This week, market participants focused on the unusual trading volume in Nebius (NBIS), an artificial intelligence (AI) infrastructure company based in Amsterdam, Netherlands. Against a backdrop of giants like Micron Technology (MU) and Nvidia (NVDA), NBIS stock suddenly exploded to record highs in terms of trading volume. This raises the question: What forces investors to review the company, until recently, was known to a small circle of experts?
The essence of this situation is not hidden in the number of purchased chips or software. Instead, the answer has to do with regional arbitrage and calculation. Here’s why Nebius turned out to be in the right place at the right time – and why the value of NBIS stock is determined not only by technology but also by country.
More news from Barchart
The New Giant in Nvidia’s Shadow
When we talk about cloud computing, Amazon’s ( AMZN ) AWS, Microsoft ( MSFT ) Azure, and Alphabet’s ( GOOGL ) Google Cloud come to mind. However, in 2026, the AI infrastructure market is clearly divided into “global clouds” and “AI factories.”
Today, Nebius is the main representative of this second category. After a major change and listing on the Nasdaq Exchange, the company ceased to be the heir of the technological assets of the past. Now, it’s a “clean bet” on AI infrastructure. But unlike its US competitors, Nebius is building its empire where competition is thin – and where the hunger for capacity is at its peak.
Recent reports and current trading volumes confirm that. The market has begun to realize that Nebius is a virtual bridge between Nvidia’s rare silicon and the European market, which has been starving for a super-powerful local computer for years.
The Physics of AI: Why ‘Just a Data Center’ Doesn’t Work Anymore
To understand the value of Nebius, you need to discard old ideas about data centers. Previously, the distributed cloud worked like this: Thousands of servers in different cities, connected by the common Internet, efficiently cope with the delivery of mail or video streaming. In AI, this model is dead.
Concentration versus distribution requires that, in order to train modern large-scale linguistic models (LLMs), thousands of GPUs need to work as one. If your chips are located in different halls or connected by slow cables, delays in data transfer are often produced. The AI industry is a monolith. Tens of thousands of Nvidia Blackwell accelerators are literally installed in one room and connected by super-fast InfiniBand.
The latency factor plays an important role. For the end user – be it a bank in Paris or a transport center in Berlin – it is important how quickly an AI agent can provide a response. When a request crosses the Atlantic to a US data center and back, the latency becomes critical. Because in the era of “agent AI” – where systems are forced to make decisions in milliseconds – the physical proximity of servers becomes as important an asset as the AI model itself.
Nebius doesn’t just rent racks. The company is building what it calls a Full-Stack AI Cloud. Its network architecture has been optimized to minimize the high number of Nvidia chips. In this regard, the company is closer to the supercomputer manufacturers than to the old hosting providers.
Regional Fractures: Why Nebius’ Connection with Europe Mattered
This is the most important point of our analysis. If Nebius had built its AI infrastructure in Texas or Virginia, it would have been one of many firms. In the United States, the market is overcrowded, from Microsoft, Google, and Oracle (ORCL) to niche players like CoreWeave (CRWV) and Lambda.
However, Europe ended up in a vacuum.
Here, we can see the lack of infrastructure. More than 500 million solvent consumers live in Europe and thousands of large companies operate there, but there are large concentrated AI clusters. US giants are building “data warehouses” in Europe, but not AI factories.
Sovereign AI and law sets strict rules. European Union (EU) regulators are increasingly requiring European data to be processed and stored within the EU. A company in the EU cannot simply send confidential client data to a data center in Ohio to be processed by a neural network without meeting strict guidelines. That has increased the need for a powerful “local brain”.
This type of property interest provides special control. The first to build a large piece of hardware in Finland or France automatically becomes the leader of those in Europe who need quick answers and compliance with the law.
Therefore, Nebius plays on regional inefficiencies. The company has created assets where demand is high and supply from US hyperscalers is physically and legally limited. This makes Nebius’s position unique.
Why Meta Platforms and Microsoft Pay Nebius?
Microsoft and Meta Platforms (META) both have AI infrastructure agreements with Nebius. At first glance, this may seem absurd. After all, Microsoft has a market capitalization of over $3 trillion, while Meta has a market cap of over $1.5 trillion. These companies have in-house construction divisions and direct contracts and services. Why would Big Tech giants lease capacity from a small European player with a market cap of $55 billion?
The answer is time.
In the United States, it is possible to build a data center quickly compared to Europe, where obtaining land permits, energy restrictions, and environmental certifications can take years. Big Tech doesn’t have this time. The AI arms race is happening right now. Nebius soaked up the infrastructure years ago, and it offers the right solution here and now.
Europe’s electricity grids are overloaded. Finding 200 to 300 MW of free energy for a single cluster in Germany or the Netherlands today is almost impossible. Nebius has retained these capabilities previously, such as its locations in Finland and France. These Big Tech giants not only rent servers from Nebius, but access to the European power grid.
The flexibility of the “pure player” allows them to move quickly. Microsoft Azure is forced to support a large number of legacy enterprise customers. Nebius, however, is building an infrastructure from scratch specifically for AI. This allows it to implement new cooling and power architectures quickly, which Big Tech firms can implement in their old data centers immediately.
That’s why we’re seeing $27 billion in contracts from Meta and up to $19.4 billion from Microsoft. For these companies, this is not just outsourcing. It is a time to buy and an important topic in the rich European market.
Strategic Alliance: Nvidia is Nebius’ Guardian Angel
It is impossible to discuss Nebius without mentioning Nvidia. In March 2026, Nvidia invested 2 billion dollars in Nebius, receiving a share in the capital. This deal legitimizes the company in the eyes of institutional investors.
What does this mean for business?
Well, the priority is on the line. With the world lining up for the next Blackwell or Rubin chips, Nebius will likely be among the first to get them. On the other hand, it is beneficial for Nvidia to have a strong independent partner of Amazon, Microsoft, and Google. This is a variation of Nvidia CEO Jensen Huang’s sales pitch.
Nebius often serves as the first place to install new Nvidia reference architectures in Europe. This creates a closed cycle – Nvidia offers the best chips, Nebius builds the best clusters, and then attracts the best customers.
Nebius’ $50 Billion Backlog
The figures from Nebius’ report for the first quarter of 2026 are thought-provoking, but investors need to know how to read them correctly.
Nebius reported revenue of $399 million in Q1, marking a whopping 684% year-over-year (YOY) increase. But the average total of Nebius’ contract backlog is approaching $50 billion. This is the amount that customers commit to paying over the next three to five years. Such cash flow forecasting is an unusual gift for a technology company.
Capex is what scares the bears the most. Nebius plans to spend $20 billion to $25 billion by 2026 to purchase equipment and construction over the next two years. Yes, this is a lot of money. But it is important to understand that this is not spending in the hope that customers will come, but increasing the production of certain customers.
The market does not inform Nebius of the current profit, but of what share of the European “AI pie” it will own in 2028.
The conclusion
The value of Nebius does not lie in the fact that the company has invented a magic chip. Rather, its strength is based on impeccable timing and local advantage.
Nebius quickly understood that AI is not just about algorithms, but also about physical infrastructure that must be embedded, localized, and managed. In the United States, its model can be one of many. But in Europe, the company has become needed by manyfrom start-ups to international companies.
Nebius is a bet on “Real Estate 2.0.” In the 20th century, the capital was created by having railroads and ports. Now, it is done by having large nodes where the “digital heart” of Europe beats. The trading volume only shows how the market is trying to appreciate this new type of strategic resource.
We don’t know where NBIS stock will trade in a month. But we know that AI factories like the one Nebius operates are essential to Europe’s AI empire.
As of the date of publication, Mikhail Fedorov did not have (directly or indirectly) positions in any of the securities mentioned in this article. All information and data in this article is for informational purposes only. This article was originally published on Barchart.com
