robots to move beyond preset code and use large-scale models to understand natural-language instructions and autonomously generate action sequences. This will significantly lower the barrier to robot programming and accelerate factory deployment.

3. Regional Competition Landscape: China Leads, India Rises, Japan and South Korea Go Deep

China remains the absolute dominant player in the Asia-Pacific physical AI market, holding a 38% share in 2025. China is not only the largest installer of industrial robots but also the largest source of AI patents globally, accounting for 69.7% of all AI patents granted worldwide. More importantly, China has formed a complete closed loop: manufacturing scale produces massive application scenarios, application data feeds back into AI training, and government capital continuously supplies ammunition for R&D and deployment.India is the fastest-growing market, with a CAGR of 50.7%. Its share has risen from 10% to 14%, showing that India, which uses manufacturing as its economic anchor, is rapidly introducing physical AI as a tool for leapfrog development. Although the base is relatively low, the high growth means that multinational companies, when planning their supply chains, will increasingly set up intelligent production lines in India.

South Korea's lead in robot density (1,220 units per 10,000 people) is no accident. South Korea has the world's highest number of manufacturing robots per 1,000 people, which is directly related to its demographic structure and high-end manufacturing strategy. South Korea also ranks first globally in per-capita AI patent applications, showing a pattern of "density-based innovation"—although its totals are smaller than China's, its per-capita innovation intensity is extremely high.

IV. What is the real challenge facing American manufacturing?

For a long time, the United States has used policy tools such as the CHIPS Act and the Inflation Reduction Act to promote the reshoring of key manufacturing capacity. But the explosion of physical AI is presenting a new challenge to this reindustrialization movement: the pace of automation in U.S. domestic manufacturing is far slower than that of its main competitors.

It can be inferred from the report that the number of industrial robots deployed in the United States in 2024 was about 33,000 units, less than one-ninth of China's. The Asia-Pacific region accounted for 74% of global newly installed industrial robots, while the United States, as one of the world's largest economies, had an installation scale even below that of South Korea. This shows that U.S. manufacturing still relies heavily on human labor at the physical level, and today, as labor costs rise and supply chain resilience issues become more prominent, this reliance is turning into a structural disadvantage.

More importantly, physical AI is not just equipment automation; it also encompasses software, data, and ecosystems. The United States still leads in AI algorithms and chip design, but in the integrated application of robots and physical systems, at least from the installation rate perspective, a clear generational gap has emerged. Without a national-level physical AI strategy, the reshoring of U.S. manufacturing will face the awkward situation of "factories are built, but production costs still cannot compete with Asia."

V. Supply chain restructuring: The degree of automation will determine locational advantages

Over the past few years, American companies have been advancing "nearshoring" and "friendshoring," trying to move supply chains from Asia back to Mexico or the U.S. mainland. But physical AI is changing the definition of so-called "low-cost manufacturing." An obvious trend is that the higher the degree of automation in a region, the greater the room for manufacturing costs to decline.

The scale of automation in Asia-Pacific means that its unit manufacturing costs will continue to fall, and possibly faster than in North America. The 295,000 industrial robots deployed by Chinese enterprises, together with the 1 trillion yuan investment direction designated by the government, will further consolidate the global supply chain position of these regions in electronics, automotive, machinery, and other fields. At the same time, India's rise offers multinational companies a "+1" option that can replace China, but India itself is also introducing physical AI on a large scale, rather than relying on cheap labor.Therefore, reshoring supply chains to the United States cannot rely solely on tariffs and subsidies; it must also compete with the same density of automation technology. Otherwise, even if factories move back to the U.S., high labor costs and lower automation efficiency will still make American manufacturing uncompetitive in price.

VI. Policy Implications: The U.S. Needs a Strategic Response at the "Physical AI" Level

U.S. investment in AI infrastructure is indeed enormous, but the report shows that in the field of Physical AI—a vertical application for manufacturing—government guidance is especially evident in China. The Chinese government has directly invested hundreds of billions of dollars in AI and robotics through national venture capital funds. Similar approaches can also be seen in South Korea and Japan (albeit with smaller amounts), while current U.S. industrial policy focuses more on chip manufacturing and clean energy, and has not yet given equal priority to Physical AI.

Although the National Science Foundation and the Defense Advanced Research Projects Agency are promoting AI research, they have not yet established a full-chain funding mechanism like China's that covers everything from basic research to industrialization and then to replication and scaling. Private companies such as Tesla and Boston Dynamics are advancing humanoid robots, but they lack large-scale federal orders and application scenarios.

Outlook for U.S. Industrial Trends: What Will Happen in the Next 3–5 Years?

Based on the current evolution path of Physical AI in the Asia-Pacific region, we can make the following trend judgments about the U.S. industrial system:

1. The U.S. manufacturing automation gap will continue to widen: Over the next five years, industrial robot installations in China and the Asia-Pacific region will continue to grow exponentially. If the U.S. maintains its current annual installation pace, its robot density may be surpassed by more countries by 2030, thereby weakening the cost basis for manufacturing reshoring. 2. Multinational enterprises will place greater emphasis on the "AI density" of supply chains: When selecting a location for a new factory, in addition to market, tariffs, and labor costs, whether the local area has Physical AI talent and infrastructure will become an investment evaluation criterion. At present, some U.S. states are relatively ahead in robotics ecosystems (such as the traditional manufacturing belt in the Midwest), but they lack systematic deployment. 3. Humanoid pilots will be carried out on a limited basis in the U.S., but commercialization will lag behind China: Given the profitable demonstrations of companies such as Unitree, U.S. companies will also accelerate the adoption of humanoid robots. However, constrained by batteries, sensors, and supply chain support, commercial deployment may be about two years later than in Asia. 4. Policy developments may give rise to new "Physical AI" promotion legislation: Just as with semiconductors, the U.S. may introduce tax incentives or special appropriations for industrial robots and Physical AI infrastructure. However, given the current political cycle, such legislation is more likely to appear after 2027. 5. Supply chain reshoring is no longer just about "moving factories": In the future, more reshored factories will be built as highly unmanned "lights-out factories," which will precisely require reliance on robots and automation equipment imported from Asia. In the short term, even if the U.S. rebuilds factories, its Physical AI technology foundation may still depend on Asia-Pacific suppliers.

ConclusionThe data from the Asia-Pacific Physical AI market reminds us that the core of modern manufacturing competition has already shifted from "labor cost advantages" to "intelligent automation advantages." China, South Korea, India, and Japan are all using Physical AI to rewrite manufacturing cost curves, while the United States finds itself in a delicate period of transition—it possesses the world's most advanced AI research, yet lacks national-level strategic direction when it comes to manufacturing deployment. If America wants its "reindustrialization" to be truly globally competitive, it must treat Physical AI as industrial infrastructure as critical as semiconductors, and build its own automation industry ecosystem as soon as possible. Over the next five years, whoever wins in the large-scale deployment of Physical AI will define the geographic landscape of next-generation global manufacturing.