Verda AI Cloud Funding Round Hits $189M as Shortage Drives V

Verda, the Helsinki-based GPU cloud provider previously known as DataCrunch, has raised $189 million in a Series B led by Emergence Capital, pushing its valuation past $1 billion. The Verda AI cloud funding round, announced on September 22, 2026, is the largest the Finnish company has closed and makes it one of Europe's newest AI infrastructure unicorns. Verda confirmed the billion-dollar threshold but declined to disclose an exact valuation.

The company rents out GPU capacity for AI training and inference, competing for the same accelerator-hungry workloads that AWS, Microsoft Azure and Google Cloud are chasing. Verda states that it runs renewable-powered infrastructure across Europe, the UK, the United States and Asia, advertises savings of as much as 90% against hyperscaler pricing, and reports an annualised revenue run rate of $165 million.

Those figures describe a business growing fast. They do not yet describe a business that is difficult to copy, and that gap is the real subject of the round.

A Funding Market Priced on Shortage

Verda is one of the providers the industry calls neoclouds: companies that buy accelerators, build data centre capacity and sell it by the hour. The category exists because hyperscalers have not added capacity quickly enough to meet demand. Cloud providers have turned away AI and cloud business for want of available GPUs, which gives challengers pricing power they would not normally hold against three of the largest infrastructure operators in the world.

That mechanism is what the Verda AI cloud funding round is buying into. Capital is flowing to any operator that can put working accelerators in front of paying customers, because the alternative for those customers is a waiting list. Verda's $189 million is a supply-side bet: the money converts into hardware, power contracts and data centre space, and revenue follows only if the machines stay busy.

The reversal in investor appetite is recent. GPU cloud was long treated as a capital-intensive business with heavy depreciation and thin margins, a profile venture investors avoided in favour of software. Sold-out capacity changed the arithmetic: when compute is committed before it is built, the risk of owning depreciating hardware sits with whoever is waiting in the queue.

Scarcity is temporary by construction. Every dollar raised by Verda and its peers is a dollar spent dismantling the constraint that justified the raise. Capacity shortages end once enough capital arrives to end them, and this round is priced against conditions the round itself is helping to erode.

The Numbers Behind the Verda AI Cloud Funding Round

The disclosed details are narrow:

Round size: $189 million, equivalent to roughly €164.8 million
Lead investor: Emergence Capital
Structure: Series B, described as oversubscribed
Valuation: at least $1 billion, exact figure undisclosed
Headquarters: Helsinki, Finland; the company was formerly DataCrunch
Operating metrics cited by the company: $165 million annualised revenue run rate, and savings of up to 90% versus hyperscaler equivalents

Against the $165 million run rate, a valuation at or just above $1 billion implies a multiple near six times revenue. For a business that must continuously buy depreciating hardware, that multiple looks reasonable rather than aggressive. The arithmetic holds if Verda keeps its GPUs utilised at close to capacity, and deteriorates fast if utilisation slips, because the assets lose value whether or not they are rented.

Verda has not published a consolidated financing table, and it declined to disclose an exact valuation. The round size is reported as $189 million, equivalent to roughly €164.8 million. Any revenue multiple derived from the disclosed figures therefore carries a wide error bar.

What the Verda AI Cloud Funding Round Does Not Buy

The most quoted number in Verda's pitch is price: up to 90% below hyperscaler equivalents. That comparison is more fragile than it looks, because GPU cloud pricing depends on what is bundled with the compute.

A headline hourly rate says little about storage, network egress, managed orchestration, support tiers or the generation of accelerator being priced. Hyperscalers sell integrated platforms with compliance certifications, global networking and enterprise contracts attached, and they can be beaten on a compute line item by a specialist offering less around it. Verda's discount is largest for customers who need raw capacity and already own the tooling to use it, and narrowest for those who want the platform.

Hardware vintage is the second factor behind the price claim. Operators competing on price typically do so with accelerator generations that hyperscalers have moved past, which is efficient for inference and fine-tuning and less so for frontier training runs. That positioning is defensible and durable, and it also caps the addressable market at customers whose workloads fit older silicon.

Residency Guarantees Versus Hyperscaler Regions

The part of Verda's pitch that is genuinely hard to copy is location. European customers in regulated industries need compute that stays inside named jurisdictions, with clear answers about who can access data and under which legal regime. A Helsinki operator with Nordic power and EU legal standing answers that requirement directly.

That advantage has a shelf life set by how fast hyperscalers build the same thing. Sovereign cloud programmes from the major providers already target European public sector and regulated workloads, and every region they open removes a reason to choose a smaller supplier. Verda's argument has to be that it reaches a given region sooner, at lower cost, or on better contractual terms than a hyperscaler will. Once the hyperscaler region exists, the argument narrows to price and service, which is the contest Verda is less equipped to win.

Verda's stated expansion across the UK, the United States and Asia points the same way. Those markets are not primarily residency plays, and they are where AWS, Azure and Google Cloud hold their deepest benches. Growth there tests the platform rather than the jurisdiction.

The Trade-Offs a Buyer Faces

For a CTO deciding where to place a training or inference bill, the round cuts both ways. A well-funded challenger is a safer counterparty than an undercapitalised one; $189 million buys runway, hardware and the credibility to sign multi-year contracts. The same round also signals that the price advantage on offer depends on a market condition its own spending is designed to end.

The practical hedge is commitment length. Short reservations against specialist capacity capture the current discount without locking a company into a supplier whose differentiation may be gone in two years. Long commitments make sense where the residency requirement is contractual and specific, or where the workload is stable enough that a 90% compute saving outweighs the cost of rebuilding tooling on another platform.

The signals worth tracking are concrete: how quickly Verda converts the round into live capacity, whether its stated footprint expands or thins out, and whether hyperscalers announce European regions overlapping Verda's strongest markets. Each of those moves pushes the question from availability back to price, which is where the incumbents prefer to compete.

Why this matters

Verda's raise shows what AI infrastructure capital is buying right now: guaranteed access to accelerators while rivals queue for them. That trade holds only while the queue exists, which makes today's discount pricing a window rather than a new baseline. Buyers should treat challenger economics as time-limited, and Verda's next two years will turn on whether it converts shortage-driven customers into ones that stay for reasons other than availability.

✔Human Verified

Researched and cross-referenced against primary sources by the Bytevyte editorial team. This article was generated with the assistance of artificial intelligence and reviewed by the Bytevyte editorial team.