Upscale AI raises $190 million, valuation hits $2 billion

Upscale AI has turned networking, not model training, into one of the most aggressively funded corners of the AI boom. The AI infrastructure company raised a $190 million extension to its early-stage round, lifting its valuation to $2 billion and bringing total funding to $500 million.

The new money came from a mix that says as much about the market as the company itself. Premji Invest, the investment arm of Indian billionaire Azim Premji, led the round, with Nvidia, Salesforce Ventures, Seligman Ventures and Singapore state investor Temasek joining alongside existing backers Maverick Silicon, Mayfield, Prosperity7 Ventures, StepStone Group and Tiger Global.

Upscale AI is built around open-standard, high-performance networking infrastructure that connects AI chips, memory and storage across fast networks. That is the less visible plumbing behind the AI race, but it is becoming central to how hyperscale training and inference systems actually work. As compute clusters grow larger and more expensive, the ability to move data quickly and efficiently has become a commercial bottleneck, and investors are treating that bottleneck as a major opportunity rather than a niche technical problem.

The company was founded by Barun Kar and Rajiv Khemani, two serial infrastructure entrepreneurs with prior ties to Palo Alto Networks, Innovium and Cavium. Reports say Upscale AI was spun out of Auradine, the blockchain and AI company also founded by the pair. Earlier coverage said the startup launched with more than $100 million in seed funding, a large starting war chest that helps explain how quickly it has reached this scale.

The pace of the fundraise is striking even by AI standards. Upscale AI has now raised $500 million in under 18 months, and its backers span semiconductors, enterprise software, growth investors and sovereign capital. That investor base suggests more than speculative enthusiasm: Nvidia’s presence links the company to the chip ecosystem, while Temasek and Salesforce Ventures signal strategic interest from institutions that see networking capacity as a constraint on broader AI deployment.

Fortune described Upscale AI as aiming to be the next Cisco for AI-era networking, a comparison that captures both the ambition and the market opportunity. If the AI wave once centered on chips and foundation models, the latest financing shows the money is now moving deeper into the stack, toward the infrastructure that determines how fast those systems can talk to one another.