Home Team to develop space satellite, humanoid robots, more AI: Shanmugam

Misryoum reports Home Team plans a 2029 satellite to detect hazardous gas plumes, expands sovereign AI infrastructure, and will launch a Humanoid Robotics Centre to train robots for high-risk missions.

SINGAPORE — The Home Team is laying out a new push that ties space technology, frontier AI and humanoid robotics to day-to-day public safety.

That direction was spelt out by Coordinating Minister for National Security and Minister for Home Affairs K.. Shanmugam on April 28, as Misryoum covered remarks made at the Milipol TechX Summit (MTX) 2026.. The centrepiece is a first satellite, codenamed Xplorer, planned to be launched in 2029.. The project aims to help the Singapore Civil Defence Force respond faster to hazardous gas situations by detecting hazardous gas plumes in the air, including gases such as ammonia.

Misryoum understands the satellite will weigh about 100kg and be placed into a near equatorial orbit.. The Home Team Science and Technology Agency (HTX) said it is partnering the National Space Agency of Singapore to explore how space technology can strengthen public safety.. The plan is not framed as a standalone capability, but as part of a broader system that includes sensors, analytics and operational response.

The summit also served as a platform for a wider message: technology development is increasingly shaped by geopolitics, and cyber risk is accelerating as AI becomes more powerful and more widely deployed.. Shanmugam said supply chains are being rapidly reconfigured, with a decoupling of hardware, software and ecosystems — which can create friction when nations acquire, integrate and deploy technology.. He added that technology is being “weaponised” into new attack vectors.

Frontier AI models were central to that warning.. Misryoum notes that Shanmugam referenced an advisory issued on April 15 about risks from such models, including their potential to identify vulnerabilities and help compromise systems soon after release.. He described how this could shrink the time between discovery and exploitation from months to hours, leaving defenders with very little room to respond.

Singapore’s emphasis on faster defence is not theoretical, the minister said.. Shanmugam pointed to an 11-month cybersecurity operation to defend critical infrastructure, and he also discussed the most recent incident in Singapore where data involving the upcoming Jurong Region Line MRT stations and the new Changi NEWater Factory 3 was compromised at Shanghai Tunnel Engineering Co (Singapore).. The Home Team’s takeaway, as presented by Shanmugam, is a need to harden systems before threats become operational.

A key part of Misryoum’s coverage here is the Home Team’s push for sovereign AI infrastructure.. Shanmugam said Singapore’s approach is to apply AI widely and responsibly, but he linked serious AI capability to having sovereign infrastructure.. HTX partnered ST Engineering, Google, NVIDIA and Nutanix to build NGINE, which the ministry described as its first sovereign, GPU-powered AI infrastructure.. Misryoum understands HTX will also sign a memorandum of understanding with NVIDIA for future iterations of NGINE.

Beyond infrastructure, the Home Team’s internal AI development is also expanding.. HTX is working with French company Mistral AI and training its own Phoenix family of large language models.. Shanmugam said Phoenix Small is already being used by officers to handle complex information in a secure environment, and that Phoenix Medium is set to be released this week.. He said the newer model can analyse images and documents and can be adapted for more advanced tasks.. A Strategic Partnership for Innovation with Mistral AI is also planned to push these efforts further.

But the minister’s argument was that AI does not operate “in the abstract.” Misryoum notes his point that it needs sensors to see, autonomous systems to act, and officers with operational concepts to work with both.. From there, the next step is humanoid robotics.. Shanmugam said HTX will launch the Home Team Humanoid Robotics Centre in September to train humanoid robots with specialised Home Team skills for operational scenarios involving high risks, including hazardous material response and fire safety..

There is also a cybersecurity “stress test” layer to this roadmap.. Shanmugam said as these systems are developed, they must be assumed to be targeted.. Misryoum understands the inaugural DEF CON Singapore — a hacking convention — will run alongside MTX.. The aim, as described by the minister, is to help Singapore grow cybersecurity talent and to test systems against world-class hackers before they are deployed operationally.

For readers, the practical impact is straightforward: if hazardous incidents can be detected and interpreted faster, response teams may spend less time guessing and more time acting.. If AI infrastructure is harder to disrupt, then critical public-safety services may be more resilient.. And if robotics training is focused on realistic high-risk scenarios, it could reduce the exposure of personnel while improving decision-making during emergencies.

The underlying shift is that public safety technology is increasingly converging across domains — space-based sensing, secure AI operations and human-centred robotics.. Misryoum expects that the outcomes of these programmes, especially the 2029 Xplorer satellite and the September robotics centre launch, will be closely watched not only for capability, but for how quickly concepts move from pilot stages into operations under real-world constraints.