Annual AI Infrastructure Bill Could Reach $1.5 Trillion

    According to research by “Bain & Company,” annual spending on artificial intelligence infrastructure could reach $1.5 trillion in 2031. This is approximately AZN 2.55 trillion. The spending includes data centers, processors, memory, and networking equipment. For investment on this scale to continue, the industry’s annual revenue would need to rise to approximately $6 trillion. In manat terms, this amount is AZN 10.2 trillion. The calculation is based on the assumption that computing infrastructure costs will equal 25% of revenue.    Approximately $4.2 trillion of the required revenue is expected to come from search, advertising, autonomous systems, and artificial intelligence products running on physical devices. This amount is equivalent to AZN 7.14 trillion. Revenue from tools that improve companies’ work efficiency could rise from $1 trillion to $1.4 trillion. In manat terms, this figure ranges from AZN 1.7 trillion to AZN 2.38 trillion. Revenue from consumer subscriptions and advertising is estimated at between $200 billion and $400 billion. This means between AZN 340 billion and AZN 680 billion.    Leading artificial intelligence laboratories are spending more than $9.75 billion on engineering work to help companies benefit from new tools more quickly. This is approximately AZN 16.6 billion. Future products could also cover fields such as drug development, mental health, and energy production. However, artificial intelligence is not yet widely used in these areas. The research shows that improvements in work efficiency alone will not be enough to cover the enormous infrastructure costs.    The size and construction costs of data centers worldwide are increasing 2-fold approximately every 12-16 months. These facilities require additional electricity generation, grid connections, advanced chips, specialized equipment, and skilled workers. As companies seek to obtain more computing power, expanding data centers is becoming the industry’s main financial burden. The key question is whether these projects will generate enough economic value to justify their costs.    According to calculations by “Epoch AI,” the capacity of Meta’s “Prometheus” center in Ohio was 600 MW in 2025. At that time, the facility was valued at $24 billion. This is AZN 40.8 billion. According to the forecast, the center’s capacity will reach 2 GW in 2027. The project’s total cost is expected to rise to $80 billion. In manat terms, this amount is AZN 136 billion.    According to the forecast, the capacity of the “Prometheus” center could reach 5 GW in 2029. At that point, the project could be worth $175 billion. This is AZN 297.5 billion. The center’s capacity is expected to reach 9 GW in 2030. The total cost could also rise to as much as $200 billion. In manat terms, this figure is AZN 340 billion. For projects like these to continue, artificial intelligence products will need to generate significantly more revenue than they do now.