Gartner reports that AI infrastructure spending could reach $7 trillion by the end of 2030. To achieve a 12% return on this capital, the industry would require $8.5 trillion in cumulative revenue. However, current forecasts suggest that enterprise and consumer AI revenue linked to data centers will not exceed $3.6 trillion by that time.
This revenue shortfall creates a risk of asset impairment, which Gartner estimates could exceed $3.5 trillion by 2030 and potentially surpass $4 trillion by 2032. The research firm notes that providers invested nearly $1 trillion in 2026, with an additional $6 trillion projected for the 2027–2030 period.
Neocloud providers, which rent out GPU capacity and AI-optimized infrastructure, face significant pressure. Gartner predicts that over 80% of these operators will cease to function independently by 2030 due to high costs and competition from hyperscalers. Spending on AI-optimized IaaS is expected to grow from $66 billion in 2027 to $174 billion in 2030.
Analyst Will Sommer points out that the asset-light business models of many AI-native companies prevent revenue from reaching infrastructure owners. This structural issue, combined with rapid hardware turnover, further complicates the path to profitability and increases the likelihood of market consolidation through mergers, acquisitions, or closures.
