Trillions of Dollars at Risk in AI Sector as Tech Giants Face Potential Financial Crisis

Whether the trillions of dollars invested in artificial intelligence will pay off in the future has become a subject of serious debate. Although technology companies are spending record amounts on data centers and powerful processors, it is still unclear where the revenue needed to recoup these investments will come from.

Ayşegül İldeniz, who held executive positions at Intel and in Silicon Valley for many years, has warned that an economic bubble could form in the artificial intelligence sector. According to her, companies need to generate $1.4 trillion in annual revenue by 2030 to justify their investments.

Bain & Company’s 2026 Technology Report, meanwhile, highlights even larger figures. According to the company’s forecast, annual spending on artificial intelligence infrastructure could reach $1.5 trillion in 2031. For these investments to be economically sustainable, the sector may need to generate approximately $6 trillion in annual revenue.

Companies such as Microsoft, Google, Amazon, Meta, and Oracle are already allocating hundreds of billions of dollars to this race. According to Bain’s estimates, their total capital expenditures could reach $780 billion in 2026. That is approximately five times the figure from just three years earlier.

The main problem is that existing artificial intelligence services do not generate enough revenue to cover such large investments. According to Bain’s forecast, consumer subscriptions and advertising could generate $200–400 billion in revenue and economic benefits in 2031, while enterprise artificial intelligence applications could generate $1–1.4 trillion. Based on these estimates, at least $4.2 trillion in additional economic value must be created to reach the $6 trillion target.

Companies are looking to new areas to close this gap. Artificial intelligence-powered search engines and advertising could create an additional market opportunity of $100–200 billion, while autonomous transportation and industrial systems could create approximately $400 billion in additional market opportunities.

Another promising area is physical artificial intelligence. Humanoid robots, digital twins, and smart manufacturing systems could reduce costs in the automotive, electronics, and other industries. Bain estimates that a 10 percent reduction in manufacturing and research costs in certain sectors would create a $900 billion economic opportunity.

However, alongside financial risks, technical constraints are also growing. Demand for artificial intelligence processors from companies such as Nvidia and AMD is increasing rapidly. Dependence on TSMC for advanced chip manufacturing and on manufacturers such as SK Hynix, Samsung, and Micron for high-speed HBM memory is creating serious pressure on the supply chain.

According to TrendForce’s forecast, the average selling prices of HBM memory could rise 121 percent year over year in 2027. This could make building already expensive artificial intelligence infrastructure even more difficult.

The scale of data centers is also growing rapidly. According to Epoch AI’s estimates, the capacity and cost of the world’s largest artificial intelligence data centers approximately double every 12–16 months.

One example is Meta’s Prometheus project in the U.S. state of Ohio. According to forecasts, the data center’s total cost could approach $80 billion in 2027, $175 billion in 2029, and $200 billion in 2030.

The energy demand of these enormous facilities has also become a separate problem. Data centers’ share of electricity consumption in the U.S. state of Virginia is reported to have exceeded 28 percent. Growing energy demand is pushing technology companies toward nuclear energy agreements and dedicated power-generation projects.

Because of existing technical and energy constraints, it is forecast that 30–50 percent of the capacity of large data centers scheduled to begin operating in 2026 could be delayed.

Nevertheless, large investments do not necessarily mean that an economic bubble has formed in the artificial intelligence sector. The main risk is not that the technology will fail, but that the revenue it generates will not grow as quickly as the costs.

Technologies such as the internet and railroads have also created major waves of investment throughout history. Although these sectors transformed the economy, not every company that invested in them made a profit.

A similar situation could arise in the artificial intelligence sector. Even if the technology becomes widespread and increases productivity across the economy, there is no guarantee that all of today’s trillions of dollars in investments will be profitable. The main question remains unchanged: Will artificial intelligence companies be able to generate enough revenue to finance the enormous infrastructure they are building?