Morgan Stanley: Data Centers Must Break Ground by October for 2027 Launch
In brief
- Morgan Stanley sets October deadline for data centers targeting 2027 operational status
- Two-to-three-year groundbreaking-to-operational timelines now standard for data centers
- 38 GW power shortfall projected for U.S. data centers through 2028
- Over 300 local moratoriums on data center construction enacted since 2023
- 35-40% of announced global capacity faces meaningful delay or cancellation risk
The October Deadline
Morgan Stanley is warning that any data center project targeting operational status before the end of 2027 needs to begin construction by October. This timeline isn't arbitrary. Groundbreaking-to-operational timelines now run two to three years at minimum, compressing the practical window for hyperscalers planning major capacity additions.
The urgency is real. Capital expenditures across Microsoft, Amazon, and Alphabet are forecast to reach roughly $785 billion in 2026. Money isn't the constraint. Time is.
Regulatory and Grid Bottlenecks
The delays stem from two overlapping pressures. More than 300 local moratoriums on new data center construction have been enacted across the United States since 2023, slowing permitting cycles and creating uncertainty for developers. Grid interconnection adds another layer. Grid interconnection queues in many U.S. regions now stretch five to seven years, forcing hyperscalers to plan years ahead just to secure power commitments.
The result: a capacity crunch. Morgan Stanley projects a 38 GW power shortfall for U.S. data centers through 2028. Separately, JPMorgan analysis found that over 60 percent of planned 2027 capacity has not yet started construction. Even more broadly, Bernstein estimates that 35 to 40 percent of globally announced data center capacity faces meaningful risk of delay or outright cancellation through 2027.
The Race Against Time
"Money is no longer the limiting factor in the AI infrastructure race. Time is." — Morgan Stanley analysis
Hyperscalers are responding by moving earlier in the development cycle than they historically have, attempting to secure sites and grid commitments before communities and regulators have organized opposition. This shift reflects a hard reality: the infrastructure race for AI is no longer a funding problem. It's a logistics and regulatory one. Projects that don't start now won't finish in time.
