Physical AI Crosses the Factory Floor: Why Asia's Next Big Bet Is Robots That Think

Physical AI Crosses the Factory Floor: Why Asia's Next Big Bet Is Robots That Think

A new report from Deloitte has declared the age of physical AI, and Asia is leading the charge. Published in April 2026 under the title "Physical AI: The moment of acceleration," the study argues that autonomous robots, smart manufacturing systems, and AI-powered^✦ logistics networks have moved decisively from experimentation to large-scale deployment across the region.

The shift coincides with two parallel developments: South Korea's internet security agency launching a national physical AI security standards project on 7 April, and a wave of physical AI demonstrations at GITEX AI Asia 2026 in Singapore this week, where companies from Japan to Europe showcased everything from brain-computer interfaces to shoe-mounted navigation for the visually impaired.

What Physical AI Actually Means for Business

Physical AI refers to artificial intelligence systems that interact with the real world through robotics, sensors, autonomous vehicles, and industrial automation, rather than existing purely in software. Think warehouse robots that learn to pick fragile items, factory quality-control systems that spot defects invisible to human inspectors, and autonomous delivery networks that reroute themselves around traffic in real time.

The Deloitte report found that physical AI deployments in manufacturing and logistics are accelerating fastest in the Asia-Pacific region, driven by labour shortages in Japan and South Korea, rapid industrialisation in Southeast Asia, and China's massive investment in smart factory infrastructure.

The merger of physical systems with AI is shifting from experimentation to large-scale deployment in manufacturing, logistics, and related sectors.

South Korea Sets the Security Standard

South Korea's move to establish physical AI security standards addresses a growing concern: as robots and autonomous systems proliferate, they become targets for cyberattacks that could disrupt^✦ entire industries. The project, launched by the Korea Internet and Security Agency (KISA) on 7 April, aims to create a comprehensive framework for securing AI systems that operate in physical environments, from factory robots to autonomous delivery vehicles.

This makes South Korea one of the first countries in Asia to address physical AI security at a national level. Singapore has taken the lead on governance with its IMDA Model AI Governance^✦ Framework for Agentic AI, released in January 2026, but South Korea's focus specifically on physical security fills a gap that most regulators have overlooked.

Baseline AI skills are increasingly becoming as fundamental as digital literacy. By equipping students with hands-on experience using AI tools, and supporting our educators to adopt them confidently, we are strengthening the foundations for Singapore's future workforce.

GITEX Showcases the Hardware

At GITEX AI Asia 2026, several companies demonstrated how physical AI is already entering consumer and healthcare markets:

• Ashirase (backed by Honda, Japan) unveiled the world's first shoe-mounted navigation system using haptic vibrations to guide visually impaired users through city streets
• Lifescapes (Japan) introduced a non-invasive brain-computer interface designed for stroke rehabilitation, using AI to interpret neural signals and guide recovery exercises
• Literal Labs (UK, presenting in Singapore) showcased logic-based AI architecture that uses 50 times less energy than conventional approaches, a breakthrough for edge computing in physical AI systems

These are not lab prototypes. Ashirase's navigation system is backed by Honda's commercialisation pipeline. Lifescapes is already running clinical trials in Japanese hospitals. The technology is production-ready, which is precisely the shift Deloitte's report identifies.

The Governance Gap

The numbers reveal a tension at the heart of Asia's physical AI boom. While 74% of companies globally plan to deploy physical or agentic AI within two years, only 21% have governance models mature enough to manage the risks. When AI systems control robots, vehicles, and industrial machinery, the consequences of failure are not just financial. They are physical.

Singapore's IMDA framework, built around the principle of "Make Humans Meaningfully Accountable," offers one approach: mandatory checkpoints, audits, and contractual accountability chains. But as South Korea's new security standards project acknowledges, governance alone is not enough. The physical layer needs its own security architecture, separate from the software governance that most frameworks address.

The rapid adoption of agentic AI across ASEAN suggests that deployment is outpacing regulation. SoftBank's $5.4 billion acquisition of ABB Robotics in late 2025 signalled that major capital is already flowing into physical AI, and the Deloitte report indicates that the investment wave is accelerating rather than plateauing.