AI data center startup Nscale has scored a huge prize, raising $2 billion at a $14.6 billion valuation in a deal backed by Nvidia.
The valuation signals the aggressive push for critical specialized infrastructure as companies race to build the massive computing clusters needed to train and deploy AI models. It also indicates that Nvidia is going beyond selling GPUs and expanding to build an entire ecosystem powering AI’s skyrocketing growth.
Nscale also announced it has added three new directors to its board: former Meta COO Sheryl Sandberg, former UK deputy prime minister and Meta executive Nick Clegg, and former Yahoo President Susan Decker.
Emerging from the shadows
Founded in 2024, the UK-based Nscale has emerged as a key player in the AI infrastructure buildout, developing data centers and operating cloud computing services. The company provides specialized data center offerings designed for the unique demands of AI workloads, making it a critical player in the AI expansion while staying under the radar.
The startup has raised a significant amount of funding in the past year as investors have flocked to the sector. Nscale announced a $1.4 billion delayed draw term loan in February and a $1.1 billion Series B last September, although it did not disclose the valuation it achieved at the time.
Nscale added $14 billion to its coffers in an “expanded partnership” with Microsoft, announced in October. Last summer, the startup partnered with OpenAI to launch a Stargate-branded AI data center in Norway.
In announcing the Nvidia deal on Monday, an elated Josh Payne, CEO of Nscale, said that the AI boom is “leading to the largest infrastructure buildout in human history … We are building this foundation that the market sits on, the engine of superintelligence.”
Nscale is mulling an initial public offering, the company told CNBC in October.
Heightened demand for AI-focused facilities
In contrast to traditional data centers built for general computing or cloud services, AI-focused facilities require substantial power delivery, state-of-the-art cooling systems, and network architectures optimized for the continual movement of data that GPU clusters require.
The timing is ideal. With companies such as OpenAI and Meta clamoring to build large language models, demand for AI compute capacity has soared. OpenAI’s Sam Altman has discussed plans for data centers consuming gigawatts of power greater than all of New York City. Traditional data centers were not designed to deliver infrastructure at that kind of scale.
Infrastructure now a hot commodity
The $14.6 billion valuation also reflects the vast amounts of capital flowing into AI infrastructure. No longer are AI applications getting the lion’s share of the attention. Lately, liquid cooling startups and companies building modular data center offerings have been flooded with cash from venture investors.
With compute capacity growing scarce, the infrastructure layer is drawing investors, gambling on the fact that as AI scales, a strong indicator of who wins the AI era will be whoever controls the data centers.
Also read: Oracle’s push to raise $50 billion for AI cloud growth shows how quickly the capital race is escalating.
