Per Woofun AI, AI infrastructure startup Upscale AI has finalized a $190 million Series A extension round, elevating its valuation to $2 billion. This latest tranche brings the company's cumulative fundraising total to $500 million. The round was spearheaded by Premji Invest, featuring new entrants such as Temasek and Seligman Ventures, alongside continued backing from existing shareholders including Maverick Silicon, Mayfield, Prosperity7 Ventures, StepStone Group, Tiger Global, NVIDIA, and Salesforce Ventures.

Upscale AI specializes in resolving the 'compute interconnect bottleneck' within large-scale AI clusters by developing a full-stack network architecture that spans chips, systems, and software. As cluster sizes expand, network performance has emerged as a critical constraint on GPU utilization and training efficiency, where inefficient interconnects directly inflate costs and waste compute power. The firm is currently collaborating with hyperscale cloud providers and 'new cloud' infrastructure partners for evaluation and deployment, though specific client details remain undisclosed. Industry analysts observe that investment focus is shifting from model layers to underlying hardware, positioning network interconnects as a pivotal infrastructure track following GPUs.