Nebius rents out online computing power, specifically designed to train and run artificial-intelligence models. Concretely, the company owns and operates data centers filled with graphics processors, the chips able to run the massive, repetitive calculations that model training requires, and it rents access to them to companies that have neither the time nor the means to build such facilities. Born from the spin-off of a large technology group, it is listed in the United States but keeps its European roots. It positions itself as an alternative to the American giants of computing rental, in a market where demand far exceeds the available supply of graphics processors. Its customers are model developers, start-ups and companies seeking quickly deployable capacity. Its challenge is to grow fast enough to fund the purchase of costly, scarce chips.
Training an artificial-intelligence model requires entire rows of graphics processors working together for weeks, a power few companies can own outright. Renting it on demand has therefore become a strategic market, today dominated by a handful of American giants. Nebius slipped in as a younger player, aiming to offer a credible alternative, notably for customers attached to a European anchoring. Its strength is a clear position in a market where demand exceeds supply, and end-to-end control of its own data centers. Its fragility is structural and serious: this business requires constantly buying very expensive, scarce chips, which means massive spending before revenue even comes in. The company faces competitors with incomparable means, and its future profitability depends on sustained growth that is not guaranteed. Some recent financial data remains to be confirmed and should be verified before acting.
Nebius holds a central link in the Compute Economy: renting online computing power built for AI. Its strength is a clear position in a supply-constrained market and control of its data centers; its fragility is the need to keep buying costly chips, against giants, with future profitability unguaranteed. Verify before acting.
Everything you need to know about NeonBridge and the compute economy.
Oil powered the 20th century. Compute powers the 21st. Every time an AI model runs or a Bitcoin block is mined, it takes energy, chips, data centers, and cloud infrastructure to make it happen. The companies building all of that form the compute economy. NeonBridge tracks 200+ of them across 7 categories.
Just like oil or electricity, compute is a raw resource that every industry needs. AI models can't train without GPU cycles. Bitcoin can't exist without hashrate. As demand grows, the companies that produce, store, and distribute compute become critical infrastructure. That makes compute the defining commodity of this century.
AI and Bitcoin are not two separate topics. They are two expressions of the same industrial revolution. AI transforms compute into intelligence. Bitcoin transforms compute into verifiable scarcity. Both need the same physical foundations: energy, chips, and data centers. That is why NeonBridge tracks them together, as one economy.
Most companies in the tracker are publicly listed stocks you can buy through any brokerage account. We highlight EU-friendly brokers like Trade Republic, Interactive Brokers, DEGIRO, and Scalable Capital to help you get started. No crypto wallet needed for the equity side.
Like any sector, compute infrastructure carries risk. Chip supply chains can be disrupted by geopolitics. Energy costs fluctuate. AI regulation is evolving fast. Bitcoin mining profitability depends on network difficulty and price cycles. Diversifying across the 7 categories of the compute economy helps reduce exposure to any single risk.
Global AI compute capacity doubles every 6-7 months, fueled by explosive AI adoption outpacing historical trends, with contributions from Bitcoin mining infrastructure repurposing and autonomous systems. Governments worldwide have committed hundreds of billions to chip manufacturing and power infrastructure. This reflects an enduring industrial shift comparable to electrification.
