The Zhitong Finance App learned that Neocloud (Neocloud), the new AI cloud computing service provider Lambda, which is supported by Nvidia, is considering raising up to 3 billion US dollars before a potential IPO, with a target valuation of up to 12 billion US dollars or higher, and has received several lists of investment terms. If the deal is completed, it will be a key capital replenishment round for Lambda as soon as it goes public next year. It is also a centralized test of the private equity market's strong demand, growth visibility, and valuation range for new cloud computing platforms.
Lambda expects revenue to exceed 1.5 billion US dollars this year, and the financing scale may reach about twice its annual revenue scale, highlighting the AI computing power business's high dependence on comprehensive funding related to GPU procurement, power capacity, and data center construction. Nvidia's shareholder status helps strengthen Lambda's computing power supply and ecological endorsement, but the computing power infrastructure and computing power leasing expansion competition between Nebius, Nscale, CoreWeave, and iREN also means that investors will eventually focus on examining its GPU utilization, customer concentration, capital efficiency, and free cash flow ability.
New cloud power Lambda is considering raising up to $3 billion before the IPO
According to media reports citing information revealed by people familiar with the matter, the AI cloud computing service provider Lambda supported by the “AI chip superhero” Nvidia (NVDA.US) is negotiating a round of financing of up to 3 billion US dollars to prepare for its initial public offering that may take place next year.
The report added that the new cloud computing service provider is discussing raising capital at a maximum valuation of 12 billion US dollars or more. Financing negotiations are still ongoing, and terms of the deal have yet to be finalized.
The report pointed out that Lambda has received several lists of investment terms for this round of financing. Some people close to the company said that this round of financing may pave the way for its initial public offering as soon as next year.
The California-based business is expected to generate more than $1.5 billion in revenue this year, according to the company.
Lambda did not immediately respond to any media requests for comment.
In November of last year, Lambda raised more than $1.5 billion in a round of financing led by TWG Global. Other supporters include Andra Capital, SGW, the family office of Scott Hassan, an early Google investor; OpenAI co-founder Andre Capassi; Ark Investments by Cathy Wood, a top Wall Street fund manager, and Nvidia's venture capital agency.
The company competes fiercely with other new cloud computing service providers such as Nebius (NBIS.US), Nscale, CoreWeave (CRWV.US), and IREN (IREN.US). According to media reports, London-based Nscale is seeking to raise up to $3 billion through an initial public offering in the US.
GPU leasing takes the main stage in the capital market! “New cloud” type AI computing power factory enters the IPO era
Founded in 2012, Lambda is a pure AI infrastructure company, not a large model developer or traditional data center real estate company; its main business is to build and operate an “AI factory” with Nvidia GPUs as the core, and rent out training, fine tuning, and inference computing power through the cloud. The product covers on-demand rental instances of 1-8 Nvidia AI GPUs, one-click computing power clusters with 16-2,000 AI GPU components, and single-tenant super AI clusters with 4,000-165,000 GPUs and a contract period of more than three years, and integrates high-density power supply, liquid cooling, high-speed optical interconnection, and network infrastructure cluster operation and maintenance.
Lambda stopped its traditional local workstation and server leasing business in 2025 and fully switched to AI cloud computing and delivery of large-scale proprietary AI computing power infrastructure; the multi-year agreement it signed with Microsoft involved the deployment of tens of thousands of Nvidia GPUs.
Lambda is a new “Nvidia First” cloud computing service provider (new cloud, or Neocloud), but there is a clear difference in business maturity and platform depth: Lambda focuses more on GPU computing power, exclusive superclusters, and joint engineering services, and the product structure is relatively simple; CoreWeave has developed into a full-stack AI cloud computing platform covering bare metal Kubernetes, Slurm scheduling, object and distributed storage, high-speed networks, dedicated and serverless reasoning, and AI agent sandboxes.
By the end of March 2026, CoreWeave had 49 data centers, more than 1 gigawatt of operating power, and more than 3.5 gigawatts of contracted power, with first-quarter revenue of US$2,078 billion and a revenue backlog of nearly US$100 billion; in contrast, Lambda expects annual revenue of more than 1.5 billion US dollars, which is significantly smaller, but contracts with investment-level customers such as Microsoft make it closer to “focusing on physical clustered AI computing power factory operators”.
Lambda plans to raise up to 3 billion US dollars and reach a valuation of 12 billion US dollars or more, highlighting that AI computing power infrastructure resources are forming a complete capital cycle of “long-term customer contracts - guaranteed loans and project financing - equity financing - IPOs”. The company has previously obtained $1 billion in syndicated guarantee credit and completed $926 million, Moody's Baa2 rated term loan financing. At the industry chain level, this means that Nvidia GPUs, HBM, NVLink/InfiniBand interconnect, optical communications, liquid cooling, and data center power still have strong order visibility.
However, the valuation of about 12 billion US dollars is equivalent to about 8 times its expected annual revenue. Next, the capital market will no longer only reward the number of GPUs, but will also examine computing power utilization, long-term contract quality, customer concentration, GPU depreciation and renewal speed, financing costs, and free cash flow. This financing proves that AI infrastructure can still obtain huge amounts of capital, but it cannot alone prove that terminal AI monetization or NeoCloud profit models have matured and grown.