Acrab has raised $350 million to build infrastructure for agentic AI: a breakthrough in computing systems.
Singapore-based technology startup Acrab has announced raising over $350 million in total funding. These funds will be directed toward developing next-generation computing infrastructure focused on agentic AI systems. Amid the rapid growth of the artificial intelligence market, where traditional solutions struggle with workloads, this move appears strategically sound.
Founded in 2024, Acrab has already established itself as a developer of full-cycle computing solutions. The company's portfolio includes AI chips, technologies for local deployment of large language models (LLMs), operating systems, multimodal interfaces, and agent management systems. The startup's key product is the GΞLIX platform, designed for local LLM deployment in agentic AI tasks.
Why This Matters for the Market
Agentic AI represents the next stage of evolution, where systems not only respond to queries but autonomously execute complex chains of actions. However, this requires a fundamentally different computing architecture capable of processing data in real time without delays. Acrab, it seems, is betting on local computing, which is critical for privacy and speed.
According to the team, the GΞLIX platform has already been tested in real-world conditions and is moving toward its first industrial deployment. The company also plans to expand into international markets and strengthen partnerships. If the pace of development continues, Acrab could become a key player in the hardware niche for agentic AI.
My comment: The $350 million investment is a strong signal to the market. Acrab has chosen the right timing: demand for local AI solutions is growing, especially in the agent segment. If the startup can scale GΞLIX to mass production, it will occupy a unique niche currently dominated by only a handful of projects. However, competition from giants like NVIDIA and AMD will be fierce—success will depend on the speed of deployment and the actual performance of the chips.