The new force in AI computing power rental, Nscale, is likely to go public in the U.S. in September! Q2 revenue surged to over $100 million, with contracts worth $51 billion, targeting a valuation of $25 billion.

According to informed sources, Nscale, a UK artificial intelligence (AI) and computing company that may conduct its initial public offering (IPO) in the US in September, is currently presenting its business growth situation to potential investors. The sources stated that Nscale's management reported that the company's revenue had reached over $100 million by the second quarter of 2026. In comparison, the revenue in the first quarter of 2026 was approximately $37 million, while the total revenue for 2025 was around $33 million. Additionally, the company reported a total contract revenue of approximately $51 billion. Sources revealed that Nscale is collaborating with Goldman Sachs Group, Inc. and JPMorgan Chase to prepare for a potential IPO. They noted that discussions are still ongoing, but the company might also postpone its IPO plans. Reports indicate that the investment banks previously proposed a valuation target of about $25 billion, up from Nscale's valuation of $14.6 billion during its funding round in March of this year. Currently, Nscale is developing data centers in locations such as Norway and West Virginia. The sources mentioned that Nscale is adding 10 gigawatts (GW) of power capacity to its AI computing centers. Prior to this, the company already had 831 megawatts (MW) of active and contracted power capacity. As of the second quarter of 2026, Nscale is also rapidly expanding its chip scale. The company currently operates approximately 25,000 chips and has about 289,000 chips in operation or under contract. One source noted that most of the chips currently in use by Nscale come from NVIDIA Corporation's Blackwell series. The company has also signed contracts to procure about 194,000 NVIDIA Corporation Vera Rubin GPUs. Nscale's business expansion is not limited to increasing computing power. In July, the company agreed to acquire the software startup Anyscale for $1.65 billion to help clients utilize AI computing resources more efficiently. The Next CoreWeave Logically, Nscale's overall business can be roughly compared to CoreWeave (CRWV.US), the cloud-based AI computing rental giant often referred to as "NVIDIA Corporation's darling," which had a market capitalization of $46.3 billion in the US stock market as of Thursday. Although Nscale is not yet publicly listed, investors have already begun to envision this new generation "AI computing rental arms dealer" as the "next CoreWeave." Nscale was spun off in 2024 from a cryptocurrency mining company and has joined the so-called "neoclouds" in the cloud computing field for AI computing rental, providing channels for enterprises needing specialized AI computing capabilities or larger AI infrastructure to obtain physical delivery or cloud-based AI computing capacity rental. Nscale has two core revenue streams. The first is long-term/reserved capacity and infrastructure service contracts, meaning customers pay for deterministic capacity (similar to reserved positions/clusters in AI GPU computing facilities' actual delivery or long-term leasing), along with service fees for operations and maintenance, high-performance networking, storage devices, and cloud hosting delivery. Nscale has publicly disclosed a multi-year arrangement for physical AI capacity infrastructure services delivery with Microsoft Corporation, which includes delivering approximately 12,600 NVIDIA GB300 AI GPU computing devices in Portugal starting Q1 2026, as well as providing Microsoft Corporation with larger and more comprehensive AI infrastructure services. The second revenue stream is on-demand billing for cloud-based AI computing rental/long-term exclusive use which typically charges based on the actual usage duration of cloud-based AI GPU computing clusters, inference call volume, bandwidth/high-performance storage devices, and other dimensions. It is understood that Microsoft Corporation is investing $10 billion to build a data center in Portugal, collaborating with Nscale and a local Portuguese partner to advance the project. At the same time, Microsoft Corporation plans to lease larger-scale AI computing capacity from Nscale in Norway and the UK. In addition to its collaboration with Microsoft Corporation, Nscale also benefits from its partnership with NVIDIA Corporation. The relationship between NVIDIA Corporation and Nscale is one of "investment and business cooperation," creating a dual deep cooperation model. Furthermore, Nscale is deeply cooperating with OpenAI on large AI infrastructure projects in Europe. Collaborations with these giants are reinforcing Nscale's market position as an "AI GPU cloud computing infrastructure cluster platform." As global demand for AI computing continues to expand, investors are keen to see if Nscale can replicate CoreWeave's growth story after its IPO. However, beyond investor enthusiasm, Nscale faces questions regarding the sustainability of rental profit margins, whether customers can generate sufficient cash flow, the limitations of power supply for AI data center expansion, and competition from rivals such as CoreWeave, Lambda, and Crusoe, all of which are also key concerns for investors.