FluidStack in talks to raise $1 billion at an $18 billion valuation after Anthropic deal

FluidStack could become one of the most valuable private players in AI infrastructure: the Oxford-founded company is in talks to raise $1 billion at an $18 billion valuation. For the market, this is not just another venture round. It is a signal that money keeps shifting from apps and chatbots to the foundational layer: compute, GPU clusters and data centers, without which the current AI boom cannot scale.

FluidStack belongs to the wave of so-called neocloud providers, new cloud companies that build their businesses around scarce AI resources and infrastructure for training and running models. Unlike classic general-purpose clouds, these players bet on access to GPUs, tight integration with data centers and faster deployment of capacity for specific AI workloads. Against this backdrop, a possible $18 billion valuation looks like investors betting not only on current revenue but on future demand from model developers and corporate customers.

The company's financial trajectory is indeed steep. According to available data, FluidStack's revenue rose from $1.8 million in 2022 to $66.2 million in 2024. That jump shows how fast an infrastructure business can grow when it hits the right spot in the market. For a startup in a capital-intensive niche, such growth matters especially: data centers, accelerators, networking equipment and power supply require large upfront investment, so investors typically look not only at current sales but at the team's ability to quickly fill new capacity with paying demand.

A partnership with Anthropic on $50 billion worth of data centers gave the company additional momentum. It was after this deal that FluidStack moved the center of gravity of its business from the UK to the US. The move looks logical: the largest AI orders, chip suppliers, financial partners and customers ready to reserve compute years ahead are today concentrated largely in the American market.

It is also reported that Jane Street and Situational Awareness are discussing the possibility of leading the new round. If these talks succeed, it will be more than a capital raise: it would fund faster build-out of sites, hardware purchases and the locking in of large contracts in the race for AI capacity.

The size of the proposed round is telling too. In AI infrastructure, hundreds of millions are no longer exceptional, because this business almost never grows organically: to take on new customers, a company must finance in advance construction, power hookups, server racks and access to accelerators. A $1 billion round gives FluidStack room for aggressive expansion, but it also sets a very high entry price for competitors. In effect, investors are paying for a place in a segment where the winner is determined not only by technology but by the speed of deploying physical capacity.

That is why capital in this segment works almost like a product in itself: whoever raises money faster reserves hardware, sites and electrical capacity sooner. For customers, this means more predictable access to compute; for the company, a chance to establish itself before the market starts to consolidate.

The main takeaway is that the market is again confirming a simple point: in AI, what is valued most is not the interface but scarce infrastructure. As long as demand for training and serving models grows faster than new capacity comes online, companies like FluidStack will command a premium in both valuation and investor attention. But a high valuation also raises the bar: the company will now be expected not just to grow fast, but to turn the AI hype into a durable, scalable and profitable infrastructure business.