Nexthop AI Raises $500 Million at $4.2 Billion Valuation — The Startup Building the Networking Backbone for the $650 Billion AI Data Centre Build-Out

Nexthop AI, a Santa Clara-based startup developing purpose-built networking hardware and software for artificial intelligence data centres, raised $500 million in a Series B funding round led by Lightspeed Venture Partners with participation from Andreessen Horowitz, Altimeter Capital, and Kleiner Perkins, valuing the company at $4.2 billion. The round comes at a moment when networking infrastructure has emerged as one of the primary bottlenecks in large-scale AI training and inference deployments. As GPU clusters scale to tens of thousands of accelerators operating in parallel, the switching and routing infrastructure connecting these chips — determining how quickly data can move between servers, racks, and facilities — becomes as important as the chips themselves. Current enterprise networking hardware from established vendors including Cisco, Arista, and HPE was designed for general-purpose enterprise and cloud workloads, not for the specific latency, bandwidth, and topology requirements of AI training clusters. Nexthop AI builds switches and network operating system software optimised specifically for AI data centre environments — enabling faster data movement between GPU clusters during training and reducing the inference latency that affects real-time AI application performance. Alongside the funding announcement, Nexthop AI unveiled three new networking switches specifically designed for AI data centre communication. The Lightspeed and a16z investment signals top-tier venture conviction that the network layer — not just GPUs, not just storage, not just power — is becoming a primary scaling constraint and an independent investment category in the AI infrastructure stack. The broader context: the world’s largest technology companies are projected to collectively spend $650 billion on AI data centre infrastructure in 2026, creating an addressable market of extraordinary scale for infrastructure companies at every layer of the stack.