Singapore plans to put AI-powered humanoid robots alongside Home Team officers by 2028, with the machines taking on dangerous or routine work. A new $100 million centre will develop and test them for jobs including hazmat response, prison cell searches and post-blast investigations.
Watch Desk analysis
What happened
The Home Team Humanoid Robotics Centre, run by Singapore’s Home Team Science and Technology Agency (HTX), opened on 30 September. The Straits Times reports that the 29,000-square-foot facility is built for end-to-end development, with spaces for motion capture, simulation, data collection and physical testing in settings such as a prison cell and chemical plant.
HTX plans to train the robots using its NGINE sovereign AI platform. The centre will also develop agentic AI capabilities for multi-step missions, including recovering from failures and adapting to changing conditions. HTX signed an agreement with ST Engineering to advance robotics and embodied AI; other collaborators include FieldAI, General Robotics and Mistral AI.
Why it matters
This is a substantial move from AI in a screen to machines intended to work in hazardous physical environments. The stated aim is to keep officers away from risks such as chemical leaks and unstable post-blast scenes, while freeing them for work that needs human judgement and interaction.
The commitment is a plan, not evidence that the robots are ready for operational duties. The centre’s simulated and physical testing environments will help show whether systems can cope with the conditions they are being built for. HTX says it wants to avoid being locked into one platform or model, a notable ambition in a field where the technology can change faster than procurement paperwork.
Our read
The interesting part is not the humanoid shape on its own. It is the attempt to build, train and test robots around specific public-safety jobs, with a stated 2028 deployment goal and a roster of partners. That gives the project something more concrete to prove than a glossy robot demonstration.
The test will be whether these systems can perform reliably in difficult environments, and whether the promised flexibility across providers survives contact with real-world integration. A robot can take on dangerous work only if the people overseeing it can trust what it does when conditions stop following the script.
What to watch
- Which tasks are selected for early operational use, and whether the 2028 target holds.
- What testing shows about reliability and recovery from failures in changing environments.
- Whether HTX brings in more partners and keeps the centre’s platform-agnostic approach.
Discussion spark: For dangerous public-safety work, should governments prioritise humanoid robots that can fit existing environments, or choose the most capable machine for each task even if that means a less flexible mix of systems?
Sources and evidence
- AI humanoid robots to be deployed with Home Team officers by 2028: Edwin Tong – The Straits Times (30 September 2026, 09:04 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.