David Wood
Embodied AI steps out of the lab but scaling challenges remain
As embodied AI moves from proof of concept into real-world pilots, industry leaders at the ATxSummit conference in Singapore warn that large-scale enterprise adoption hinges on safety, cost and data governance
At Singapore’s ATxSummit tech conference, industry leaders noted that while embodied artificial intelligence (AI) in the form of robots is becoming more capable thanks to advances in hardware, simulation and sensors, broader adoption will depend on reliability, safety, cost, data availability and stronger governance standards.
William Dally, chief scientist and senior vice-president of research at Nvidia, said the next breakthrough in robotics lies in using AI to help robots perform tasks they have not been explicitly programmed to do. He cited the example of a humanoid robot assembling a model car from a text prompt, demonstrating how robotic foundation models allow machines to act autonomously in the physical world by translating visual inputs into motor actions.
However, Yutaka Matsuo, a professor at the University of Tokyo, cautioned that the industry remains in its infancy. “We are not in the full adoption phase at the moment,” he said, adding that better architectures, algorithms, data, compute resources, cost efficiency and safety systems are still needed.
Om Nalamasu, senior vice-president and chief technology officer at Applied Materials, noted that the industry has moved from asking whether such systems can be built to how they can be deployed safely and reliably. He described 2024 as a proof-of-concept phase, 2025 as a year of demos, and 2026 as a period for pilots.
For robotics to scale, Nalamasu said systems must achieve lower latency, greater energy efficiency and cost-effectiveness. Sensors will also be critical, as robots depend on sensor fusion to understand and respond to the physical world. Data remains another major constraint, with real-world robotics data far scarcer than the internet-scale text data used to train software-based AI models.
“We need to be thinking about standards, interoperability and the governance model,” he added.
Real-world deployments
Commercial deployments are already emerging in structured environments. Zhao Yuli, chief strategy officer at Galbot, said the company has deployed more than 1,000 robots in China across humanoid-operated stores, logistics facilities and pharmaceutical chains. Galbot uses both real-world and synthetic data to train its systems.
Yet generalisation remains a challenge. Zhao noted that while robots perform well in known or semi-structured environments, they struggle in unfamiliar settings. For this reason, Galbot is focusing first on semi-structured scenarios where robots can learn from real-world deployments.
Suthen Thomas Paradatheth, chief technology officer at Grab, pointed out that robotics, unlike software, involves significant physical concerns and “does not have near-zero marginal cost”.
He explained that each robot requires hardware, maintenance, supply chains, sensors and physical integration with buildings. Updating robots is also harder than updating cloud software, especially if a fleet requires new sensors or hardware modifications.
Because of these challenges, Grab conducts extensive testing before scaling. “Before scaling to hundreds of robots, we make sure we crack it first in simulation and with a few robots, while building a data flywheel to monitor, learn from and improve each deployment,” said Paradatheth.
Establishing trust is also essential. Grab’s autonomous vehicle pilot clocked 40,000km and involved months of testing, stakeholder engagement and community consultation before public deployment.
Paradatheth said Grab’s approach begins with the customer problem, rather than the technology. “Fall in love with the customer problem, but don’t fall in love with the solution set,” he said.
Public-private collaboration
The panellists also highlighted the need for public-private collaboration. In China, Zhao said government support has helped create testbeds, strategic projects and long-term funding for embodied AI. In Japan, Matsuo pointed to the AI Robot Association (AIRoA), an open data initiative targeting 100,000 hours of robotics data for researchers and companies developing robotic foundation models.
Safety standards will be critical because embodied AI can affect the physical world directly; unlike purely digital AI systems, robots can cause physical harm if they fail. Matsuo suggested that Japan and Singapore could help shape global standards for safety, interoperability and governance.
Nalamasu added that progress in robotics will not be linear, as advances in hardware, software and data will reinforce one another in a “multiplicative” way.
Dally stressed that embodied AI will only become practical if intelligence can run efficiently on the device itself. “We need to run them on real robots, and these can’t be tethered with an umbilical cord back to the datacentre. They have to be carrying the intelligence on them,” he said. This will require more efficient chips, software frameworks and model architectures.
The promise of embodied AI remains significant. Speakers pointed to ageing populations, labour shortages, healthcare, manufacturing productivity and city operations as key areas where the technology could deliver value. In the near term, industrial and semi-structured environments are likely to lead adoption, but over time, autonomous robots are expected to move deeper into public spaces and homes.
Read more about AI in APAC
- With half of all GenAI projects failing after the proof-of-concept stage, organisations are discovering that technology alone isn’t a silver bullet. Here’s how to turn them into lasting competitive advantages.
- Nutanix CEO Rajiv Ramaswami talks up the company’s agentic AI play, the growing demand for sovereign cloud capabilities, and why decoupling storage from HCI is hastening migrations from VMware.
- At Qualtrics Experience Live in Sydney, leaders from Zip Co, Fonterra, Swyftx and Commonwealth Bank shared how AI is accelerating research, breaking down data silos and turning feedback into measurable business value.
- Monday.com’s Globster service brings OpenClaw to consumers and businesses in a bid to democratise access to agentic AI capabilities.
