JPMorgan has concluded that an acceleration in revenues among artificial intelligence companies has strengthened the economic rationale for the significant infrastructure spending underpinning the sector.
Strategist Nikolaos Panigirtzoglou told clients that recent revenue momentum makes the AI capital expenditure cycle look more economically viable than it did six months ago, and that faster top-line growth can make the sector's heavy capital intensity more sustainable - on the condition that revenue gains ultimately convert into durable margins and acceptable returns on invested capital.
The bank highlights the wide variance in estimates for cumulative AI data center capital spending through 2030. Its credit research group projects about $5.5 trillion in cumulative capex, while some external estimates reach as high as $10 trillion. JPMorgan frames a midpoint of roughly $7.5 trillion when accounting for that range.
On the revenue side, JPMorgan's equity analysts expect AI cloud providers, model providers and neoclouds to reach a combined revenue run-rate near $1.6 trillion by the end of 2026. That base is projected to grow at annual rates between 10% and 20%, taking the combined run-rate to a range of $2.5 trillion to $3 trillion by 2030.
Much of the projected demand for AI infrastructure is anticipated to come from large enterprises. A JPMorgan survey of Asia Pacific companies found that average AI spending as a share of expenses plus capex rose from 4.5% over the past 12 months to an expected 5.8% over the next 12 months. When applied globally, JPMorgan says that 5.8% share implies roughly $1.7 trillion of AI spending.
The bank notes that reaching a combined AI revenue run-rate of $2.5 trillion by 2030 would require that enterprise AI share to increase to around 6.5% to 7.0%. Panigirtzoglou described that step-up as a meaningful increase from the near-term 5.8%, but not implausible if AI moves from experimental projects to scaled deployments and if enterprises are able to fund AI budgets through productivity savings, labour substitution, revenue uplift, or reduced spending on legacy technology.
JPMorgan's analysis therefore ties the viability of the AI capex cycle to two linked outcomes: continued above-trend revenue growth from AI providers, and the translation of that growth into durable profitability and satisfactory returns on invested capital. The bank's varied capex scenarios and its revenue run-rate forecasts underline how the sector's infrastructure story hinges on enterprise adoption and economic returns.
Contextual note: The bank's figures and survey results form the basis of its view; JPMorgan presents both upside and downside possibilities depending on how revenues and enterprise budgets evolve.