Upscale AI, an AI networking startup based in Santa Clara, is reportedly in talks to raise a new funding round at a valuation of about $2 billion, according to Bloomberg. The company is seeking between $180 million and $200 million. Final terms have not been set, and the company has not commented.
If this round goes through, it will be Upscale AI’s third fundraising effort since it launched seven months ago. The company raised $100 million in seed funding in September 2025, followed by an oversubscribed $200 million Series A in January 2026.
Tiger Global Management, Premji Invest, and Xora Innovation led the Series A, with other investors including Maverick Silicon, StepStone Group, Mayfield, Prosperity7 Ventures, Intel Capital, and Qualcomm Ventures. So far, Upscale AI has raised over $300 million.
Barun Kar, the CEO, and Rajiv Khemani, the Executive Chairman, co-founded Upscale AI. Both have started multiple companies and worked at firms such as Palo Alto Networks, Innovium, and Cavium, which Marvell later acquired. They developed Upscale AI at Auradine, Khemani’s AI and blockchain infrastructure company, before launching it publicly.
Upscale AI believes that traditional data centre networking, which connects general-purpose servers, is not well-suited to AI workloads. Its platform brings together AI clusters by connecting GPUs, accelerators, memory, storage, and networking into one synchronised system. The SkyHammer architecture focuses on scaling up networking at the rack level to reduce latency and bottlenecks that slow AI training and inference.
The platform uses open standards like ESUN, Ultra Accelerator Link, Ultra Ethernet, SONiC, and the Switch Abstraction Interface. This makes it an alternative to networking systems from established companies like Cisco and Broadcom.
Upscale AI is also involved in groups such as the Ultra Accelerator Link Consortium, the Ultra Ethernet Consortium, the Open Compute Project, and the SONiC Foundation.
