London-based AI startup Nscale is making an audacious move. Less than a year after emerging from stealth mode, it’s now seeking a massive $2.7 billion in funding to build a network of global AI data centers. This bold plan hinges on a pending partnership with ByteDance, the Chinese tech powerhouse behind TikTok.
According to an offering document seen by Bloomberg, Nscale aims to set up data centers worldwide, packed with Nvidia chips, and lease them to companies like ByteDance to train and operate AI models. This development comes just four months after Nscale secured $155 million in Series A funding, led by Sandton Capital Partners with backing from Kestrel 0x1, Blue Sky Capital Managers Ltd, and Florence Capital. That initial capital was geared toward expanding across Europe and North America.
But Nscale isn’t just chasing scale; it’s building smarter. The company specializes in sustainable AI-focused data centers designed to maximize GPU performance for AI workloads. Thanks to its full-stack approach, Nscale handles both the design of the data centers and the deployment of GPU superclusters, achieving greater speed, efficiency, and reliability.
Its platform is tailored for every stage of the generative AI lifecycle. Developers can train and run models on bare-metal or virtualized GPU nodes, leverage Kubernetes-native services (NKS), schedule AI workloads with SLURM, and pick between serverless or dedicated inferencing options. Nscale also plans to roll out its first public AI cloud service by Q1 2025, offering flexible access to AI-specific training and inference tools.
Since its launch from stealth in May 2024, demand for Nscale’s solutions has exploded. The company’s pipeline of greenfield data center projects has soared from 300MW to 1.3GW. In 2025 alone, Nscale expects to deliver 120MW of new capacity. These new sites are purpose-built to host giant superclusters capable of supporting hyperscalers and companies running next-generation language models.
To stay ahead, Nscale is betting on advanced cooling technologies like closed-loop direct liquid cooling. This strategy not only boosts performance but also reduces the environmental footprint. CEO Joshua Payne emphasized that Nscale’s ability to manage the entire value chain—from designing data centers to building GPU superclusters—gives it an edge over traditional players.
“The AI market is scaling rapidly, and so are we,” Payne said. “We manage every layer of infrastructure needed to meet the intensive demands of large-scale AI customers. The biggest bottleneck to scaling is securing the massive electricity required for these GPU superclusters.”
He added, “With a 1.3GW pipeline, we can design, build, and deliver bespoke GPU clusters faster than competitors, offering better unit economics. Thanks to the strong backing of our investors, we’re positioned to reach more customers globally with sustainable, cost-effective AI infrastructure tailored to unlock new AI capabilities and services.”
If Nscale succeeds in securing the $2.7 billion it’s targeting, it would rapidly ascend as one of the most ambitious players in the booming AI infrastructure space—a sector that is only getting more competitive by the day.
