Second Green AI Summit at Harvard and Boston University Successfully Convened
How Data Centers Are Planned, Approved, and Built in the Age of AI
As the demand for AI-ready digital infrastructure accelerates, data center development has become one of the most complex, resource-intensive sectors in global industry. Companies must now navigate a wide array of siting, permitting, energy, and workforce challenges—while meeting aggressive timelines and sustainability goals.
This chapter breaks down the real-world practices shaping where, how, and why data centers are being built today. From application timelines to incentive strategies, it offers an insider’s look at what drives decision-making in the age of hyperscale AI infrastructure.
Before construction begins, developers face a multi-year, multi-agency approval process involving:
Site selection
Permitting and zoning
Design review
Energy interconnection agreements
Typical development timelines range from 21 to 48 months, although large companies often fast-track projects through early-stage co-development with local governments and utilities.
Key bottlenecks include:
Grid access delays (often years-long waitlists)
GPU and chip supply shortages
Environmental review and local opposition
Example: Meta’s Project Cosmo in Wyoming agreed to clean up waterways, use renewable power, and limit water use as part of its local approval process.
Companies prioritize sites that offer:
Affordable, undeveloped land
Cool climates (to reduce cooling costs)
Low disaster risk
Access to fiber-optic infrastructure
Power grid stability and reliability
States like Wyoming are emerging as hotbeds for development thanks to:
No state income or corporate taxes
Data center-specific sales tax exemptions
Low property taxes
Workforce training grants and utility cost reimbursement
Greenfield development—building from the ground up—is often preferred over retrofitting existing structures, as it allows custom design for AI workloads, power density, and cooling.
Local governments are often eager to attract data centers as engines of economic development. These facilities can:
Create thousands of construction jobs
Support hundreds of ongoing roles
Generate millions in tax revenue
Stimulate tech ecosystem growth
Impact Example:
A single hyperscale facility can inject over $240 million into a local economy during construction and support $30+ million in ongoing economic activity annually.
However, there are trade-offs:
Pressure on land and water resources
Traffic and noise during construction
Impacts on neighboring agricultural or residential land
Rising tensions around aesthetics and zoning
Some companies are addressing this by:
Building elevated solar farms for dual land use (e.g. grazing animals)
Investing in local training programs (e.g. Microsoft’s Data Center Academy in Wyoming)
Using closed-loop cooling systems to reduce water demand
Permitting and power grid delays are now among the biggest challenges in data center siting.
States with pro-business tax codes, stable power grids, and cool climates are winning hyperscale investment.
Local government support is crucial and often tied to jobs, training, and clean energy commitments.
As public scrutiny grows, developers must balance economic benefits with environmental and community impacts.