Crusoe’s hunt for billions in fresh capital has produced the initial close of an anticipated $3.9 billion AI infrastructure funding round.
The AI infrastructure provider announced on Sept. 17 the initial close of an anticipated $3.9 billion Series F round at a $30.9 billion post-money valuation. Nvidia joined the round alongside investors including Founders Fund, GIC, Qatar Investment Authority and TPG, while Atreides Management, Mubadala Capital and Valor Equity Partners co-led the financing.
The new numbers also reveal how quickly Crusoe has expanded. According to Crusoe’s funding announcement, the company now has more than $140 billion in total contracted value and more than 6 GW of gross contracted capacity across its data center and cloud businesses.
Crusoe lands billions after IPO reports
The funding follows reports that Crusoe was exploring an IPO and seeking billions of dollars in additional capital to expand its infrastructure footprint.
In August, Channel Insider reported that Crusoe was discussing a potential pre-IPO raise of roughly $3 billion at a valuation around $35 billion. The newly announced Series F carries a $30.9 billion post-money valuation.
Crusoe said the round was oversubscribed. In addition to Nvidia, the investor list includes ARK Invest, Salesforce Ventures, accounts advised by T. Rowe Price Associates, Inc., Tiger Global and Robinhood Ventures Fund I, among others.
The company said it plans to use the capital to expand its vertically integrated AI infrastructure business, which stretches from power generation and data centers to Crusoe Cloud.
Crusoe says it now has more than 6 GW of gross contracted capacity across its data center and cloud businesses, with approximately 1 GW of that capacity delivered and operational.
The scale-up comes as other AI infrastructure providers race to secure the capital, power and computing hardware needed to support increasingly demanding AI workloads. Groq raised $350 million to expand its AI cloud, while Gimlet Labs secured $300 million as it builds out its multi-silicon infrastructure business.
Crusoe Cloud bookings jump more than 20-fold
Crusoe’s cloud business is also expanding alongside its physical infrastructure.
The company said Crusoe Cloud bookings have grown more than 20-fold year-over-year to date in 2026. Its Managed Inference business, launched late last year, has already contracted more than $100 million in annual recurring revenue, according to the company’s Series F announcement.
Crusoe is also building infrastructure for some of the AI industry’s most prominent companies.
In a multi-year agreement with Perplexity, announced Sept. 15, the AI search company selected Crusoe to support model training and production inference workloads. Perplexity will use dedicated Nvidia GB300 NVL72-powered clusters for frontier-model development and Crusoe’s Managed Inference service for production workloads.
Its expansion goes beyond cloud services. Crusoe opened a second manufacturing facility in Tulsa, Oklahoma, on Sept. 14, adding 400,000 square feet and bringing its national manufacturing footprint to roughly 1 million square feet.
The Tulsa facilities manufacture medium-voltage switchgear, low-voltage switchboards, electrical enclosures, industrial controls and custom-fabricated copper busbar used to support AI infrastructure.
AI infrastructure is becoming a capital race
Crusoe’s anticipated $3.9 billion round reinforces a challenge facing the emerging AI infrastructure market: adding GPU capacity is only one piece of the puzzle.
Providers increasingly need access to enormous amounts of capital, electricity, data center capacity and advanced chips before they can deliver additional compute to customers. That makes financing and infrastructure execution increasingly important differentiators alongside raw cloud performance.
For MSPs, resellers, and infrastructure partners, Crusoe’s latest funding provides the company with additional capital to expand capacity, but funding alone does not guarantee that the promised compute will be available when customers need it.
Partners evaluating newer AI infrastructure providers should pay attention to how much capacity is operational rather than merely planned, where that capacity is located, what hardware is available, and how easily workloads can move between providers.
Crusoe says approximately 1 GW of its more than 6 GW in gross contracted capacity is currently delivered and operational. As billions more flow into AI infrastructure, the gap between capital raised, capacity contracted and computing capacity actually available to customers will be one of the numbers worth watching.
More News: Google and Nvidia are tackling the AI data center power crunch with a new alliance designed to make massive computing facilities more flexible on the electrical grid.