Acrab raised $350 million to build infrastructure for agentic AI — a new challenge for cloud giants
Singapore-based startup Acrab, founded in 2024, has announced raising over $350 million in total funding. These funds are aimed at creating a new generation of computing infrastructure focused on agent-based artificial intelligence systems. The company intends to use the capital to develop its own platform, conduct research in computing systems, expand its partner network, and enter international markets.
Acrab is betting on a vertically integrated full-cycle architecture. Its development portfolio includes specialized AI chips, solutions for local deployment of large language models (LLMs), operating systems, multimodal interfaces, and AI agent management technologies. Its flagship platform GΞLIX has already been tested in real-world conditions and is in the preparation stage for its first industrial deployment and mass production.
Of particular interest is the focus on local LLM deployment — this is a key differentiator for Acrab from competitors, who primarily rely on cloud servers. Local deployment ensures lower latency, enhanced data privacy, and reduced data transmission costs, which is critically important for real-time agent-based AI systems.
My expert commentary: Raising $350 million at such an early stage (the startup is only a year old) indicates a high level of investor confidence in Acrab's concept. However, success will depend on the company's ability to scale chip production and establish partnerships with major corporations. If GΞLIX truly enters the mass market, it could radically change the AI infrastructure landscape, shifting the focus from cloud computing to decentralized local solutions.