Birmingham Humane Society Says Nebius AI Factory Cost It $18 Million

Nebius reached Google News for a conflict far beyond computing capacity. A Birmingham humane society says its planned 300-megawatt AI factory erased $18 million from a fundraising campaign.

The Greater Birmingham Humane Society, or GBHS, had spent years planning an animal welfare campus near the proposed data center. CEO Allison Black Cornelius now says uncertainty surrounding the facility drove away major donor commitments, according to the fundraising account. The reported loss remains the organization’s claim, not an independently audited finding.

That distinction matters, but it does not make the dispute less significant. Nebius says its design will control sound, recycle cooling water, and avoid raising residential electricity bills. GBHS argues that unanswered environmental questions already changed donor behavior before the facility began operating.

The primary conflict is therefore not simply Nebius against an animal shelter. It is the economic promise of AI infrastructure against the local uncertainty created while that infrastructure moves through planning and construction.

Google News readers encountering the headline may see an unusual fight involving dogs and servers. The more important story concerns who absorbs risk when an AI project advances before its neighbors accept the evidence behind its assurances.

What Changed Beside the Humane Society Campus

A proposed technology investment became an immediate financial problem for an organization that had planned its neighboring campus for a decade.

Nebius plans to develop its BHM01 AI factory at 201 Milan Parkway in Birmingham’s Oxmoor Valley. An AI factory is a data center designed around dense computing systems that train and operate artificial intelligence models.

The site covers roughly 75 to 80 acres near Lakeshore Parkway. Public reporting describes a planned capacity of 300 megawatts, enough to place the project within the hyperscale category.

GBHS owns 27 nearby acres intended for a consolidated animal welfare and medical campus. The organization began planning that development in 2016, well before the current data-center controversy.

Its proposed campus was designed to include an adoption center and a teaching hospital for fourth-year veterinary students. Plans also included animal enrichment areas, public walking trails, a cat café, and connections to Red Mountain Park.

The first phase, a Jefferson County Animal Services facility, has already been built. That makes the dispute more consequential than an argument over an undeveloped concept.

Cornelius has said the arrival of the AI factory created uncertainty among donors. The organization attributes an $18 million fundraising reduction to concerns about the project and its effects on the planned campus.

That figure should be understood precisely. GBHS is describing money it expected to raise or had associated with donor commitments, not a regulatory fine imposed on Nebius.

No public independent audit has established that the entire gap resulted solely from the data center. Donors can change plans for several reasons, including economic conditions and project uncertainty.

Still, fundraising depends on confidence. A donor considering a hospital or adoption campus must assess whether animals, visitors, students, and staff can use the property as intended.

Questions about continuous mechanical noise, vibration, construction disruption, backup generators, and future expansion can alter that assessment. The effect can arrive years before engineers switch on the first computing rack.

GBHS initially raised its concerns through public hearings and direct requests for more information. In February, Cornelius told local officials she was watching the planned campus “vaporize” as donors reconsidered their participation.

By July, the organization had moved from public opposition to litigation. Its complaint seeks judicial review of approvals and permits connected to the hyperscale facility.

The lawsuit names the City of Birmingham and development entities associated with the project. It represents a request for court scrutiny, not a final ruling that any approval was unlawful.

Construction activity has nevertheless begun at the site. That sequence created the central tension now spreading through Google News: physical development is advancing while the neighboring nonprofit says its own project is losing financial support.

The Humane Society cannot evaluate the conflict like a typical commercial neighbor. Dogs and cats can detect frequencies outside normal human hearing ranges, while sick or traumatized animals require controlled environments.

A sound level acceptable for an office may therefore produce different concerns beside outdoor kennels, veterinary facilities, and animal enrichment spaces. Whether the completed facility would cause that harm remains disputed.

The immediate change is easier to establish. The shelter’s campus plan now carries a new perceived risk, and that perception has entered its fundraising conversations.

Why the $18 Million Claim Raises the Stakes

The reported fundraising loss turns an environmental disagreement into a measurable example of how AI infrastructure can shift costs onto neighboring institutions.

AI data centers are often presented through direct benefits. Developers emphasize construction employment, permanent technical jobs, tax revenue, grid investment, and computing capacity.

Those benefits are easier to document because companies and development boards can model them. Local opportunity costs are harder to calculate, especially when they emerge through abandoned donations or delayed community projects.

GBHS has supplied one striking number. Its CEO says the data-center controversy cost the organization $18 million in fundraising for a campus designed around veterinary medicine and animal services.

That assertion does not establish legal liability. It does reveal how risk travels through a community before regulators or courts settle the underlying technical questions.

A donor does not need conclusive proof of future noise harm to withdraw. The donor only needs to believe that uncertainty makes another charitable project safer or more useful.

This mechanism creates an asymmetry. Nebius can continue developing its site while commissioning studies, designing barriers, and defending permits. The Humane Society must reassure contributors before it can finish its own campus.

Delay also affects the two sides differently. A large infrastructure company can incorporate legal and permitting disputes into a multiyear construction schedule.

A nonprofit depends on fundraising momentum, donor deadlines, and visible progress. A prolonged dispute can weaken a campaign even if later testing shows that some environmental fears were overstated.

This is why the fundraising claim belongs in a technology story. It offers a concrete measure of externalized uncertainty, meaning a risk experienced by people or institutions outside the project owner.

Birmingham’s broader policy response reinforces that point. The city imposed a temporary pause on certain new data-center applications and later developed rules addressing siting, environmental safeguards, infrastructure, and public review.

The city’s data-center guidance says the updated framework includes 20 protections for hyperscale facilities. Those protections cover compatibility with surrounding areas and environmental responsibility.

However, the Nebius project received permits before the newer framework took effect. Reporting indicated that the revised requirements would generally govern future developments, not undo existing approvals.

That timing leaves GBHS confronting a facility under earlier rules while Birmingham develops stronger standards for whatever comes next. It also fuels the argument that regulation followed the project instead of preceding it.

The city has an economic case for welcoming data centers. Birmingham has available land, transmission infrastructure, regional connectivity, and a workforce that can support large construction projects.

Officials also want the city to participate in the physical expansion behind artificial intelligence. Every widely used model depends on buildings, power systems, cooling equipment, network links, and specialized processors.

Yet the local bargain cannot be evaluated only through regional revenue. The relevant question is whether claimed public gains exceed losses imposed on specific neighbors.

GBHS says its planned campus would serve animals, veterinary students, adopters, and the wider public. If that campus shrinks or fails, the loss extends beyond one nonprofit’s balance sheet.

The $18 million figure remains contested in causal terms. Public evidence does not show which donors withdrew, what conditions accompanied their commitments, or how the organization calculated the total.

Those missing details should limit definitive conclusions. They should not erase the reported fundraising response, which is itself relevant evidence about confidence in the site.

For companies building AI infrastructure, the lesson is uncomfortable. Technical compliance alone may not protect a project from reputational or community costs.

Trust must exist early enough to influence decisions. A mitigation plan disclosed after neighbors organize, donors retreat, and lawsuits arrive is less persuasive than evidence shared before approvals.

The Google News Conflict Is Economic Promise Versus Local Risk

Nebius promises a valuable computing asset, while its neighbors argue that the project transferred uncertainty to them without adequate public review.

Nebius describes Birmingham as a strong location for a modern AI factory. The company highlights existing infrastructure, local skills, economic development, and demand for artificial intelligence computing.

Its public project overview says the facility will use a closed-loop cooling design. Such systems recirculate cooling fluid instead of continuously consuming fresh water through evaporation.

Nebius says annual water consumption will remain low and below that of the office complex previously located on the property. It also says the system will avoid fluid discharge into sewers or the environment.

The company promises to fund the new power infrastructure that Alabama Power builds for the facility. It says the project should not affect residential electricity bills.

Noise is another central concern. Nebius says an acoustic analysis found that planned mitigation would keep facility sound within existing ambient levels.

Its proposed measures include acoustic walls, equipment canopies, low-noise fans, and tree buffers. The company also says the site will comply with applicable noise ordinances.

Those are specific commitments, not empty branding. They offer clear criteria that regulators, residents, and independent engineers can test as construction and commissioning proceed.

However, they remain prospective claims. The finished Birmingham facility is not yet operating at full capacity, so actual performance has not been measured across seasons and operating loads.

GBHS disputes whether conventional acoustic analysis adequately represents conditions for animals. Its medical leadership has focused on low-frequency hum, vibration, and higher frequencies that humans may not notice.

That challenge does not prove that Nebius’s design will harm animals. It identifies a mismatch between the population used to define acceptable sound and the population living beside the facility.

An environmental test built around human hearing can satisfy one regulatory question while leaving an animal-welfare question unresolved. Outdoor kennels create another variable because barriers and buildings do not protect every activity equally.

Residents have voiced related concerns about construction noise, power demand, water use, air emissions, traffic, and property values. Their objections place the Humane Society within a larger neighborhood response.

Supporters see a different balance. Business and development advocates argue that strict rules can push investment toward competing cities, costing Birmingham jobs and long-term tax revenue.

Nebius representatives have said the facility will create more than 100 permanent jobs. Public presentations have also projected substantial annual government and education revenue from the development.

Economic development officials approved significant tax abatements for the project. The decision intensified debate over whether the expected public return justified the incentives and local disruption.

The dispute therefore has one primary opponent map: economic promise versus local risk. Nebius and GBHS are the visible parties, but they represent competing methods for valuing the same land.

The company measures output through computing capacity, investment, employment, and tax receipts. The Humane Society measures the site through animal health, veterinary education, donor confidence, and public services.

Neither framework is inherently illegitimate. The governance problem arises when one moves forward without producing evidence that the other side considers credible.

This is also why a Google News headline about dogs belongs beside stories about chips and cloud capacity. The physical AI buildout increasingly collides with institutions that never considered themselves technology stakeholders.

A hospital, school, neighborhood, farm, or shelter can become part of an AI supply-chain dispute because it shares a power grid, water system, road, or soundscape.

Once that happens, the question is no longer whether society wants artificial intelligence. It is where its supporting infrastructure goes, who approves it, and which risks count.

What Nebius Says the Numbers Do Not Show

The company’s mitigation plan answers several common objections, but it cannot yet resolve questions that depend on full-scale operating evidence.

Nebius says the project will fit into its surroundings through engineering controls. That position deserves direct consideration because some public criticism treats all hyperscale facilities as technically identical.

Cooling designs vary substantially. A closed-loop system can use far less ongoing water than an evaporative system, particularly after the initial fill.

Sound profiles also depend on equipment placement, fan design, barriers, operating loads, and distance from property lines. A study based on the actual layout is more informative than recordings from unrelated facilities.

Power effects depend on utility agreements and infrastructure funding. Alabama Power told residents that it had arrangements intended to prevent the project from raising rates through 2027.

These facts weaken claims that every feared impact is inevitable. They do not independently verify Nebius’s predictions across the facility’s complete operating life.

The 300-megawatt scale matters because later phases can change the acoustic and electrical profile. Measurements from an initial installation may not represent a fully built campus.

Operating conditions also change with temperature and computing demand. Cooling equipment can work harder during hot periods, while backup generators require scheduled testing.

Nebius says its noise controls will keep the completed facility within existing background levels. The most useful verification would measure performance at several boundaries after commissioning.

Those measurements should include the frequencies relevant to animals, not only aggregate levels designed around human annoyance. The testing protocol should also explain how it handles vibration and intermittent events.

Transparency is as important as the final number. Residents need access to methods, monitoring locations, operating conditions, and enforcement procedures.

A statement that a project complies with an ordinance tells the public little if the underlying measurements remain inaccessible. An independent review can reduce suspicion even when it confirms the developer’s analysis.

GBHS faces its own evidentiary burden. The organization should explain how it calculated the fundraising reduction without exposing confidential donors.

Useful disclosure could separate signed commitments, verbal pledges, expected campaign revenue, and contributions formally withdrawn. It could also identify whether donors cited the data center as their primary reason.

That information would not settle the environmental case. It would make the financial claim easier to evaluate and distinguish direct losses from delayed fundraising.

The legal process will examine different questions. GBHS’s court challenge seeks review of approvals and permits, including whether Birmingham followed the required procedures.

The complaint is one of several legal challenges associated with the project. At least two nearby residents also sued over the development and its construction effects.

A judge allowed one resident case to proceed while construction continued. That ruling did not determine the ultimate merits of every allegation.

For readers following the story through Google News, the safest conclusion is narrower than either side’s strongest position. Nebius has presented a detailed mitigation case, while opponents have identified unresolved verification and process questions.

The facility has not been shown to cause the feared operational harm because full operations have not begun. Conversely, proposed engineering controls have not established harmlessness under real, full-scale conditions.

The fundraising effect occupies a different category. GBHS says financial damage has already occurred through donor decisions, regardless of what later environmental measurements find.

That creates the story’s reversal. The infrastructure’s physical impact remains prospective, but one neighboring institution says the economic impact arrived first.

Birmingham’s Fight Mirrors a Wider Data-Center Backlash

The conflict is not an isolated dispute about one shelter, but Birmingham gives the national debate an unusually tangible community cost.

AI companies are racing to secure electricity, land, cooling systems, and network capacity. That demand has moved data centers from remote industrial zones into public debates across the United States.

Communities increasingly ask how many permanent jobs these sites create, who funds grid upgrades, and whether promised tax revenue justifies abatements. They also want enforceable limits on noise and resource use.

Birmingham illustrates the policy timing problem. Officials considered a moratorium and new rules after the Nebius project had already advanced far enough to remain outside key restrictions.

The city’s temporary pause applied to new large-scale applications, while projects with completed permits remained exempt. The policy acknowledged public concern without resolving the conflict that produced it.

In June, Birmingham adopted a new data-center ordinance after crowded hearings. It included requirements involving setbacks, environmental protections, noise studies, cooling systems, and infrastructure review.

According to local ordinance coverage, the council approved those rules on a 6-to-3 vote. GBHS argued they would not alter the existing Nebius development.

That divide resembles other disputes where communities strengthen rules only after a large proposal reveals gaps. The first project becomes both an exception and the reason for reform.

A nearby precedent emerged in Nashville during 2026. The Nashville Zoo opposed a proposed data center beside property it wanted for conservation education and animal programs.

The technical details differ, and the Nashville proposal was much smaller. However, both cases involve animal institutions arguing that conventional development review overlooks their specialized needs.

The comparison strengthens one GBHS argument. Animal welfare is not automatically captured by standards built around human residents, offices, or industrial users.

It also gives developers a clearer path. Early consultation with animal-care specialists can identify frequency, vibration, lighting, traffic, and emergency-power concerns before a plan hardens.

The national trend does not mean every data center is unacceptable. It means projects face a higher burden to connect technical claims with independently reviewable local evidence.

For enterprise buyers, developers, and AI users, this is not merely a public-relations issue. Delays, lawsuits, permit disputes, and community resistance can affect when computing capacity becomes available.

The physical layer behind cloud AI is becoming a deployment risk. Companies buying capacity must assess whether announced megawatts have secure power, durable permits, community acceptance, and realistic construction schedules.

Knowledge workers also have a stake. Every AI assistant, model query, generated image, and automated workflow relies on infrastructure located in a real community.

Tools can make digital work feel detached from geography. The Birmingham fight shows that compute always has an address, neighbors, and a regulatory history.

Following complex disputes requires more than saving headlines. A personal knowledge base can connect permits, public statements, court filings, and later measurements without treating every new claim as settled fact.

That method matters because stories change quickly. An early article may describe a proposal, while later coverage reports construction, litigation, incentives, or revised rules.

Google News helps surface those developments, but aggregation does not reconcile contradictions. Readers still need to distinguish company projections, nonprofit claims, court allegations, and independently measured results.

Birmingham’s lesson is therefore broader than opposition to one AI factory. Infrastructure planning now needs a local evidence strategy as carefully designed as its power and cooling systems.

Three Signals Will Decide What Happens Next

The next phase should be judged through operating measurements, court decisions, and the Humane Society’s documented fundraising trajectory.

The first signal is independent environmental testing. Nebius expects initial commissioning and energization as development proceeds, followed by a longer buildout.

Testing should measure noise at the Humane Society property and nearby homes under meaningful computing loads. It should cover high frequencies, low-frequency sound, vibration, and generator testing.

Water use should also be compared with the company’s closed-loop projections. Public reporting should distinguish initial filling from recurring consumption.

If independent results remain within promised limits, Nebius’s argument will strengthen. GBHS would still have a legitimate process complaint, but claims of unavoidable operational harm would become harder to sustain.

If measurements exceed predictions, the company will face pressure to add controls, restrict operations, or compensate affected neighbors. Such a result would also influence rules for future Birmingham projects.

The second signal is the litigation. Courts will determine whether the Humane Society and resident challenges can proceed and whether the disputed approvals followed Alabama law.

A ruling against the permits would weaken the assumption that early approvals insulated the project from later scrutiny. It could delay construction and reshape how Birmingham reviews similar developments.

A ruling favoring the city and developers would narrow opponents’ legal options. However, it would not independently validate every environmental prediction or erase the reported fundraising loss.

The third signal is donor behavior. GBHS must show whether its campaign stabilizes, loses additional commitments, or recovers after new studies and court decisions.

A transparent update would help establish whether the $18 million figure represents a temporary delay or a durable loss. It would also show whether mitigation agreements can restore confidence.

These signals should appear in that order because physical evidence addresses the central operational dispute. Legal decisions address procedural legitimacy, while fundraising reveals the continuing community consequence.

Several possible outcomes can coexist. Nebius could comply with environmental limits while GBHS still proves that a poor public process damaged its campaign.

A court could uphold the permits while later measurements require stronger mitigation. Donors could also return if independent experts validate the site design and binding protections emerge.

Readers should resist treating each hearing or filing as a final verdict. The project’s actual performance will matter more than either side’s current forecasts.

The Humane Society’s claim nevertheless changes the standard for evaluating AI infrastructure today. It gives Birmingham a concrete question that other cities will soon face.

When a technology project promises regional benefits but imposes uncertainty on an established neighbor, who must pay to resolve that uncertainty?

Nebius can answer with accessible studies, independent monitoring, and enforceable remedies. Birmingham can answer through transparent approvals that apply before projects become exceptions.

GBHS can strengthen its position by documenting donor decisions and defining the animal-specific standards it believes the site must meet. Courts can determine whether the approval process stayed within the law.

For anyone tracking the dispute through Google News, those are the developments worth saving and comparing. Do not measure the story only through construction progress or courtroom rhetoric.

Watch the boundary measurements, the permit rulings, and the fundraising record. Together, they will show whether Birmingham gained an AI asset, displaced a public-serving campus, or found a workable way to keep both.