Fluidstack operates at the infrastructure layer of the AI boom, supplying the large-scale GPU compute that frontier labs and enterprises need to train and serve models. Founded in 2017 and based in London, the company provides cloud infrastructure purpose-built for AI: dense GPU clusters paired with orchestration and monitoring designed for intensive, long-running training jobs and high-throughput inference. As demand for compute has outstripped supply, specialized AI clouds like Fluidstack have become strategically important.
The platform's value is in reliably standing up and operating clusters at a scale that is hard to achieve. Training large models requires not just GPUs but careful networking, scheduling, and observability so that expensive hardware stays utilized and jobs do not fail silently. Fluidstack packages this into a managed AI cloud, letting customers focus on model development rather than the operational complexity of running massive GPU fleets.
Fluidstack's customer base underscores its position. The company is trusted by organizations including Anthropic and Google, a signal that it can meet the reliability and scale demands of the most sophisticated AI buyers. That credibility has drawn major strategic interest, with reports of large prospective investments and partnerships tied to expanding distribution of advanced AI accelerators through third-party data-center partners.
In February 2025, Fluidstack raised a $200 million Series A led by private-equity firm Cacti, building on an earlier seed round backed by Seedcamp and others. The capital supports continued expansion of its AI cloud capacity. Reports through 2025 and into 2026 indicated further fundraising discussions at substantially higher valuations and potential strategic investment from Google, reflecting the intense demand for dedicated AI compute and Fluidstack's role in supplying it.