This week, market participants are focusing on the anomalous trading volume in Nebius (NBIS), an artificial intelligence (AI) infrastructure company based in Amsterdam, the Netherlands.
Against the backdrop of giants like Micron Technology (MU) and Nvidia (NVDA), NBIS stock has suddenly burst toward the top in terms of trading volume.
This raises a question: What is forcing investors to aggressively revalue a company which, until recently, was known only to a small circle of specialists?
The Essence of the Phenomenon
The essence of this phenomenon is hidden not in a quantity of purchased chips or even software.
Instead, the answer has to do with regional arbitrage and compute.
Here's why Nebius has turned out to be in the right place at the right time — and why the value of NBIS stock is being determined not only by technology but also geography.
A New Giant in the Shadow of Nvidia
When we speak 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 has clearly split into "universal clouds" and "AI factories."
Today, Nebius is a main representative of this second category.
After a massive transformation and listing on the Nasdaq Exchange, the company ceased to be an heir to technological assets of the past.
Now, it is a "clean bet" on AI infrastructure.
But unlike U. S.
competitors, Nebius builds its empire where competition is minimal — and where hunger for capacity is at a max.
Recent reports and current trading volumes confirm as much.
The market has started to realize that Nebius is a physical bridge between the scarce silicon of Nvidia and the European market, which for years has been starved in terms of super-powerful local compute.
To understand the value of Nebius, you need to throw away old ideas about data centers.
Previously, the distributed cloud worked like this: Thousands of servers in different cities, connected by ordinary internet, perfectly coped with the delivery of mail or video streaming.
For AI, this model is dead.
