Acrab Raises $350M to Build Agentic AI Compute Infrastructure in Singapore

Acrab, a Singapore-headquartered technology company building the hardware and software stack for agentic artificial intelligence, has secured over US$350 million in cumulative financing during its early development phase. The raise positions the startup as one of the most heavily funded early-stage AI infrastructure plays to emerge from Southeast Asia.

Founded in 2024 by Ken Phua, who previously held leadership roles at Arm UK and served as co-CEO of Arm China, Acrab is developing a full-stack compute architecture spanning AI silicon, local large language model (LLM) inference, operating systems, multi-modality human-machine interfaces, and agent orchestration technologies. Its first-generation compute platform, GΞLIX, is designed to support local LLMs for agentic AI workloads and has been validated in demanding real-world deployment environments as it heads toward mass production.

Vertex, the global venture platform backed by Temasek, was among the earliest investors through its Vertex Ventures SEA & India and Vertex Growth funds, and consistently increased its investment as Acrab hit critical technology milestones. Vertex Holdings CEO Kee Lock Chua said the firm backed Acrab on the conviction that the next wave of AI will run at the edge rather than in the cloud, a thesis that has only deepened as Acrab translated it into a validated platform.

With the new capital, Acrab plans to accelerate platform development, deepen research in agentic compute systems, expand collaborations with global technology partners, and strengthen its presence in key international markets. The company's approach — building dedicated silicon for agentic workloads rather than repurposing general-purpose chips — addresses a growing recognition that today's GPU-centric infrastructure is not optimised for the real-time, local inference demands of autonomous AI agents.

Why it matters for Singapore: Acrab represents one of the few Singapore-founded companies building at the semiconductor and systems architecture level, an upstream segment of the AI value chain that the city-state has long sought to develop. If its thesis that agentic AI will shift from cloud to edge computing proves correct, Singapore could emerge as a critical node in the next compute paradigm rather than remaining a consumer of hyperscaler AI. The Vertex-Temasek connection also signals that Singapore's investment ecosystem is willing to back capital-intensive deep tech plays, not just application-layer startups.