BCG Institute: By 2030, the AI ecosystem will require $2.6 trillion in annual spending

According to BCG Institute, by 2030, the AI ecosystem will require about $2.6 trillion in annual end-market spending on data centers, chips, model development, and returns to investors. Enterprises will account for $1.8 trillion, consumers for $366 billion, major cloud operators for internal use for $204 billion, and new technologies like physical AI for $200 billion. The estimate covers the world excluding China; infrastructure costs are mainly concentrated in the US.

The $1.8 trillion is a measure of how much the AI ecosystem will demand from enterprises, not a forecast of actual spending. In one BCG scenario, companies cover this amount by reallocating $857 billion from existing technology budgets and adding $937 billion in new spending. Their IT budgets would then be about 13% higher than the International Data Corporation (IDC) forecast, with AI's share growing from 5% to 25% by 2030.

The scenario shows how much AI growth depends on corporate spending. "The question is not just whether AI brings immediate financial returns," says Clark O'Neil, BCG's global head of AI, infrastructure, and cybersecurity. "The question is also what is the cost of falling behind." But the fear of falling behind should not replace a business case.

BCG considers the development of AI infrastructure more sustainable than past booms: major cloud operators finance investments mainly from cash flow rather than debt. However, the ability to finance infrastructure does not guarantee adequate returns. O'Neil identifies two indicators: how infrastructure investments are financed—a sustained shift to debt signals risk—and whether deployed computing power generates sufficient revenue.

BCG assumes an average lifespan of AI chips of five years and estimates that each gigawatt of power introduced in 2026 should generate about $15 billion in revenue. Weaker demand could trigger price wars and losses. O'Neil advises most buyers to avoid long-term price commitments: flexible contracts will allow taking advantage of future cost reductions. "$2.6 trillion is achievable," he says, "but not guaranteed." And even if enterprises make these expenditures, the question remains whether they will gain enough economic value.