AI · SEMICONDUCTOR · FACT CHECK
SKD WIRETrue
Korean AI company Besl AI has announced a plan to secure 200,000 GPUs by 2030, declaring its conversion to a "neocloud" — a cloud computing model focused on providing GPU infrastructure for AI workloads rather than general-purpose cloud services.
The company's plan centers on building out a large-scale GPU fleet to serve AI training and inference demand. The neocloud model it is adopting has gained attention in the industry as AI companies seek dedicated GPU capacity outside the traditional hyperscaler cloud platforms, which have faced supply constraints amid surging demand for AI compute.
Besl AI framed the 200,000-GPU target as the foundation of its new business direction, positioning the company as a specialized GPU infrastructure provider for the AI era.
Securing GPUs at this scale is a defining challenge for AI infrastructure companies, as advanced GPUs remain in tight global supply. Besl AI's roadmap places the buildup on a multi-year horizon through 2030, giving the company time to phase in procurement as capacity becomes available.
The announcement adds Besl AI to a growing list of companies repositioning themselves around AI-specific compute infrastructure, reflecting a broader industry shift in which GPU access, rather than storage or general computing, has become the core value proposition of cloud services.
The claim, based on the company's announcement, is True.
