Cancer Alley's Warning for the AI Data Center Boom

Oct 4

Oct 6, 2026 · Vital Biomedical News Staff

Louisiana spent decades learning what happens when heavy industry arrives faster than the health science. AI data centers are now being built at record speed, with many of the same open questions.

What "Cancer Alley" means

Along an 85-mile stretch of the Mississippi River between Baton Rouge and New Orleans sit more than 150 chemical plants, in mostly Black communities (NBC News). Residents and activists have called it "Cancer Alley" since the 1980s.

The best-documented case is St. John the Baptist Parish. In 2018, EPA data showed a census tract there faced the highest cancer risk from toxic air pollution in the nation, with 85% of that risk traced to chloroprene emissions from the Denka neoprene plant (Earthjustice). EPA had classified chloroprene as a likely human carcinogen in 2010. Air monitoring at nearby Fifth Ward Elementary found children exposed at levels 11 times what EPA considers acceptable, and the agency concluded the risk fell disproportionately on Black residents (ProPublica).

The plant cut chloroprene emissions by about 85% after a 2017 settlement with the state. The federal government sued Denka in 2023 to force deeper cuts, then dismissed that case in 2025. Denka maintains that EPA overstated the risk (Louisiana Record). In May 2025, Denka suspended production at the site, citing financial losses (Earthjustice).

Does Louisiana have the highest cancer rate?

Not quite, though it is close to the top. In 2024, Louisiana's age-adjusted cancer death rate was 158.3 per 100,000 people. Kentucky (175.6) and Mississippi (167.8) were higher (KFF). Nationally, the rate was 141.5 in 2023 (USAFacts).

Statewide rates reflect many causes, including smoking, obesity and access to care. That is why the more precise question is not whether Louisiana has high cancer rates, but whether people living near industrial pollution have higher rates than they otherwise would.

The fight over the evidence

For years, state officials pointed to Louisiana Tumor Registry reports showing a mix of higher and lower cancer rates across the industrial corridor (Tulane Law). Those reports did not measure pollution exposure.

In 2022, researchers at Tulane's Environmental Law Clinic published a peer-reviewed study that did. They compared neighborhood-level cancer data from the registry with EPA estimates of cancer risk from air toxics. Among poor neighborhoods, those with heavy air pollution had higher cancer rates. Poor neighborhoods with little pollution had rates close to the state average, and affluent neighborhoods showed no link at all (WWNO).

A 2025 study in Frontiers in Public Health found a similar pattern: cancer incidence tracked with air toxics, compounded by poverty and minority status. It also found some areas with the opposite pattern, which the authors attributed to old air-quality data, long cancer latency and possible underreporting (Frontiers). LSU Health, which runs the registry, has publicly objected to how its data has been characterized (LSU Health).

The core lesson is timing. Cancer can take decades to develop. By the time registries can detect a pattern, the exposure has already happened.

The data center boom comes to Louisiana

In December 2024, Meta announced a $10 billion AI data center in rural Richland Parish, in northeast Louisiana. The project, called Hyperion, has since grown dramatically. Entergy Louisiana will build 10 gas-fired power plants to supply it, producing more than 7 gigawatts (Fortune). State regulators approved the first three plants in August 2025 and put the other seven on a fast-track schedule, with a final vote expected in December 2026 (Union of Concerned Scientists).

One of the first three plants is being built at Entergy's Waterford site in St. Charles Parish (DCD). St. Charles is part of the same river corridor at the center of the Cancer Alley debate.

What has happened elsewhere

Memphis, Tennessee. Elon Musk's xAI built its Colossus supercomputer in South Memphis and powered it with dozens of gas turbines. The company avoided air permits by classifying the trailer-mounted turbines as "non-road engines." University of Tennessee researchers found the turbines added to air pollution, and EPA closed that loophole in January 2026 (CNBC). The turbines emit smog-forming nitrogen oxides and formaldehyde (Southern Environmental Law Center). The nearby Boxtown neighborhood is historically Black and already sits beside a refinery and a power plant (Climate & Capital Media). A city-commissioned analysis found the air safe, but the researcher acknowledged large uncertainty because the model relied on emissions data supplied by xAI rather than independent monitoring (TIME).

Northern Virginia. The world's largest data center cluster depends on more than 13,500 permitted backup generators, nearly half of them in Loudoun County. Residents report constant noise and lost sleep (Technical.ly). Most are diesel generators without pollution controls (Virginia Mercury). Beginning July 2026, new permits are generally expected to meet cleaner Tier 4 standards (Encino Environmental). The industry points out that a state legislative review found backup generators run only rarely (WVIR).

What we know, and what we don't

What is known. A data center's health footprint comes mainly from burning fuel: gas power plants, on-site turbines and diesel generators. The pollutants involved, including nitrogen oxides, fine particulate matter and formaldehyde, have well-established links to asthma, heart and lung disease, and some cancers.

What has been estimated. Researchers at UC Riverside and Caltech estimate that pollution tied to data center power demand could cause up to 1,300 premature deaths a year by 2030, with public health costs near $20 billion annually (UC Riverside). A 2026 update found more than 90% of the estimated harm comes from the power plants supplying electricity, not from the data center sites themselves (Environmental Protection Network). Some analysts argue these figures may be overstated (Dev Sustainability).

What is unknown. No long-term study has followed the health of people living beside hyperscale data centers. Nearly every figure above is a model, and some rely on emissions numbers reported by the companies themselves. The effects of round-the-clock noise on sleep and heart health near these sites have not been studied. And if there is a cancer signal, it would not appear in registries for 10 to 30 years.

Can we learn the lessons this time?

Cancer Alley offers a practical checklist:

Measure before building. Independent air monitoring should establish a baseline before operations begin, rather than relying on models built from company data.
Count what's already there. Boxtown and the Mississippi River corridor were heavily polluted before any data center arrived. Permits should consider cumulative exposure, not each source alone.
Close loopholes before, not after. In Memphis, turbines ran without permits for over a year before federal rules changed.
Watch where projects land. Cancer Alley and Boxtown are both majority-Black communities that absorbed industrial risk with little say.
Track health over time. Long-term health monitoring near large facilities would let problems surface in years, not decades.
Push for cleaner power. If most of the harm comes from power plants, the energy source matters more than anything else.

The bottom line

Data centers are not chemical plants. A server hall does not release chloroprene, and the evidence does not show that data centers are creating a new Cancer Alley. But the power behind them burns fuel, often in communities that already carry heavy pollution burdens, and the projects are moving faster than the science.

Louisiana's experience shows the cost of answering health questions decades after the permits are signed. The open question is whether regulators will ask them now.