Acrab startup raised over $350 million to build a platform for agentic AI: betting on decentralized computing
Singapore-based company Acrab, specializing in developing computing infrastructure for artificial intelligence agent systems, has announced raising over $350 million in total funding. These funds will be directed toward building a full-fledged technology platform, as well as research in computing systems, expanding its partner network, and entering international markets.
Full-cycle development: from chips to AI agents
Founded in 2024, Acrab positions itself as a developer of full-cycle computing architecture. The startup's portfolio includes AI chips, solutions for local deployment of large language models (LLMs), operating systems, multimodal interfaces, and AI agent management technologies. The company's key product is the GΞLIX platform, designed for local execution of agent AI tasks. According to the Acrab team, the solution has already been tested in real-world conditions and is at the stage of transitioning to its first industrial deployment and mass production.
Notably, the emphasis on local deployment of LLMs and agent systems reflects a growing trend toward decentralization of computing power in the AI field. In an environment where cloud solutions are becoming increasingly costly and centralized, such initiatives could significantly reshape the market landscape.
In my view, raising this amount of funding at an early stage is not only evidence of investor confidence in the Acrab team but also a signal that the market for high-performance computing for AI is rapidly shifting toward specialized hardware solutions. If the startup successfully implements its ambitious plans, we may witness the emergence of a new standard for agent AI that combines the performance, autonomy, and security of local computing.