Forecasted global data centre caapacity growth through to 2030 will require around $3 trillion in total investment, marking what has been described as the largest infrastructure investment “supercycle” in modern history, according to a new report from real estate services firm JLL.

In its 2026 Global Data Center Outlook report, which examines market data, regional forecasts and strategic implications for the sector, JLL said nearly 100GW of new data centre capacity is expected to be added worldwide between now and the end of the decade.

This expansion would effectively double global capacity and demand eye-watering levels of capital investment.

Americas lead growth as APAC and EMEA expand

JLL forecasts that the data centre sector will grow at a compound annual growth rate of 14% through to 2030, driven primarily by hyperscale cloud expansion and the unquenchable thirst for AI. Between 2025 and 2030, global data centre capacity is projected to increase by 97GW, reaching a potential total of 200GW by the end of the period.

The Americas currently represent the largest data centre region, accounting for around half of global capacity, according to the report.

The region is also forecast to experience the fastest growth, with a projected supply CAGR of 17% through to 2030, maintaining its position as the dominant global market. The United States underpins the majority of this growth, representing approximately 90% of capacity in the Americas.

In Asia-Pacific, data centre capacity is expected to expand from 32GW to 57GW by 2030, achieving a CAGR of 12%. Growth in the region is being led by colocation, which is forecast to grow at 19%, while on-premise capacity is projected to decline by 6% as enterprises continue to migrate workloads to the cloud.

The EMEA region is forecast to grow at a CAGR of 10%, supported by government backing for AI infrastructure and increasing demand for sovereign AI clouds designed to meet data privacy requirements.

The region is expected to add 13GW of new supply, with growth concentrated in established European hubs alongside emerging markets in the Middle East pursuing broader digital transformation strategies.

AI workloads shift from training to inference

Unsurprisignly, AI is expected to account for an increasing share of data centre workloads over the coming years.

While AI represented around a quarter of all data centre workloads in 2025, largely driven by training requirements, a shift is anticipated in 2027 as workloads begin to overtake training as the dominant demand driver.

Once an AI model has been created, inference supports ongoing revenue generation through application usage, with sustained demand increasing as adoption grows. However, this growth is dependent on the development and rapid uptake of inference applications that have yet to emerge at scale.

The report also highlights that inference demand requires geographically distributed infrastructure in order to reduce latency and serve end users effectively, a shift that is expected to influence future capacity planning and deployment strategies.

Power constraints reshape data centre strategies

Energy availability has emerged as a critical constraint on data centre expansion. With average wait times for grid connections in primary data centre markets now exceeding four years, operators are increasingly turning to behind-the-meter power arrangements and colocated battery storage.

In the United States, natural gas is projected to play a significant role in addressing grid constraints, both as temporary bridge power and as a permanent on-site generation solution, reflected in rising global turbine orders. However, some of the largest data centre tenants remain resistant to natural gas due to sustainability concerns.

Natural gas solutions are less prominent in EMEA and APAC, where renewable energy sources such as solar and wind are seeing increased uptake. In parts of EMEA, projects combining renewables with private wire transmission have been shown to reduce power costs for tenants by up to 40% compared with grid-supplied electricity.

Utility interconnection delays have also prompted some operators to move beyond power purchase agreements to directly fund their own generation assets, a trend reinforced by ‘bring your own power’ mandates in markets including Ireland and Texas.

Construction costs rise amid rapid expansion

Construction costs are rising alongside the rapid pace of sector expansion. JLL said average global data centre construction costs increased from $7.7 million per MW in 2020 to $10.7 million per MW in 2025, representing a CAGR of 7%. For 2026, the firm forecasts a further 6% increase to $11.3 million per MW.

However, it should be noted that these figures cover shell and core construction only, with tenants typically responsible for technical fit-out costs, which can reach up to $25 million per MW for AI infrastructure.

Recommended reading

• AI Council Looks to Energy Grid Reform to Power Growth Zones
• AI Energy Crunch Demands “Urgent” Nuclear Reform, Warns Report
• New Coalition Targets Clean Power for UK Data Centres

Speed to power remains the primary driver of site selection, followed by community support, latency considerations and proximity to customers. However, as project sizes continue to grow, variations in construction costs are expected to play a more significant role in location decisions.

Sector consolidation accelerates as investment surges

JLL estimates that the projected 100GW of new capacity by 2030 equates to approximately $1.2 trillion in real estate asset value creation and a requirement for around $870 billion in new debt financing.

When combined with the estimated $1 trillion to $2 trillion that tenants will invest in GPUs and networking infrastructure, total data centre expenditure over the next five years could approach $3 trillion.

The scale and complexity of new data centre developments are also driving further consolidation across the sector. Rising development costs and the increasing sophistication required to build and operate modern facilities are raising barriers to entry, reducing speculative projects and accelerating those backed by established and credible operators.

For these groups, JLL said debt markets are expected to remain accessible.

According to the report, as AI-driven expansion reshapes power, technology and real estate markets, the transition from centralised AI training to distributed inference is expected to fundamentally alter how and where capacity is deployed.