MRCTI Data Center Framework Puts Local Costs Ahead of AI Promises

The Mississippi River Cities and Towns Initiative says 75% of its members have faced AI infrastructure interest, prompting an MRCTI data center framework for local reviews.

The coalition unveiled two evaluation resources during its annual meeting in Natchez, Mississippi, on September 16. One outlines questions that communities can ask developers. The other helps officials calculate public costs, resource demands, and community benefits before approving a project.

That intervention changes the balance between developers and smaller municipalities. Companies usually arrive with technical consultants, confidential utility arrangements, and ambitious investment forecasts. Local governments must judge those proposals while protecting water supplies, electricity customers, public budgets, and residents living near construction sites.

The framework does not oppose Mississippi River data centers. It instead challenges the idea that attracting a large AI project is automatically a local economic win. The central conflict is now the developer’s promised investment versus the community’s full, long-term cost.

The MRCTI Data Center Framework Gives Cities a Common Starting Point

The coalition is turning scattered local negotiations into a shared regional test for AI infrastructure proposals.

MRCTI represents communities along a river system that crosses ten states. Its members range from major metropolitan governments to smaller cities with limited planning staff. That difference matters when a municipality receives a proposal involving industrial-scale electricity, water, land, roads, and tax incentives.

According to the mayors’ announcement, more than three-quarters of member cities have been approached about hosting some form of AI infrastructure. Coalition co-chair Melisa Logan, the mayor of Blytheville, Arkansas, said members were asking how to protect water and utility service while evaluating those offers.

The organization responded with two resources rather than a single regional verdict. Its data center “Wishlist,” developed with the Sierra Club, identifies risks and benefits that elected officials should examine. A separate AI Infrastructure Community Evaluation Framework focuses on fiscal returns, public cost protections, resource use, and community benefits.

The second resource originated with the Future Caucus of the Missouri House of Representatives. Lindenwood University now hosts the AI evaluation framework, giving local officials a structured way to compare what a developer promises with what the public must provide.

That distinction is important. A project can represent a large private investment while still imposing costs on roads, water systems, emergency services, housing, and the electric grid. Gross investment does not reveal who funds those systems, who owns the improvements, or what happens if demand changes.

Data center developers also vary widely. Some facilities use water-intensive cooling, while others use closed-loop or air-cooled systems. Some require dedicated generation and transmission infrastructure. Others can locate near existing capacity or schedule workloads around grid conditions.

A checklist cannot settle those technical differences. It can ensure that officials identify them before approving permits, incentives, or service agreements.

The MRCTI data center framework therefore functions as a negotiating baseline. It encourages communities to request comparable information about projects that might otherwise be presented through different assumptions and confidential models.

It also gives mayors political cover to ask difficult questions. A small city can point to a regional standard when requesting water studies, financial guarantees, or enforceable hiring commitments. That is harder for a developer to dismiss as the concern of one unfriendly municipality.

The immediate change is procedural, not regulatory. MRCTI cannot approve or reject a facility, set utility rates, or force a company to publish confidential contracts. It can give local leaders a common vocabulary before those binding decisions occur.

That may be the framework’s most practical contribution. Communities rarely suffer from a complete absence of promises. They struggle to compare those promises with obligations that can last decades.

Why Mississippi River Data Centers Are Reaching Smaller Communities

AI infrastructure is moving toward places with land, energy access, water, and supportive governments, putting rural municipalities at the center of national computing expansion.

The United States already has more than 3,000 operating data centers, according to the report on MRCTI’s announcement. Illinois alone has 139 operating facilities and more than 100 under development, Alton Mayor David Goins said during the Natchez meeting.

The expansion is no longer concentrated around established technology corridors. A national location analysis found that roughly three-quarters of planned facilities were located in the South or Midwest. Many proposed sites were in rural areas.

Several forces make the Mississippi River Basin attractive. Communities offer large tracts of land, access to transmission corridors, industrial water sources, transportation networks, and state economic-development incentives. Some also have existing industrial sites that can accommodate high electricity demand.

AI has raised the scale of that demand. Training and operating large models requires dense clusters of computing equipment. Those systems run continuously, produce heat, and depend on electrical and cooling infrastructure that must work reliably.

Developers consequently look beyond traditional data center markets where land, grid connections, and construction capacity are constrained. River communities can appear ready for development, particularly when state leaders want technology investment outside major cities.

For local governments, the appeal is understandable. A large project can expand the tax base, fund infrastructure, create construction work, and attract related suppliers. Rural communities that have lost population or industrial employers have strong reasons to consider those benefits.

However, a data center is not interchangeable with a factory. A facility can require enormous capital investment and significant public infrastructure while supporting a comparatively limited permanent workforce. Construction employment may surge, then fall when the campus becomes operational.

The tax value also depends on state law. Sales-tax exemptions, property-tax abatements, and negotiated incentives can substantially change how much revenue reaches the host community. A headline investment figure does not show the taxable portion.

Local capacity creates another pressure point. A proposal might require a small planning department to review engineering, utility, environmental, and financial issues at the same time. Consultants working for the developer may possess far more information than the city receiving the proposal.

The MRCTI data center framework addresses that imbalance by separating project size from project value. It asks whether public obligations remain manageable after incentives, infrastructure costs, resource consumption, and long-term risks are counted.

This approach does not assume that rural communities should reject AI facilities. It recognizes that those communities should not accept weaker protections simply because they have fewer staff or less bargaining experience.

The issue extends beyond the host city. Water withdrawals can affect shared aquifers and river systems. New generating plants and transmission projects can affect utility customers across an entire service territory. Tax decisions can shift costs to neighboring jurisdictions.

That is why Mississippi River data centers have become a regional governance question. The facility sits in one place, but its electricity, water, transportation, and financial consequences rarely stop at the municipal boundary.

The Real Contest Is Investment Promises Versus Public Exposure

A credible data center agreement must explain who carries the downside, not merely who celebrates the initial investment.

Louisiana provides a clear example of both sides. Meta announced a major AI data center in Richland Parish, an agricultural area in northeast Louisiana. The original project was described as a four-million-square-foot facility with completion expected in 2030.

The project carried significant economic promises. Meta said it expected 5,000 temporary construction jobs and 500 operational positions. It also pledged infrastructure investment for roads and water systems in the parish.

Power requirements introduced a more complicated calculation. Entergy proposed three natural gas generating plants capable of producing 2,262 megawatts. That capacity would support the data center while adding major long-lived assets to the utility system.

The Louisiana Public Service Commission approved the initial generation and transmission agreement in August 2025. Entergy said Meta would pay its share of infrastructure costs and that other customers would be protected.

The company’s utility approval also authorized an expedited process for procuring up to 1,500 megawatts of solar resources. Two gas plants were expected to begin operating in late 2028, with a third targeted for the end of 2029.

Supporters view that arrangement as more than a dedicated power supply. They argue that a large customer can support generation and transmission upgrades that strengthen the broader grid. New investment can also distribute fixed utility costs across a larger volume of electricity sales.

Critics focus on obligations that remain outside the developer’s direct payment. Confidential agreements can obscure cost allocation, performance requirements, and exit provisions. Transmission, maintenance, fuel, or replacement costs can still affect other customers, depending on contract terms and regulatory decisions.

The disagreement is not simply about whether Meta pays a bill. It concerns which risks the contract transfers to households and businesses if construction changes, demand falls, or the company leaves before utility assets reach the end of their useful lives.

An Associated Press examination of the Richland Parish project found that confidentiality restrictions prevented the public from seeing how much Meta would pay toward the electricity infrastructure. Regulators reviewed the contract, but key details remained undisclosed.

The report also found that customers were expected to help fund a transmission line serving the facility. Consumer advocates questioned whether the public could inherit costs if the developer did not renew its electricity agreement.

Those questions illustrate why the AI infrastructure community impact cannot be measured at the groundbreaking ceremony. The relevant period extends across construction, operation, contract renewal, plant retirement, and possible site closure.

A responsible agreement must address several points. It should identify which party funds dedicated infrastructure, how cost overruns are handled, and what financial security exists if the project is delayed. It should also explain whether other customers pay before the facility begins operating.

Communities need similar clarity on employment. Construction-job estimates should be separated from permanent positions. Hiring commitments should identify the skills required, the geographic area covered, and whether the obligations are enforceable.

Public revenue projections deserve the same treatment. Officials need taxable values after exemptions, not just total capital spending. They also need estimates for road maintenance, emergency response, housing pressure, and utility expansion.

This is the primary tension behind the MRCTI data center framework. Developers offer scale, speed, and private capital. Communities must determine whether public systems absorb a disproportionate share of uncertainty.

Water Policy Is Becoming Part of AI Infrastructure Policy

Water availability alone does not answer whether a data center’s withdrawals are sustainable, transparent, or compatible with competing local needs.

Data centers use water directly for cooling and indirectly through electricity generation. The amount varies according to climate, facility design, workload, cooling technology, and the water intensity of the regional power supply.

That variation makes universal claims unreliable. A large facility using evaporative cooling can consume far more water than one using a closed-loop system. Even the same facility may use different amounts during hot weather and cooler periods.

Some published estimates place a large data center’s daily use near five million gallons. That figure can help illustrate scale, but it should not be treated as a forecast for every project. Local leaders need site-specific engineering data.

The critical questions concern source, timing, and scarcity. A withdrawal from a resilient surface-water system has different consequences from pumping a stressed aquifer. Consumption during drought or peak agricultural demand can matter more than the annual average.

MRCTI is linking those project-level reviews with a broader river policy. The coalition wants other Mississippi River states to consider rules modeled on Missouri’s restrictions on large-scale water exports.

Missouri’s law requires a permit when water is withdrawn in the state and transported more than 30 miles beyond its borders. The state’s water export permit includes exceptions for bottled or packaged water intended for individual use.

Lawmakers from Mississippi, Louisiana, and Minnesota committed to pursuing related legislation, according to MRCTI. Coalition members ultimately want a federally recognized Mississippi River Compact modeled on the Great Lakes Compact.

That proposal is broader than data center cooling. It seeks to establish regional control over water before scarcity or commercial pressure produces conflicting state policies. AI development has added urgency because many communities are receiving infrastructure proposals simultaneously.

Still, an export restriction would not resolve every local concern. A data center can place pressure on a water system without sending water across state lines. Cities would still need disclosure, conservation requirements, drought plans, and monitoring.

A meaningful review should identify expected withdrawal and consumption separately. Withdrawal measures water taken from a source, while consumption measures the portion not returned to that source. The distinction affects how communities evaluate long-term supply.

Officials should also examine wastewater. Cooling systems can concentrate minerals or require chemical treatment. A proposal must explain where discharged water goes, what treatment is necessary, and who funds any municipal system upgrades.

Emergency conditions deserve explicit terms. A facility could agree to reduce water use during drought, switch cooling modes, or rely on stored water. Those promises need measurable triggers rather than general statements about conservation.

The same principle applies to public reporting. Annual estimates offer little protection if actual use remains confidential. Communities need meter data, reporting schedules, and enforcement options when consumption exceeds approved levels.

AI infrastructure community impact therefore depends on governance as much as engineering. A technically efficient cooling design can still create conflict when residents cannot verify its use. Transparent data can expose problems early and distinguish real constraints from speculation.

The mayors’ regional effort reflects that reality. Water is not merely an input that a developer purchases. It is a shared resource governed across municipal and state boundaries.

A Framework Cannot Replace Disclosure or Enforceable Contracts

The MRCTI resources improve the questions cities ask, but their value depends on access to reliable answers and authority to act on them.

The coalition’s approach has an immediate limitation. Its framework is guidance, not law. A developer can decline to disclose information unless permits, incentives, utility proceedings, or state regulations require it.

Commercial confidentiality makes that problem harder. Utilities and technology companies argue that service agreements contain sensitive forecasts, pricing terms, and operational details. Consumer advocates answer that secrecy prevents the public from testing claims about rates and risk.

Mississippi offers a warning about that information gap. Amazon has announced multiple data center projects in the state, backed by large infrastructure commitments. Entergy Mississippi says those customers will strengthen the grid and reduce long-term pressure on bills.

A Synapse Energy Economics analysis reached a different conclusion. Based on public filings, the study estimated that residential customers had already been charged about $38 million for data center-related investments by March 2026.

The study estimated a monthly residential effect of $10.60, although its authors acknowledged uncertainty because major contract details were not public. Entergy disputed the analysis and said data center customers were fully funding their infrastructure needs.

The Mississippi rate analysis demonstrates the core verification problem. Both sides make claims about who benefits, but incomplete disclosure prevents outsiders from independently reconstructing the calculation.

A checklist cannot overcome that barrier by itself. Cities need contractual rights to information, and utility regulators need access to cost-allocation data. The public also needs enough aggregated evidence to evaluate whether promised protections are working.

Enforcement is equally important. A developer may promise local hiring, water efficiency, renewable energy procurement, or road improvements. Those commitments have limited value without deadlines, measurement methods, remedies, and a responsible enforcement body.

Officials must also account for changes after approval. AI computing requirements evolve quickly. A campus can expand, install denser hardware, or shift its cooling strategy. A review based only on the first construction phase may understate the eventual load.

Phased approvals offer one response. A city can connect later expansions to verified performance during earlier phases. Water, power, noise, traffic, and employment commitments can then be tested before additional capacity receives approval.

Financial security provides another safeguard. Deposits, minimum-payment provisions, parent guarantees, and decommissioning funds can reduce public exposure if a project stops. The correct instrument depends on state utility law and the infrastructure involved.

Communities must avoid overconfidence in projected benefits as well. Permanent jobs may materialize, local tax revenue may increase, and new transmission can improve reliability. None of those outcomes is guaranteed solely by the project’s announced size.

The framework should therefore be treated as the beginning of due diligence. It identifies the records and protections officials need. It does not certify that a particular project is safe, economical, or beneficial.

That skeptical reading does not weaken MRCTI’s work. It clarifies the next institutional step. Regional guidance becomes consequential when cities incorporate it into ordinances, development agreements, utility reviews, and incentive conditions.

Three Signals Will Show Whether the Mayors’ Plan Has Teeth

The next test is whether regional principles become measurable state rules and enforceable local agreements.

The first signal is legislation in Mississippi, Louisiana, and Minnesota. Lawmakers in those states said they would pursue measures based on Missouri’s water-export restrictions. Filed bills, committee hearings, and enacted requirements would show that MRCTI can move beyond voluntary coordination.

The details will matter more than the number of bills. Policymakers must define covered withdrawals, exemptions, permit standards, drought protections, reporting duties, and enforcement. Weak rules with broad exemptions would do little to change local bargaining power.

A path toward a Mississippi River Compact would be more significant. It would require cooperation among several states and eventually federal recognition. Progress would strengthen the mayors’ argument that water governance must match the river’s regional scale.

The second signal is the content of new local agreements. Cities using the MRCTI data center framework should begin requesting detailed water budgets, utility cost protections, infrastructure schedules, hiring measures, and exit guarantees.

Public access to those terms will determine whether the framework creates accountability. A municipality cannot demonstrate successful protection by saying that confidential documents contain adequate safeguards. It needs publishable metrics that residents can track.

Watch for agreements that distinguish temporary construction work from permanent employment. They should also separate developer-funded improvements from costs assigned to utilities, states, parishes, or cities.

The third signal is measured performance at existing projects. Rate changes, water withdrawals, tax receipts, construction impacts, and permanent hiring will reveal whether developer forecasts match real outcomes.

Louisiana and Mississippi will be especially informative because large projects are already testing their utility systems and regulatory models. If residential bills remain protected and promised infrastructure arrives, supporters will gain evidence for their position.

If public costs appear before benefits, the coalition’s caution will look increasingly justified. Unexplained rate changes or unavailable data would also strengthen the case for mandatory disclosure.

The same evidence should inform future proposals across the basin. Cities do not need to repeat every negotiation independently. They can compare real results with the assumptions embedded in new applications.

For developers, this process creates a clearer route to community acceptance. Companies that can document efficient water systems, pay for dedicated infrastructure, and accept enforceable commitments will have stronger projects.

For residents, the issue is not whether AI services are useful. It is whether the physical infrastructure behind those services is governed with the same care applied to other large industrial developments.

Knowledge workers and businesses also have a stake. The applications they use depend on expanding computing capacity, but hidden infrastructure risks can surface through electricity costs, water conflicts, construction delays, and political opposition.

The MRCTI data center framework offers cities a practical starting point for managing those risks. Its success will not be measured by how many checklists officials download. It will be measured by contracts, public data, and state rules that allocate costs before construction makes them difficult to reverse.

As more Mississippi River data centers move from proposals to operating facilities, communities should ask one direct question: do the agreements protect residents when the optimistic forecast fails? The answer will determine whether the AI buildout becomes durable regional investment or another cycle of privatized gains and public obligations.