Acrab has raised $350 million to build infrastructure for agentic AI.
Singapore-based startup Acrab has announced raising over $350 million in total funding. The funds will be used to develop computing infrastructure focused on agentic artificial intelligence systems. The company plans to allocate capital to build its own platform, conduct research in high-performance computing, expand its partner network, and enter global markets.
Full-Cycle Development
Founded in 2024, Acrab is developing a full-cycle computing architecture. Its portfolio includes AI chips, solutions for local deployment of large language models (LLMs), operating systems, multimodal interfaces, and AI agent management technologies. The first platform, named GΞLIX, has already been validated in real-world conditions and is aimed at local LLM deployment for agentic AI tasks. According to the team, the solution is moving toward its first industrial implementation and mass production.
Notably, Acrab is betting on local computing, which is particularly important in the context of data privacy and latency reduction. While many giants such as NVIDIA and AMD focus on cloud solutions, Acrab is trying to carve out a niche in edge infrastructure for agentic AI. This could become a key growth driver, especially in the corporate sector, where data security is a critical factor.
From my perspective, raising $350 million at such an early stage is a strong signal to the market. However, the main question is whether Acrab can truly compete with established players in the semiconductor and AI infrastructure space. Given the current boom around agentic AI, the bet on local computing looks promising but requires significant investment in manufacturing and development. If the startup manages to scale GΞLIX to mass production, it could become a serious competitor to existing solutions.