Proponents of data center development point to several revenue streams for local communities:

Property taxes: A large data center campus can add tens of millions of dollars per year to a township’s property tax base. The Gaines Township parcel under consideration (168 acres) would become one of the highest-assessed properties in Kent County. Supporters point to property tax as a stable, recurring revenue source that benefits schools, roads, and county services.

Construction-phase jobs: A major data center requires 1,000–3,000 construction workers during the build phase, typically lasting 2–4 years. These are well-paying trades jobs that flow through local union halls and subcontractors.

Utility tax and sales tax: Data centers pay significant Michigan use tax on equipment purchases. A $1B facility build generates roughly $60M in state use tax, a portion of which returns to the local community through revenue-sharing formulas.

Permanent jobs: A hyperscale campus typically creates 50–200 permanent positions. These are generally high-skill, high-wage IT operations jobs, plus support roles. Microsoft has announced local hiring commitments as part of its Michigan expansion.

The revenue case changes significantly when subsidies are factored in. Data center developers routinely negotiate significant tax reductions before breaking ground:

Personal property tax exemptions: Michigan provides a 100% personal property tax exemption for data center equipment under PA 328 of 2015 (the Data Center Equipment Tax Exemption). This is the single largest subsidy available — a major data center’s server and networking equipment can be worth hundreds of millions of dollars, all exempt from local property tax.

MEGA credits and MEDC incentives: The Michigan Economic Development Corporation can offer additional incentives through the Michigan Strategic Fund — tax credits, grants, and site development assistance. These are typically tied to job creation commitments that, once met, are not recaptured if the jobs later disappear.

Real property assessment agreements: Townships can enter into development agreements that affect how property is assessed and taxed. These agreements are not always made public before a board votes on a permit.

The net calculation: Critics of aggressive incentives point to studies in Virginia (the largest data center market) showing that, net of subsidies, some local governments receive far less than headline numbers suggest. Virginia has begun requiring impact statements that calculate net fiscal benefit before permits are issued.

The job creation narrative deserves scrutiny. Data center employment patterns differ significantly from traditional manufacturing or commercial development:

The permanent job count: Hyperscale data centers are highly automated. A 100 MW campus requires roughly 50–150 permanent employees. For context, a mid-size manufacturing plant of similar footprint might employ 300–800 workers. The gap matters when communities are comparing data centers to other industrial uses competing for the same sites.

What the jobs are: Permanent data center roles are predominantly IT infrastructure and operations jobs that require specialized credentials. These are well-paying positions — median salaries of $75,000–$100,000+ — but they are not entry-level roles accessible to residents without prior tech industry experience. Local hiring commitments in developer agreements often cover construction, not permanent operation.

The construction multiplier: Construction-phase employment is real and significant. Local union halls in West Michigan are likely to see meaningful work from a major build. However, this employment ends when construction does — and construction timelines for hyperscale facilities can be 2–4 years.

Indirect economic effects: Economic developers argue that anchor investments like Microsoft attract supply chains, vendors, and ancillary businesses. Critics note that data centers buy most inputs (hardware, software, cooling systems) from national and global suppliers — local indirect effects are smaller than for manufacturing.

Questions to ask: How many permanent jobs? What is the local hire commitment and for how long? What are the wage floors? What happens if the facility is sold or the tech cycle changes?

Approving a data center is also a zoning decision about what alternative uses cannot occupy the same land. This is the opportunity cost question — what futures get foreclosed?

Land use: A hyperscale campus consumes 100–500 acres of industrial-zoned land, typically for a single tenant. That land cannot simultaneously serve manufacturing, mixed-use, housing, or agricultural purposes. Once rezoned and developed, conversion costs are high.

Grid capacity: A 100 MW data center requires grid infrastructure upgrades that, once built to serve the data center, constrain what else can be powered in the same area. Manufacturers, commercial developers, and EV charging infrastructure all draw from the same grid. In tight grid conditions, serving the data center’s load can slow or block competing development.

Water: Cooling-water users consume and return water to the watershed. In Great Lakes communities, water withdrawals above certain thresholds require state permits and must meet the Great Lakes Compact anti-diversion standards. Communities that permit significant water use for data centers may face constraints on water for other users.

Assessment: Communities negotiating their first data center deal often lack comparable experience. The developer’s team has done dozens of these negotiations. Townships writing model ordinances — like Cascade Township — are trying to codify the lessons before the next application arrives.

Whether you support or oppose data center development, these are the questions every community should be able to answer before a vote:

1. Net fiscal impact over 10 years: After all exemptions, abatements, and infrastructure costs borne by the public, what does the community net in tax revenue?
2. Who pays for the grid upgrade? Is the developer funding transmission and distribution improvements, or are those costs allocated to ratepayers through the utility’s capital recovery process?
3. What are the binding job commitments? How many permanent jobs, at what wages, with what local hire percentage — and what happens if those commitments are not met?
4. What are the water withdrawal numbers? Daily and annual maximum withdrawals, discharge location, and what the state permit process will require.
5. What is the exit scenario? If the operator sells or abandons the facility, what environmental remediation obligations attach to the land? Who is responsible for decommissioning?
6. What is off the table? If this parcel goes to a data center, what alternative uses become impossible or significantly harder over the 20-year horizon?

Cascade Township is building answers to these questions into its model ordinance. Gaines Township is voting tonight without that framework in place.

Virginia (Northern Virginia “Data Center Alley”): The largest data center market in the world. Loudoun County hosts over 150 data centers consuming >70% of the county’s electricity. Property tax revenue is significant, but grid costs, noise complaints, and viewshed concerns have prompted new ordinances limiting scale and requiring fiscal impact analyses. Virginia law now requires localities to produce fiscal impact statements for large projects.

Oregon (Umatilla County): Early mover in data center hosting — cheap hydropower attracted Google, Amazon, and Meta. Counties discovered that data center personal property (servers) qualified for enterprise zone exemptions that dramatically reduced the advertised tax benefit. Oregon has since modified enterprise zone rules for data centers.

Iowa: Heavy state subsidies attracted Microsoft, Google, and Meta data centers. A 2021 Iowa study found the state’s data center sales tax exemption cost $133M in forgone state revenue in a single year — while generating approximately 1,000 permanent jobs statewide. The per-job subsidy calculation prompted legislative debate over whether the math pencils out.

The emerging consensus: Communities that have gone through a full cycle of negotiation, construction, and operation generally report that the economic case is real but overstated in initial developer projections. The communities that fared best negotiated from a position of knowledge: completed fiscal impact analyses, secured binding infrastructure commitments in the development agreement, and required disclosure of ongoing energy and water use.