en.Wedoany.com Reported - Upscale AI announced the completion of a $190 million extension to its Series A-1 funding round, bringing the company's total funding to $500 million and its valuation to $2 billion. The round was led by Premji Invest, with new investors including NVIDIA, Salesforce Ventures, Seligman Ventures, and Temasek, and continued participation from existing investors Maverick Silicon, Mayfield, Prosperity7 Ventures, StepStone Group, and Tiger Global.
Headquartered in Santa Clara, the startup positions itself as a full-stack AI network infrastructure company, covering chips, systems, and software. Its architecture focuses on creating an open-standard AI fabric that connects accelerators, memory, and storage into a unified infrastructure platform, aiming to reduce network bottlenecks in large-scale AI training and inference clusters. The company stated that it is currently collaborating with multiple hyperscale cloud service providers and emerging cloud providers for customer evaluations and deployments in both scale-up and scale-out AI network environments.
Upscale AI stated that with the growing demand for AI-native network infrastructure, it will use the new funds to accelerate product development and expand operational scale. Open standards, interoperability, and ultra-low latency networking are considered key elements of its strategy, positioning it as an alternative to more vertically integrated AI infrastructure approaches. This funding round comes as investors increasingly focus on networking as a critical constraint in AI infrastructure, alongside computing, memory, power, and cooling.
Industry analysis points out that AI infrastructure is driving the largest network platform transformation in decades. Large-scale AI training and inference clusters impose new requirements for deterministic latency, congestion management, fault tolerance, collective communication, and synchronous data transfer between accelerators. The industry is responding with new technologies and standards such as Ultra Ethernet, UALink, Scale-Up Ethernet (including ESUN and SUE), Multi-path Reliable Connection (MRC), and co-packaged or near-packaged optics for higher-radix AI switching. New network platforms are being designed around AI fabrics rather than traditional enterprise or cloud traffic patterns, creating opportunities for startups focused on AI-native switching chips, systems, and software.
Upscale AI's leadership team has deep experience in the networking chip industry. CEO Barun Kar previously held senior positions at Broadcom, and Executive Chairman Rajiv Khemani co-founded Innovium, which developed the Teralynx Ethernet switching chip later acquired by Marvell. This background helps explain investor interest in the company's full-stack AI networking strategy, although the company has not yet publicly disclosed customer names or volume production results.
This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com
