AI industry faces $3 trillion return on investment challenge

The technology sector is grappling with the financial justification of its substantial investments in artificial intelligence (AI). TechCrunch reports on the widening gap between the capital poured into AI infrastructure and the actual economic returns generated, posing a critical question for the industry's future.

David Cahn, a partner at Sequoia Capital, initially calculated in 2023 that $200 billion in annual revenue was needed to cover AI infrastructure costs, particularly GPU acquisitions and operational expenses. By 2026, Cahn's updated estimate for AI infrastructure spending reaches $1.5 trillion. This necessitates a total industry revenue of $3 trillion to validate these investments.

These figures may even be conservative, as rising costs for memory and specialized inference chips, alongside construction bottlenecks, are increasing the required revenue per installed gigawatt. The industry must find ways to accelerate revenue generation to keep pace with escalating expenses.

While some AI companies, such as Anthropic and OpenAI, have achieved notable revenue figures – Anthropic reportedly reaching $60 billion in ARR and OpenAI reporting $13 billion for 2025 (though it later stated $20 billion ARR in November 2025), with alleged data center commitments of $1.4 trillion – a significant disparity between investments and profits persists.

Financial analysts are also monitoring the situation. Torsten Slok, chief economist at Apollo Asset Management, suggests that AI's return on investment may materialize more slowly than anticipated, underscoring the need for strategies to enhance profitability and demonstrate concrete economic value.