Fluidstack company profile: valuation, funding & revenue

This Fluidstack company profile covers the company's valuation, revenue, funding rounds, investors and history. Fluidstack works in AI, with a focus on cloud computing infrastructure. It was founded in 2017 and is headquartered in London, United Kingdom. Its latest valuation on record is US$7.5 billion (January 2026). Every figure links to a public source, and anything the company hasn't disclosed is marked as such.

Revenue

Funding history

Fluidstack timeline

Key dates from the sources on this page, oldest first. Metric snapshots are listed after the funding events.

• 2017 — Fluidstack is founded in London by César Maklary (CEO) and Gary Wu.
• Jan 2026 — Series A: US$830M, valuing the company at US$7.5B. Investors: Situational Awareness (lead).

Investors

Overview

Fluidstack provides large-scale computing infrastructure for artificial intelligence. It builds and operates clusters of graphics processing units (GPUs) in data centres and rents them to AI labs and enterprises that need to train and run large models. Companies like Fluidstack are often called "neoclouds": specialised cloud providers focused on AI computing, competing with traditional hyperscalers such as Amazon Web Services, Microsoft Azure and Google Cloud. Fluidstack has become known for delivering very large clusters quickly, and for its role in major data centre projects for leading AI developers.

Fluidstack was founded in 2017 by César Maklary and Gary Wu, who met as students at the University of Oxford, and is headquartered in London. It announced an US$830 million Series A at a US$7.5 billion valuation, led by Situational Awareness, the AI-focused investment fund founded by Leopold Aschenbrenner, with the round closing in January 2026, according to the company and its legal adviser Mishcon de Reya. In April 2026 Bloomberg reported the company was in talks to raise about US$1 billion at an US$18 billion valuation; we treat that as reported talks until confirmed.

Products and services

• Dedicated GPU clusters: large clusters reserved for a single customer, often for training frontier models.
• On-demand GPUs: cloud access to GPUs for smaller workloads.
• Managed infrastructure: networking, storage, orchestration and support for AI workloads.
• Data centre development: partnerships to build new AI data centre capacity in the US and Europe.

Funding

For much of its history Fluidstack grew with relatively little venture capital, financing GPUs largely through debt and customer contracts. Its Series A of US$830 million, led by Situational Awareness, valued it at US$7.5 billion. Earlier reports had suggested a raise of about US$700 million at around US$7 billion; the final round was larger. GPU-backed debt facilities provide much of the capital needed to buy hardware.

Business model

Fluidstack earns revenue by renting computing capacity, often under multi-year contracts with AI labs and large companies. It buys or leases GPUs and data centre space, financed with a combination of equity, debt and customer prepayments, and earns a margin on the difference between its costs and what customers pay. Long-term contracts with creditworthy customers make it easier to borrow against the hardware.

Competition

Fluidstack competes with neoclouds such as CoreWeave, Lambda, Crusoe, Nscale and Together AI's infrastructure, as well as with hyperscalers. Customers often choose providers based on availability of the latest GPUs, cluster size, speed of delivery, reliability and price.

Fluidstack at a glance

Fluidstack is a London AI cloud company founded in 2017 by César Maklary and Gary Wu. It builds and operates large GPU clusters for AI labs and enterprises. It raised an US$830 million Series A led by Situational Awareness at a US$7.5 billion valuation, closing in January 2026, and was reported in April 2026 to be in talks at an US$18 billion valuation.

Timeline

• 2017: founded by Oxford students César Maklary and Gary Wu.
• 2023–2024: grows rapidly supplying GPU capacity to AI companies.
• 2025: announced as a partner in large AI data centre projects in the US and Europe.
• January 2026: closes an US$830 million Series A at US$7.5 billion.
• April 2026: reported to be in talks to raise about US$1 billion at US$18 billion.

Why neoclouds emerged

The boom in generative AI created enormous demand for GPUs, particularly Nvidia's data centre chips. Hyperscalers could not supply enough capacity quickly, and AI labs wanted large, dedicated clusters configured for training. Neoclouds filled the gap, buying GPUs, securing data centre space and power, and offering capacity faster and sometimes cheaper. Nvidia has supported neoclouds by allocating chips to them, which diversifies its customer base beyond the hyperscalers that are also developing their own chips.

The financing model

AI clusters are extremely expensive: a single large cluster can cost billions of dollars. Neoclouds finance them with debt secured on GPUs and on customer contracts. This allows rapid growth but creates risk: if customers cancel or GPU values fall faster than expected, lenders and equity holders could suffer losses. Investors watch customer concentration, contract length and the pace of hardware obsolescence closely.

Who funds Fluidstack

Fluidstack's Series A was led by Situational Awareness. Lenders have provided GPU-backed debt facilities.

Valuation context

A US$7.5 billion valuation at Series A, and reported talks at US$18 billion months later, reflect the extraordinary demand for AI computing. Our guides to AI cloud providers, AI infrastructure companies and the largest UK startups compare Fluidstack with peers.

Risks

• Customer concentration: a few AI labs may account for most revenue.
• Debt and leverage: GPU financing creates significant obligations.
• Hardware obsolescence: new chip generations reduce the value of older GPUs.
• Power constraints: data centre growth depends on securing electricity.
• Reported talks: the US$18 billion figure was unconfirmed in our sources.

Questions for customers

• Which GPU generations and cluster sizes are available, and when?
• What uptime and support commitments are offered?
• Where are the data centres, and how is power secured?
• How flexible are contract terms?

Building AI capacity at speed

Delivering a large AI cluster involves securing GPUs from Nvidia or its partners, finding data centre space with enough power and cooling, installing high-speed networking that connects thousands of chips, and running software that keeps the cluster healthy. Delays at any step can cost AI labs months in a race where timing matters. Fluidstack's reputation rests on assembling these pieces quickly by working with data centre developers, hardware suppliers and financiers. It typically does not own the buildings itself but partners with developers that provide powered space, while Fluidstack installs and operates the computing.

This partnership model allows faster expansion than building data centres from scratch, but it requires strong relationships and careful coordination. As AI labs plan clusters requiring hundreds of megawatts or even gigawatts of power, the ability to deliver at that scale becomes a key competitive factor among neoclouds.

Europe's AI infrastructure ambitions

European governments have sought to build domestic AI computing capacity to reduce reliance on US providers and to support local AI companies. Fluidstack, as a London-based provider, has been involved in European infrastructure discussions, including large projects in France announced around the 2025 AI Action Summit in Paris. Such projects can bring public support and long-term demand, but they also involve political considerations and long planning timelines. Fluidstack's mix of US and European projects gives it exposure to both markets.

What to watch

Watch confirmation of its next round, new data centre projects, customer diversification and debt financing.

Takeaways

AI developers can access large GPU clusters quickly through neoclouds. Founders can see how a long-running infrastructure business can be transformed by a market shift. Investors should watch leverage and customer concentration.

Sources

Every figure on this page comes from the public sources below. Where sources disagreed, we used the company's own announcement or a primary news report. Private-company data changes; check the date beside each figure.

1. Mishcon de Reya
2. Fluidstack