The risky dance
Today, capital flows toward digital infrastructure at the speed of light. Driven by the AI hype, Wall Street and Private Equity (PE) are pumping billions into a sector previously viewed as dull real estate: data centers.
However, beneath the concrete and cooling systems bubbles a fundamental question: will the return on investment (ROI) ever catch up to the hype, or will this system eventually crack badly?
Does the return follow the hype?
The core of the current overheating lies in a capital mismatch. While PE giants like Blackstone and KKR invest billions in physical data centers, the ball is in the court of their tenants: the Big Tech companies. They must prove that the end user, from the local baker to the multinational, is willing to pay for AI services to justify those gigantic rents.
The numbers are staggering. Capital expenditures (CAPEX) for data centers have exploded due to the need for specialized liquid cooling and massive power supplies for heavy Nvidia GPUs. However, the great uncertainty lies in operational costs (OPEX). Experts are extremely cautious about the margins on AI. There is a dangerous gap between the American construction frenzy and actual market adoption. While Wall Street expects linear growth, practice shows a complex stagnation.
Reality
A dangerous gap is widening between the construction boom and actual market adoption. While Wall Street anticipates linear growth, what we are seeing in practice is a complex stagnation:
The Financial & Economic Reality
- Negative ROI: Data centers depreciate faster than they generate revenue. Much hardware is already outdated before the high development costs are recouped.
- Existential anxiety: Billions are being "burned" out of fear of losing the AI race to competitors, regardless of current returns.
- Market valuation: The global data center market is expected to grow from $92 billion in 2024 to double that size by 2030. In the long term, analysts even estimate total spending within the broader AI ecosystem to reach into the trillions.
- Private Equity role: PE funds are no longer just investing in the bricks, but in the entire chain (energy, cooling, software) and are effectively acting as project developers for Big Tech.
Social & Ecological Impact
Tax loss: At least ten American states miss out on $100 million in yearly revenue each as a result of aggressive tax incentives for data center developers.
Power demand: In tech hubs like Phoenix (USA), power demand is increasing by as much as 500%. Power grids sometimes need to triple to meet the demand.
Water consumption: One data center of 100 megawatts (MW) consumes an average of 2 million liters of water per day for cooling, which is equivalent to the consumption of 6,500 households.
Local resistance: Noise pollution from cooling installations and the heavy burden on the public power grid are causing increasing conflicts with local residents.
Geopolitics
Simple use: The general public is currently using AI mainly for simple text and productivity tasks. The deep, profitable integration within business processes still faces significant human, technical, and legal barriers.
European Sovereignty: Under pressure from the EU AI Act and strict GDPR regulations, an increasing number of European companies are rethinking their position in the US cloud. The demand for local, independent solutions is rising.
Difference between the US and Europe: In the US, the priority is purely on power availability (often via gas), while European policy forces developers to use green power and mandates the reuse of waste heat.
The Power Hunger
The battle in the data center market is no longer fought over bricks, but over power rights. Investing in locations has changed to land & power banking. Because the waiting times for grid connections in America have now increased to 5 to 8 years, PE parties buy land purely and solely because of the allocated power rights.
Whoever controls the power, owns the market.
To financially bridge the long, unproductive construction periods, Private Equity employs three smart strategies:
Mark-to-Market revaluation: Once the power commitment (Grid Allocation) is finalized, an interim revaluation of the land takes place. The enormous increase in value raises the net asset value (NAV) of the fund, which allows for favorable refinancing to free up capital for new projects.
Credit-Based Financing (Pre-leasing): Binding lease agreements are signed with so-called Hyperscalers (such as Microsoft, Google, or AWS) even before the first stake is driven into the ground. The PE party uses the strong creditworthiness of these tech giants as collateral for cheap project financing.
Yield-Exit: PE parties focus on the risky development phase where the highest margins lie. Once the data center is operational and generating stable income, the fund sells the project to the 'yield market' (such as pension funds or Infrastructure REITs) that are satisfied with lower, inflation-resistant returns.
American overkill
This dangerous dance between Private Equity, Wall Street, and Big Tech is a gigantic gamble. The business model relies entirely on the assumption that the scarcity of power rises faster than the interest rate. In the worst-case scenario where the commercial ROI of AI for Big Tech fails to materialize, the pre-lease contracts will disappear and the market will collapse.
What happens on the other side of the ocean almost always impacts the European economy. If the American AI bubble bursts, the shockwaves will be felt all the way to Flanders.
Fortunately, Europe is choosing a different path. With strict legislation and a focus on human values and sustainability, we protect our market. To truly be part of the AI revolution without the risks of American overkill, European companies must reclaim their technological independence. Building on our own, local, and sovereign infrastructure is no longer a defensive choice, but the smart growth strategy for the winners of the future.
Source/ Morgan, Lewis & Bockius LLP - The newyork times