Photo by Tom Barrett / Unsplash
In Meridian, Mississippi, a $10 billion AI data center arrived with a decade of tax holidays — while the city it neighbors has been federally ordered to stop leaking raw sewage into its creeks since 2019.
The manhole at 65th Avenue has a case number. The EPA assigned it: SSO 7, E25-037. In 2024, it overflowed eighteen separate times. The contents — untreated sewage — ran into Okatibbee Creek, which cuts through Meridian's east side. There is no dramatic moment to describe here. The overflow happens, and the creek carries it, and the families whose yards back up to the waterway know the smell before they check the news.
Meridian is a city of about 34,000 people, located in east-central Mississippi, two hours northeast of Jackson. The city has a median household income of $36,562. More than one in three residents lives below the poverty line. The population has been declining — slowly, steadily, about one percent per year — for long enough that the schools have consolidated, the tax base has narrowed, and the maintenance schedule for the sewer system has been moved to "when possible" for longer than anyone cares to document officially.
The West Meridian Trunk Line has collapsed — twice. The B Street sewer main, running from 8th Avenue to St. Andrews Street, has experienced what the city's own engineering reports call "numerous pipe failures over the last thirty years." The South Wastewater Treatment Plant, which processes sewage for a significant portion of Meridian's households, has components that are pushing seventy-five years old. David Ruhl, the program manager overseeing the remediation effort for Waggoner Engineering, said it plainly to the local paper: "Every year you're going to be sinking money into this plant."
In March of 2024, in a single month, the city's sewage system discharged over two million gallons of raw waste. That is not a metaphor for dysfunction. That is a measurement, recorded by city engineers, filed with federal regulators, entered into a legal record that now runs to five annual reports.
The residents near 65th Avenue did not choose this infrastructure. They inherited it, the way you inherit debt you did not incur — because the people with the authority to fix it kept finding other priorities.
The mechanism here is not corruption in the transactional sense. It is structural — a sequence of decisions, each made by actors responding to their actual incentives, that produces an outcome no individual actor had to intend. The sequence is worth naming precisely, because it is repeating in other cities and will continue to repeat until the naming interrupts it.
Step one is disinvestment. Meridian's sewer system did not collapse suddenly. It was undermaintained for decades — not because local officials were uniquely negligent, but because the city's fiscal capacity contracted as its population shrank and its poverty rate rose. A city where 34.3% of residents are below the poverty line cannot generate the property tax revenue to fund infrastructure replacement at the scale a federal consent order requires. The city entered a fiscal trap: the revenue base is too small to maintain the infrastructure, so the infrastructure degrades, so it costs more to fix, so it consumes more of a shrinking budget.
The EPA/DOJ Consent Decree (Case No. 3:19-cv-00427), signed August 6, 2019, identified the root cause of Meridian's 46+ annual sewage overflows explicitly as "decades of deferred maintenance and lack of investment in the infrastructure." This is not an analysis from an advocacy group — it is the federal government's own finding, submitted under penalty of law in the city's Annual Report No. 5 (February 28, 2025). The city recorded 63 SSOs discharging 14.3 million gallons in 2023; 46 SSOs discharging 3.3 million gallons in 2024. The trend is improving under federal enforcement, but total remediation cost is estimated at $150–250 million over 20 years. The city has spent $17.8 million in ARPA funds on the effort so far.
Step two is federal enforcement. The EPA, DOJ, and MDEQ imposed the consent decree in 2019. The city is now legally compelled to spend money it does not have. This creates a dependency cascade: federal appropriations, pandemic relief funds, bond financing, and grant applications become the operating model for basic civic infrastructure. The city is not governing; it is grant-writing.
Step three is selective development dealmaking. On January 9, 2025 — five years into the consent decree, while the manhole at 65th Avenue was preparing for its eighteenth overflow of the year — Governor Tate Reeves stood before cameras to announce that Dallas-based Compass Datacenters would build an eight-center, 500-megawatt hyperscale AI campus in Lauderdale County. The deal's price tag: a ten-year exemption from all Mississippi state income, franchise, and corporate taxes, coordination of local property tax abatements up to 66% for thirty years, and state-funded site preparation. The statutory threshold that qualified Compass for this package: twenty new jobs and a $20 million capital investment. Translation: a $10 billion company gets a decade of tax-free operation for promising to hire roughly two full shift crews.
To power the Compass campus's 500-megawatt draw, the Mississippi Public Service Commission approved extending the life of a coal-fired generating unit at Plant Victor J. Daniel (Jackson County) into the mid-2030s — a plant previously scheduled for closure by 2027. PSC Commissioner De'Keither Stamps defended the decision: "We can't stop economic development because we've got to wait, you know, 15 years for some nuclear power in the service area." The power demand externalization is structural: 500 MW of new load on Mississippi Power's grid — serving 192,000 residential customers — will place upward pressure on electricity rates that fall disproportionately on low-income households in the same region the deal is supposed to help.
Step four is the transparency suppression that makes accountability structurally impossible. Mississippi is one of twelve states that do not publicly disclose data center tax subsidy amounts, according to Good Jobs First's 2025 analysis. The Compass deal was announced, celebrated, and entered into state law with essentially no public cost disclosure. By November 2025 — ten months after groundbreaking — a Lauderdale County Supervisor told the local press that they had still not received fiscal projections on the deal. So we're running billion-dollar infrastructure for a Texas AI company on public land… and calling it economic development. Cool. Now explain who pays.
The answer, as the mechanism chain makes clear, is that Meridian's existing residents pay — in rate increases, in deferred remediation timelines, in a tax base that will not benefit from the Compass campus during the precise decade when consent decree infrastructure costs peak. The timeline overlap is precise: the Consent Decree runs approximately 2019–2039. The Compass tax exemption runs 2025–2035. These are not separate stories.
What is happening in Meridian is not unique to Mississippi. It is a specific, well-documented variant of a pattern that researchers in economic development, regional planning, and public finance have been tracking with increasing precision: the structural tendency of state-level economic development incentives to flow toward large capital investments in depressed regions while simultaneously leaving the public infrastructure burden on the communities with the least capacity to carry it.
The mechanism has been studied at national scale. Good Jobs First, a nonpartisan research organization that tracks economic development subsidies across all fifty states, has documented that data center incentives in particular are among the most lopsided in the industry — high capital value, low job creation, high energy demand, low local multiplier effect. Bloomberg's infrastructure reporting has found that wholesale electricity costs in regions with dense data center development have increased by as much as 267% — a cost that eventually reaches residential ratepayers through utility rate cases. Mississippi, as of the 2025 Compass deal, has positioned itself at maximum exposure on all three dimensions: maximum subsidy, minimum job threshold, maximum power demand.
The Center for Economic Accountability named the Compass/Mississippi deal the "Worst Economic Development Deal of 2025." CEA President John C. Mozena was precise in his framing: "Nothing good happens when you have CEOs and politicians making massive deals with public money behind closed doors." The CEA's analysis cited the statutory threshold directly — Mississippi Code §57-113-21 — noting that the minimum job requirement of twenty positions is so low as to be functionally meaningless as a community benefit standard. "I'm sure plenty of employers in Mississippi would be happy to hire 20 new workers if that meant they could operate in the state virtually tax-free for a decade," Mozena noted.
This is not a problem of bad intentions at the individual level. State politicians responded rationally to the incentive structure they operate within: announce a headline investment number ($10 billion), accept a corporate tax exemption that generates no immediate budget cost, deliver a press moment, and move on. The long-run fiscal consequences — a compressed county tax base, rising residential electricity rates, no Compass tax contribution during the peak infrastructure remediation window — will be someone else's problem to manage in 2028 or 2030 or 2034.
What the Meridian case adds to this national pattern is specificity and simultaneity. The infrastructure failure is not theoretical. It is a federal consent decree with GPS-mapped overflow coordinates, five years of annual reports, and a manhole that overflowed eighteen times in a single calendar year. The deal that exempts a major corporate tenant from contributing to the tax base is not a future risk — it was signed the same month the city filed yet another sewage remediation progress report with federal regulators. The juxtaposition is not rhetorical. It is a documented, dateable, legally verifiable fact.
The broader implication: where state economic development frameworks reward capital attraction over civic fiscal capacity, infrastructure decay and selective corporate subsidy are not competing priorities — they are the same policy, operating simultaneously, on the same community.
A reasonable counterargument holds that the Compass investment will generate downstream fiscal benefits — construction activity, supplier contracts, property value increases in adjacent commercial zones, and eventual post-exemption tax revenue — that more than offset the ten-year exemption window. Under this reading, the deal and the consent decree are sequential problems, not simultaneous ones: Meridian gets the economic activity now and the tax revenue later. This argument has structural merit and should be taken seriously. It fails primarily on two counts: (1) the transparency architecture that would allow this claim to be verified does not exist — Mississippi does not disclose subsidy costs, and no fiscal projection has been provided to county supervisors ten months post-groundbreaking — and (2) the job-creation threshold (twenty positions) is too low to generate the wage-multiplier effect that would translate economic activity into residential income gains in a 34.3%-poverty-rate city. The additive argument requires trusting a deal structured to avoid accountability.
A second counterargument frames the consent decree remediation shortfall as a federal and state funding insufficiency rather than a misallocation of available resources. Under this view, state leaders cannot be blamed for courting Compass when the consent decree's $150–250 million price tag exceeds what any realistic municipal bond issuance or grant cycle can cover; the infrastructure failure and the corporate incentive deal operate in entirely different budget categories. This is also a structurally coherent argument. It breaks down on the simultaneity problem: the state's economic development apparatus produced a $10 billion deal in the same legislative session and using the same statutory toolbox it could have deployed to create a data-center community benefit fund, infrastructure contribution requirement, or fiscal transparency mandate. The tools existed. The choice not to apply them to community benefit was itself a policy decision. The alternative explanation acknowledges the fiscal gap; it cannot explain the absence of any policy instrument connecting corporate capital attraction to local infrastructure investment.
What is not known: The direct fiscal timeline — specifically, how much of Lauderdale County's tax base will be consumed by consent decree obligations during the 2025–2035 Compass exemption window versus what Compass would have contributed absent the exemption — has not been calculated at the CPA level. This analysis would sharpen the signal significantly and is the highest-priority research gap.
What is not known: Mississippi Power's post-January 2025 rate case filings, which would document projected residential rate increases attributable to Compass's 500 MW power demand, have not been fully analyzed. This is a direct, quantifiable cost to Meridian's low-income households and remains inferred rather than verified.
What is not known: The racial geography of the SSO sites — whether the overflow locations (65th Avenue, B Street, Sowashee Creek corridor) disproportionately affect Black residential neighborhoods — is geographically plausible but has not been verified with Census block data. A FOIA request to MDEQ for SSO location mapping overlaid with demographic data would confirm or refute this dimension.
What would change this signal: If post-2025 Compass employment actualization reports documented substantially more than 20 jobs with above-median wages in Lauderdale County's residential workforce, the distributional argument weakens. If county budget documents showed a dedicated infrastructure contribution from the Compass deal, the structural critique would require revision. Neither condition is currently in evidence.
[1] City of Meridian. Wastewater Consent Decree Annual Report No. 5. February 28, 2025. meridianms.org (PDF) — Tier A
[2] U.S. Environmental Protection Agency. Meridian, MS Clean Water Settlement Information Sheet. EPA Region 4 Enforcement. epa.gov — Tier A
[3] Mississippi Public Service Commission. Mississippi Power Utility Filing — Compass Datacenters Coal Unit Life Extension Request. January 9, 2025. psc.ms.gov — Tier A
[4] Meridian Star. "City Makes Progress on Consent Decree as Bond Issue Looms." April 17, 2025. meridianstar.com — Tier B
[5] Mississippi Today. "State, MS Power Extend Life of Coal Unit to Energize Data Centers." February 14, 2025. mississippitoday.org — Tier B
[6] Data USA / U.S. Census Bureau. Meridian, MS Profile. 2024. datausa.io — Tier B
[7] Meridian Star. "Todd Calls for County to Cut Taxes, Hike Pay with Datacenter Revenue." November 10, 2025. meridianstar.com — Tier B
[8] Center for Economic Accountability. "Mississippi Data Center Project Named 2025's Worst Economic Development Deal of the Year." December 29, 2025. economicaccountability.org — Tier C
[9] WBHM Public Radio. "Data Centers Bring Billions to Mississippi — Are the Investments Worth the Risk?" 2025. wbhm.org — Tier C
[10] All World PM. Meridian Consent Decree Program Overview. 2025. allworldpm.com — Tier C