UK artificial intelligence infrastructure company Nscale is planning to raise up to $3 billion through its first initial public offering (IPO) in the United States, joining the wave of AI data center companies that are rushing to list on capital markets. Insiders say the company could potentially start the listing as early as September this year, and it is currently working with Goldman Sachs and JPMorgan Chase to advance the related plans.
Nscale, headquartered in London, has expanded its artificial intelligence data center business aggressively in recent years. The company is currently building data centers in regions including Norway and West Virginia in the United States. With global technology companies rapidly increasing their demand for AI computing power, the company hopes to raise more funds through the capital markets to support its large-scale infrastructure expansion plans.
Insiders previously said that Nscale has already told potential investors that, prior to listing, the total contract revenue it holds is about $51 billion, providing important business support for its IPO.
However, discussions about an IPO are still ongoing. The specific fundraising size and timing of the listing may still change, and there is also a possibility that the IPO could be delayed. Nscale declined to comment on this.
Under the currently discussed plan, Nscale’s U.S. IPO could raise up to about $3.0 billion. If completed successfully, it would become one of the larger IPO fundraising projects in the AI infrastructure sector in the near term.
Nscale’s board lineup is also drawing considerable market attention, including Sheryl Sandberg, former chief operating officer of Meta Platforms (META.US), and Nick Clegg, Meta’s former global affairs president and former UK deputy prime minister.
Nscale’s business model mainly centers on building large AI data centers and purchasing large quantities of AI chips and other infrastructure equipment, then renting out computing capacity to customers that need to train and run AI models.
As generative AI develops rapidly, global technology companies have seen a sharp increase in demand for GPUs, data centers, and power resources. AI infrastructure firms such as Nscale are investing billions of dollars to expand computing capacity.
Insiders previously said the company plans to add about 10 gigawatts (GW) of power capacity for its AI computing centers. Before that, Nscale already had about 831 megawatts (MW) of power capacity either in operation or already contracted.
By way of reference, 1 gigawatt of power can, at any time, meet the electricity needs of about 750,000 U.S. households. This means that if Nscale ultimately achieves its goal of adding 10 gigawatts, its AI infrastructure scale would expand significantly.
The massive power demand also reflects how capital-intensive current AI infrastructure buildouts are. Companies not only need to buy large quantities of AI chips, but also must invest huge sums in building data centers and secure sufficient power supply and related infrastructure resources.
In addition to expanding data center scale, Nscale is also moving into AI software and compute management.
This July, the company agreed to acquire software startup Anyscale for $1.65 billion, hoping to use its technology to help customers use AI computing resources more efficiently.
By combining data center infrastructure with a software platform, Nscale aims not only to provide customers with computing power but also to improve the efficiency of using AI compute, thereby expanding its business coverage across the AI infrastructure industry value chain.
As Nscale prepares to go public, AI infrastructure firms are driving a new wave of fundraising in the capital markets.
In May this year, data center acquisition platform Blackstone Digital Infrastructure Trust raised $2.0 billion through an IPO; in July, Csquare, backed by Brookfield, went public and raised $1.21 billion.
Meanwhile, data center operator Switch has secretly filed for an IPO and could potentially do so as early as November this year. Prior reports indicated that the valuation the company sought in a round of financing may be close to $50 billion, including debt.
Behind this series of fundraising activities is the rapid surge in global capital demand for AI infrastructure. As large AI models require more and more computing resources, data center operators need to invest huge sums to buy chips, build server rooms, and secure sufficient power supply—prompting more and more companies to seek financing in equity and debt markets.