Upscale AI weighs new funding round at $2 billion valuation

A seven-month-old Santa Clara startup is in talks to raise up to $200 million at a roughly $2 billion valuation, a pace that captures both the appetite and the risk now shaping AI infrastructure investing. Upscale AI, founded by Barun Kar and Rajiv Khemani, has already raised more than $300 million since it launched.

The company raised more than $100 million in seed funding in September 2025 and then secured a $200 million Series A on Jan. 21, 2026. That round brought in Tiger Global, Premji Invest, Xora Innovation, Maverick Silicon, StepStone Group, Mayfield, Prosperity7 Ventures, Intel Capital and Qualcomm Ventures, a list that shows how aggressively capital has moved into the infrastructure layer of the AI boom.

Upscale AI says it is building networking infrastructure for AI compute clusters, with its SkyHammer architecture designed for scale-up AI networking. The company says its system is built on open standards including ESUN, Ultra Accelerator Link, Ultra Ethernet, SONiC and SAI, and that it participates in the Ultra Accelerator Link Consortium, the Ultra Ethernet Consortium, the Open Compute Project and the SONiC Foundation. In March, Upscale AI said it was working with NVIDIA Spectrum-X Ethernet switch silicon and AI-optimized SONiC software, and it says its networking solutions are slated to ship in 2026.

The fundraising surge has put a sharp question in view: how much of Upscale AI’s price is tied to durable customer demand, and how much reflects scarcity in a market where investors are chasing anything that can move data faster between AI systems. Upscale AI says it has already drawn customer interest from hyperscalers and AI infrastructure operators, but the valuation talk comes before the company has fully demonstrated scale in the field.

The startup has also leaned into visibility. After its seed financing, Upscale AI said it was featured on the NASDAQ Tower in Times Square, a signal designed to project momentum as it courts another round of capital. For investors, the bet is that networking will remain one of the bottlenecks in AI expansion. For the market, the test is whether that bottleneck justifies a $2 billion price tag this early in the company’s life.