Artificial intelligence is advancing faster than the infrastructure and organisational changes needed to deploy it, creating new bottlenecks for businesses, according to a report by the McKinsey Global Institute.
The report, released this month, found that the complexity of tasks AI can reliably perform has been doubling roughly every four months since 2023. Infrastructure needed to support that growth, including data centres, power capacity and chip manufacturing, is expanding at a much slower pace.
“The tools are here. But the reorganization is not,” McKinsey said, highlighting the widening gap between rapid AI development and the slower pace of change within businesses.
AI adoption has become widespread, with 89% of organisations surveyed in 2026 using AI in at least one business function. However, only 46% have moved beyond the pilot stage, showing that adoption has not necessarily translated into large-scale deployment.
The report also found that, outside a small group of high-performing companies, only around one-quarter of respondents had redesigned their workflows around AI. Most organisations were instead adding AI tools to existing processes rather than changing how work is structured.
That distinction could have a major bearing on the productivity gains businesses are able to achieve from AI. While AI can already help employees complete individual tasks faster, McKinsey said broader productivity improvements will depend on redesigning workflows and implementing those changes across operations.
Infrastructure Faces New Constraints
The bottlenecks facing AI adoption are also changing. Access to advanced chips and computing capacity emerged as early constraints on AI expansion. As demand increased, electricity supply and grid connections became increasingly important challenges.
McKinsey expects the next set of constraints to come increasingly from applications, workforce skills and organisational workflows.
Investment in the physical infrastructure supporting AI is nevertheless accelerating. The report estimates that global investment in data centres could reach USD 7 trillion cumulatively between 2025 and 2030, with a significant share driven by AI demand.
But the rapid pace of investment also carries a risk. Infrastructure could expand faster than actual demand, potentially resulting in excess capacity.
For businesses, the challenge is therefore extending beyond access to AI technology. AI capabilities could advance faster than companies are able to adapt their technology, workflows and organisational structures.
McKinsey said business leaders need to build the ability to adapt quickly and use AI not only to automate existing tasks, but also to create new products, businesses and sources of growth.
The report's broader conclusion is that AI's economic impact will depend not only on how quickly the technology advances, but also on how infrastructure, investment, regulation, organisations, markets and workers respond to that pace of change.
