AI infrastructure company Crusoe announced on Thursday that it had raised US$3.9 billion in a Series F round, pushing its valuation to US$30.9 billion. The financing matters because it signals a strategic pivot: rather than focusing on large-scale expansion, Crusoe said it will transition its large-scale vertically integrated campuses to modular “Crusoe Spark units” that support the growth of its Cloud service. That refocusing shifts what Crusoe is known for, as the developer behind projects including Oracle’s Abilene data-centre campus, which rents out computing capacity to OpenAI.
The mechanics of the new approach are concrete. Spark units can be built in Crusoe’s factories and delivered by heavy-duty transportation to areas with available power. According to CEO Chase Lochmiller, this will cut construction time down from years to weeks. The portable units can be used for inference because they do not need access to thousands of chips the way training models often do. Lochmiller told The Wall Street Journal: “You don’t actually need an Abilene to do that,” adding that running inference on the same scale as training “can be a bit of overkill.” Crusoe’s manufacturing facility in Colorado has the potential to produce up to 1 GW of Spark capacity per year, and Spark units are already operational in Nevada.
The evidence comes from a single full-text report by Capacity Media, a secondary trade publication. The dossier identifies the co-leads as Atreides Management, Mubadala Capital, and Valor Equity Partners, with backing from new and existing investors including Founders Fund, Nvidia, and Qatar Investment Authority. The report also notes Crusoe’s history with the Abilene site: the company expanded it for Oracle and OpenAI, but leased the buildings to Microsoft instead after the pair went with other developers. Lochmiller estimates the site has around 2.1 GW of planned power capacity, including a 900 MW on-site gas plant for Microsoft. Crusoe’s strategies have included developing and leasing data centres, renting out GPUs, and running customers’ AI models while charging for tokens.
The sector implications are significant. Moving toward modular units could address several concerns tied to expanding large-scale AI infrastructure: access to large amounts of electricity, labour requirements, and environmental concerns. It could also help solve bottlenecks caused by planning and insurance hold-ups. Crusoe is not alone in this direction; the report says it joins the ranks of Huawei, Schneider Electric, ABB, and others investing in portable, modular data centres. With public attention currently focused on large-scale data centres, a shift to smaller modular setups may alleviate worries about the rapid spread of large infrastructure.
The main limitation is that this analysis rests on one secondary source, and the dossier does not include Crusoe’s full announcement, financial terms beyond the round size and valuation, or independent confirmation of production capacity and operational claims. What to watch is whether Crusoe can scale Spark manufacturing and delivery at the stated pace, how customers respond to modular inference capacity, and whether the Abilene campus remains a meaningful part of Crusoe’s business after the Microsoft lease arrangement.