Meta plans to spend more than $10B to build a 1GW data center campus in Lebanon, Indiana, expected to be operational at the end of 2027 or in early 2028
Meta Platforms Inc. said it will spend more than $10 billion to build a data center campus in Lebanon, Indiana, ranking it among …
Context & Ripple Effects
Meta had already committed $10 billion to a Louisiana AI data center in 2024, establishing a pattern of large, owned campus builds rather than a one-off Indiana project. The later expansion of the Louisiana campus into a planned 5GW Hyperion site underscores how quickly Meta’s infrastructure ambitions can scale once a campus is underway.
The Indiana commitment adds a geographically separate 1GW build to that program, with a stated operating window that makes delivery, power availability, and construction execution central constraints rather than simply capital allocation.
First-order effects
- Meta commits more than $10 billion to a new Indiana campus and adds 1GW of planned compute capacity to its internal infrastructure pipeline for late 2027 or early 2028.
- The project creates immediate demand for site construction, power interconnection, and data-center equipment around Lebanon, while making on-time delivery a key operational dependency for Meta.
Second-order effects
- A second large Midwestern campus increases competition for long-duration power capacity, construction labor, and data-center supply-chain capacity among hyperscale builders.
- Meta’s choice to fund another major owned campus reinforces the pressure on cloud and infrastructure rivals to secure sites and power earlier, rather than rely solely on incremental capacity additions.
Third-order effects
- If Meta continues pairing multibillion-dollar regional campuses with larger flagship sites, frontier AI infrastructure will become more concentrated among companies able to finance and execute multi-year power-and-construction programs.
- The economic value of AI buildouts will increasingly depend on physical deployment discipline—power, permitting, equipment delivery, and financing structure—not just access to chips; delays remain a material uncertainty.
The trend: This is another step in the shift from discrete data-center projects to geographically distributed, power-constrained AI compute networks funded by the largest platforms.