ai automationdigital transformationai strategy

Physical AI: APAC Factories Writing the Competitive Playbook

26 May 2026·10 min read

405M robots today, 1.3B by 2035. BYD's Xi'an factory runs at 97% automation. Physical AI isn't coming—it's here. Learn from APAC's playbook.

Physical AI: APAC Factories Writing the Competitive Playbook

Physical AI — Why APAC Factories Are Writing the Next Competitive Playbook

> TL;DR: 405 million robots are working globally today, scaling to 1.3 billion by 2035. BYD's Xi'an factory already operates at 97% automation with 95% grasping success rates. Physical AI has moved from pilot to production in APAC, creating a competitive advantage through four maturity stages: Automation → Collaborative Digitalisation → Digital Twin → Physical AI. While 80% of enterprises globally remain in stages 1-3, APAC manufacturers are rapidly progressing to full Physical AI integration. The question isn't whether this technology is coming—it's whether you'll adapt APAC's proven playbook or develop your own strategy from scratch.

Excerpt: There are 405 million robots working right now. By 2035, there will be 1.3 billion. A BYD factory in Xi'an already runs at 97% automation. Physical AI is not coming — in APAC, it is already here. The question is whether you will learn from their playbook or write your own from scratch.

Slug: physical-ai-apac-factories-competitive-playbook

Categories: ai-automation, digital-transformation, ai-strategy

There are 405 million robots deployed globally today. By 2035, that number will reach 1.3 billion. Not prototypes. Not pilot programmes. Working robots in factories, warehouses, hospitals, and logistics networks — a threefold increase in a single decade.

And a BYD plant in Xi'an, China, is already running at ~97% automation. Grasping success rates exceed 95%. This is not a laboratory experiment. It is a production facility building millions of vehicles a year with almost no human hands on the line.

Physical AI — the integration of artificial intelligence with physical systems like robots, autonomous vehicles, and smart infrastructure — is no longer theoretical. In Asia-Pacific, it is operational. And the companies not paying attention are about to find themselves competing against factories that never sleep, never make errors, and cost a fraction of their human-run equivalents — a competitive dynamic we explore in the agentic AI market growth trajectory.

The numbers behind the transformation

Deloitte's 2026 APAC Physical AI Whitepaper provides the most comprehensive look at where physical AI stands today and where it is heading. The acceleration is remarkable.

Physical AI is transforming 5% of enterprises today. That will jump to 41% within three years. We are moving from early adoption to mainstream deployment at a pace that mirrors the cloud adoption curve — except faster.

For "extensively integrated" physical AI — the deepest level of deployment — the numbers go from 3% today to 18% within two years. A sixfold increase in the most mature form of physical AI, in just 24 months.

The hardware scale is equally dramatic. 500,000 industrial robots were deployed in 2024 alone. By 2028, annual installations are forecast to reach 700,000. That is a factory floor being built in real time, mostly in APAC, mostly in China, South Korea, and Japan.

In the United States, robotics funding tells the same story from the investment side: $10.3 billion was raised in the first 11 months of 2025, a 61% year-on-year increase. Capital is flowing into physical AI at accelerating rates.

The BYD case — what 97% automation looks like

BYD's Xi'an plant is the benchmark. At roughly 97% automation, it represents the near-endpoint of the physical AI journey. Robots stamp, weld, paint, and assemble. AI systems monitor quality in real time, catching defects that human inspectors would miss. Logistics robots move parts between stations without human guidance.

The grasping success rate — the percentage of times a robotic arm successfully picks up and places a component — exceeds 95%. This is a metric that, just five years ago, was below 80% in most facilities. The improvement is not incremental. It is transformational.

What makes BYD different is not just the technology. It is the integration. The factory operates as a single system — robots, AI, logistics, quality control, and production scheduling all connected and optimised in real time. This is not automation layered onto an existing process. It is a process designed for automation from the ground up.

That distinction matters. Most Western companies are trying to automate existing processes. BYD rebuilt the process for automation. The results speak for themselves.

Four stages of maturity

Deloitte identifies four stages of physical AI maturity. Understanding where your organisation sits — and where your competitors sit — is the first step to building a strategy.

Stage 1: Automation

Robots perform repetitive, predefined tasks. No AI involved. Think of a traditional assembly line where robotic arms weld the same joint on every car. This is where most Western manufacturing still operates.

The limitation is rigidity. Stage 1 automation cannot adapt to variation. If the part is slightly different, the robot stops. If the process changes, the robot needs reprogramming. Efficiency gains are real but bounded.

Stage 2: Collaborative Digitalisation

Humans and robots work side by side, with digital systems coordinating their interactions. AI begins to play a role — not controlling the robots, but scheduling their tasks, predicting maintenance needs, and optimising workflow.

This is where many APAC factories are today. The human-robot collaboration creates efficiency gains without requiring a full redesign of the production process.

Stage 3: Digital Twin

The entire physical operation is mirrored in a digital model. AI can simulate changes, test scenarios, and optimise performance before any physical change is made. The digital twin becomes the planning layer — every decision runs through the simulation first.

This stage is where APAC's leading manufacturers are heading. The digital twin does not just optimise current operations. It enables rapid prototyping of new processes without the cost and risk of physical experimentation.

Stage 4: Physical AI

Full integration. AI controls physical systems in real time, adapting to changing conditions without human intervention. The BYD Xi'an plant operates at this level. Robots do not just execute predefined tasks. They adapt to variation, learn from errors, and optimise their own performance.

80% of enterprises globally are in Stages 1 through 3 — Automation through Digital Twin. Only a small minority have reached full Physical AI. But the velocity of movement through the stages is accelerating — and as our cloud maturity stages analysis shows, the companies that move through stages systematically always outperform those that try to skip ahead.

Why APAC is writing the playbook

Several factors explain why APAC factories are leading in physical AI adoption.

Scale. China, South Korea, and Japan have the world's largest manufacturing bases. Scale provides the volume needed to justify the capital investment in robotics and AI. A factory producing 100,000 units a year can amortise automation costs differently from one producing 10,000.

Government support. APAC governments are actively subsidising and incentivising physical AI adoption. China's "Made in China 2025" initiative, South Korea's Manufacturing Innovation strategy, and Japan's Society 5.0 plan all include significant support for industrial robotics and AI.

Talent availability. APAC has a deep pipeline of engineering talent in robotics and automation. India alone expects its AI talent pool to grow from 600,000 in 2022 to 1.25 million by 2027. The talent supply matches the demand.

Cultural readiness. APAC manufacturing culture is more receptive to automation than many Western markets. The cultural resistance that slows adoption in Europe and North America — concerns about job displacement, union pushback, regulatory caution — is less pronounced in APAC, though EU AI Act compliance is increasingly shaping how European manufacturers approach AI deployment.

The barriers — and they are real

Physical AI is not without obstacles. Deloitte identifies the top barriers, and they are instructive for any organisation considering the journey.

41% cite cost and resources as the primary barrier. Physical AI requires significant capital investment — robots, sensors, AI systems, digital twins. The payback period can be long, and many organisations cannot fund the upfront cost.

36% struggle with use case identification. They know physical AI exists but cannot identify the specific applications that would deliver value in their operations. This is a strategy problem, not a technology problem.

33% face talent gaps. They cannot find or afford the people needed to implement and manage physical AI systems. The talent market for robotics engineers and AI specialists is tight globally.

31% point to technology and data challenges — legacy systems that cannot integrate with AI, data quality issues, and infrastructure that was not designed for real-time AI inference.

These barriers are real, but they are not permanent. Cost decreases as the technology matures. Use cases become clearer as early adopters publish results. Talent expands as training programmes scale. The barriers are high today. They will be lower next year. The question is whether you start climbing now or wait until the queue gets longer.

What this means for your organisation

You may not run a factory. But if you have a supply chain, a logistics network, or any physical operations, physical AI is coming to your competitive landscape. Your suppliers, your competitors, and your customers will be affected.

Three practical steps:

Assess your supply chain exposure. Which of your suppliers are in APAC? Which are investing in physical AI? If your key supplier achieves 97% automation, your cost base benefits — but your dependency on that supplier increases. Understand the exposure before it becomes a risk, using a structured approach similar to our post-merger technology integration methodology and tech due diligence checklist.

Identify your Stage. Where does your organisation sit on the four-stage maturity model? Most companies assume they are more advanced than they are. An honest assessment is the starting point.

Learn the APAC playbook. The use cases, implementation patterns, and failure modes are all visible. BYD, Foxconn, Samsung, and others are publishing their results. You do not need to invent physical AI strategy from scratch. You need to adapt what is already working.

The bottom line

Physical AI is not coming. In APAC, it is already here. 405 million robots today. 1.3 billion by 2035. Factories running at 97% automation. Annual installations accelerating. Capital flowing at 61% year-on-year growth.

The playbook is being written in Xi'an, Shenzhen, Seoul, and Tokyo. The question for every other organisation is simple: will you learn from their playbook, or will you try to write your own from scratch when it is already too late?

Key Takeaways

Physical AI adoption is accelerating rapidly: From 5% of enterprises today to 41% within three years, with extensively integrated deployments jumping from 3% to 18% in just two years
APAC factories are setting the benchmark: BYD's Xi'an plant operates at 97% automation with 95% grasping success rates, demonstrating what full Physical AI integration achieves
Four maturity stages define the journey: Automation → Collaborative Digitalisation → Digital Twin → Physical AI, with 80% of global enterprises still in stages 1-3
Scale, support, talent, and culture drive APAC's advantage: Government incentives, engineering talent pipeline, and cultural readiness to automation create competitive advantages
Barriers exist but are temporary: Cost (41%), use case identification (36%), talent gaps (33%), and technology challenges (31%) are real but decreasing over time
Supply chain strategy is critical: Understanding supplier exposure to Physical AI and your organization's maturity stage enables proactive competitive positioning

Is your supply chain ready for Physical AI? TechLevity helps companies assess their physical AI readiness — from supply chain exposure to internal maturity mapping, supported by AI-native engineering expertise.

[Book your Physical AI assessment →]