[MEDIA RELEASE] Protecting Singapore and the planet: HTX signs MoU with Firmus Technologies to boost sustainable computing

SINGAPORE, 27 May 2025 – HTX (Home Team Science and Technology Agency) and Firmus Technologies have signed a Memorandum of Understanding (MoU) to collaborate on advanced AI infrastructure design and research, supporting HTX’s mission to enhance Singapore’s public safety systems while meeting ambitious energy efficiency goals.

The MoU outlines joint research into liquid-cooled AI infrastructure, focused on reducing land and energy use, and advancing GPU performance density. The two organisations will also explore design principles and technologies for next-generation AI factories – with a view to uplift Singapore’s sovereign capability in mission-critical compute, particularly for public safety and emergency response systems.

The move to partner with Firmus Technologies is aligned with HTX’s commitment to incorporate sustainability technologies into its operations and help with the effort of its parent ministry – the Ministry of Home Affairs – to contribute to the Singapore Green Plan 2030.

“HTX is determined to empower the Home Team with AI, but we are also cognisant of the fact that AI models use enormous amounts of power. This partnership with Firmus Technologies allows us to pursue AI innovation while also prioritising energy efficiency,” said HTX Chief Executive Chan Tsan.

“Protecting the public is paramount. But we also believe in doing our utmost to protect the environment for our future generations.”

“As demand for AI grows exponentially, how we build the infrastructure to support it matters more than ever,” said Tim Rosenfield, Co-CEO of Firmus and SMC. “Our focus is on designing platforms that move from electrons to tokens with maximum efficiency. We’re excited to collaborate with HTX on pioneering infrastructure approaches that deliver high-performance AI while making the most of Singapore’s limited land and power footprint.”

Firmus Technologies has deployed AI Factories in Singapore and Australia, optimised for high-density GPU workloads using direct-to-chip liquid cooling with real-time orchestration software. The company’s integrated infrastructure stack has helped push the boundaries of thermal and power efficiency while delivering enterprise-grade services for LLM training, inference, and agentic workloads.