Kelsey Birch remembers the crushing pressure, the sharp jerks, and the feeling of hands inside her abdomen. "I felt like I was being torn apart," she said. "I couldn't handle it."

Birch, a then 28-year-old emergency dispatcher in Colorado, hadn't gone into labor expecting surgery. She was healthy. Her first pregnancy had been going well, and her full-term baby was positioned face down — an ideal candidate for a vaginal birth. As her contractions grew stronger, a delivery room nurse urged Birch to get an epidural before the hospital anesthesiologist finished his shift. Birch agreed.

The American College of Obstetricians and Gynecologists recommends that a woman with an epidural in first-time labor push for up to four hours. Birch pushed for just over two before a doctor intervened. Her labor had failed to progress, her medical records noted, so, like one in three American women who give birth, Birch underwent a cesarean section.

The half-hour surgery began months of recovery. For weeks, Birch struggled to hold and breastfeed her newborn daughter. Her incision site was searingly painful; first as a fresh wound and again as it healed, scar tissue adhering to skin and muscle. Like many women who undergo a C-section, she learned to cope with intrusive memories of the surgery.

C-sections are now the most common inpatient surgery in America. Business Insider analyzed years of hospital delivery data from Florida, Mississippi, and Iowa and found that C-section rates can differ sharply between hospitals just miles apart — even when caring for similar, low-risk patients. Those gaps, experts say, point to how a hospital's policies, culture, and financial incentives can push surgery rates higher, helping explain why the US consistently performs far more C-sections than global health authorities consider necessary.

In 1970, just over 5% of babies in the US were delivered by C-section nationwide — too low to protect some women and their babies. That rate tripled within a decade. Now, for nearly 20 years, over 30% of babies in the United States have been born by C-section, around double the rate the World Health Organization says is "ideal" for maternal and infant health. US doctors perform these surgeries 60% more frequently than doctors in France, and nearly double the rate of Finland and Sweden.

Many of them, maternal health experts told Business Insider, are medically unnecessary. Federal health agencies consider women with low-risk pregnancies — those who are pregnant for the first time, are at full term, are not delivering twins, and whose babies are head-down rather than breech — the least likely group to require surgery to safely deliver their babies. Yet more than a quarter of those women delivered by C-section in the US in 2023, the most recently available year with national data.

One large study estimates that as many as 19% of all births should be C-sections to protect women and their babies, leaving as many as 13% possibly performed needlessly in the US.

That suggests roughly one in 10 pregnant American women — nearly half a million women each year — are operated on unnecessarily, undergoing major abdominal surgeries that carry a higher risk of hemorrhage, blood clots, and infection, and leave women more likely to develop dangerous complications in their future pregnancies.

It's not because women are asking for more C-sections. Only an estimated 2.5% of US babies are born by elective C-section each year. Instead, experts say, the reason behind the high C-section rate largely comes down to the structure of the American medical system.

C-sections are more profitable, predictable, and are perceived to be more protective against lawsuits than vaginal births, according to experts and multiple studies. They say health systems looking to keep operating rooms full, revenue up, and liability risk minimized are indirectly incentivized to keep surgery rates high.

Put plainly, a high C-section rate may just be better for business.

When C-sections hurt more than they help

C-sections are often critical, sometimes lifesaving procedures.

In medical emergencies, like when the placenta completely blocks the cervix or an umbilical cord slips down past the baby, a C-section is mandatory. Failure to promptly perform the surgery can disable or kill the woman, her baby, or both. Other factors — including a pregnant woman's age, whether she's obese, or if she's diagnosed with diabetes or pre-eclampsia — all increase the chance that a C-section is the safest way to deliver a baby.

To deliver a baby by C-section, a surgeon dissects through seven layers of body tissue: skin, fat, fascia, muscle, peritoneum, uterus, and amniotic sac. Doctors first use a scalpel; later, they may use their fingers or a blunt tool. Like any abdominal surgery, it also presents immediate risks, including infections, hemorrhages, blood clots, or injury to other organs.

The risks increase with additional C-sections, and a woman is far more likely to have another C-section once she's had her first. In 2021, 86% of women who previously had a C-section underwent the surgery again to deliver their next baby. While studies show that as many as 80% of women who have undergone a prior C-section can safely deliver their next baby vaginally, many hospitals nationwide ban women who have had a C-section from attempting a later vaginal birth, citing safety and liability concerns. Each additional surgery will increase her risk of developing severe complications, like placenta accreta, which almost always requires a hysterectomy to prevent hemorrhaging.

These complications can sometimes turn deadly. While rare, the United States has the highest rate of maternal mortality among high-income countries — 18.6 deaths per 100,000 live births in 2023 — and a woman who receives a C-section is up to five times more likely to die due to pregnancy-related causes than a woman who delivers vaginally.

Even when the surgery goes as expected, some women describe the experience as extreme, even traumatic. Women are anesthetized, but most choose to remain conscious to witness the birth of their child. While some women who spoke with Business Insider said their surgeries were unremarkable and felt little of the procedure, others told Business Insider they still felt the snagging pull of the scalpel, intense pressure, and sometimes shocking pain. Their memories of the surgery are punctuated with specific detail: their arms held outstretched, pinned in restraints secured to the operating table; their bodies trembling; and their newborns held amid IV lines and the surgical drape.

After a C-section, some women endure weeks or months of painful recovery, much longer than women who deliver vaginally. Women told Business Insider they struggled to breastfeed, shower, and get out of bed to care for their newborns. Some suffered painful infections along their 4- to 6-inch surgical incisions. Others weathered panic attacks, nightmares, and lingering anxiety.

The hospital you choose can affect how likely you are to get a C-section

Doctors delivering babies have to make tough calls. If a fetal heart rate monitor alerts that a baby is in distress, or if a woman's labor is so prolonged that she or her baby is at risk of suffering severe injury, a doctor might feel compelled to intervene.

Even if the medical necessity of C-section is ambiguous, "I can't really gamble," said Dr. Elizabeth Langen, an OB-GYN and a clinical associate professor at the University of Michigan who studies C-section rates. Doctors may choose to perform a C-section because "it's devastating when a baby is born and isn't doing well," she said. "You keep thinking, 'Is there something else I could have done?'"

The health and safety of a woman and her baby aren't always the only influences at play.

Controlling for a constellation of factors — hospital obstetric care levels, delivery volume, urban or rural location, maternal age, race, health, and income — multiple studies show that one of the strongest predictors of whether a woman will get a C-section is which hospital she delivers in.

Business Insider requested data on hospital delivery data from all 50 states and DC, and analyzed data from three states that responded first to the request or had already published the data publicly: Florida, Mississippi, and Iowa. Business Insider's analysis reveals huge variations in C-section rates from one neighboring hospital to another — suggesting a woman's chance of undergoing a C-section is affected by her choice between two close-by hospitals.

A 15-minute drive is enough to swing outcomes. In 2019, the average rate of C-sections at one Palm Beach County hospital in Florida was 26% higher than at a hospital less than 5 miles away in a similarly majority white, wealthy area. Between 2010 and 2019, Business Insider found similar instances in Broward, Duval, Miami-Dade, and at least five other Florida counties. The hospitals all had similar levels of obstetric care and numbers of babies delivered.

Similar patterns were found in Mississippi, where C-section rates are among the highest in the country, as well as in Iowa, where C-section rates are among the lowest nationwide. Doctors at a hospital in northwest Iowa, performed surgeries on women who were pregnant for the first time at a rate 65% higher than doctors at a comparable hospital 15 minutes away, according to Business Insider's analysis of aggregate five-year hospital birth data from 2019 to 2023.

For C-sections performed on low-risk pregnancies, differences in hospital rates may be more telling, experts told Business Insider, as surgeries performed in this category are more likely to have been medically unnecessary.

In Mississippi, which has the highest rate of C-sections on low-risk pregnancies in the country, some hospitals had low-risk pregnancy cesarean delivery rates that swung as low as 14%. Doctors at other hospitals performed C-sections on 45% of all women with low-risk pregnancies, according to the analysis of data from 2022.

How a hospital operates — its workplace culture, administrators' priorities, and the type and number of staff employed, particularly nurses or midwives — can dramatically impact how many cesarean deliveries its doctors perform, said Dr. Emily White VanGompel, a family medicine doctor and professor at the University of Illinois Chicago who studies how organizations impact C-section rates.

Hospital administrators and care staff can foster cultures more supportive of vaginal birth, White VanGompel said. Conversely, they can also implement policies and care standards that cause doctors to intervene more in a woman's labor.

Some hospitals that treat women with high-risk pregnancies may require a high C-section rate to most safely deliver their patients' babies. A risk-adverse hospital may also lead doctors to perform more medical interventions in part to minimize its risk of a lawsuit, said Louise Roth, a sociology professor at the University of Arizona who studies how organizations and medical malpractice influence maternal health, including C-section rates.

Doctors may then practice medicine in ways that evidence-based studies show can increase the chance of a C-section: induce labor more often, continuously monitor more fetal heart rates in low-risk pregnancies, and interrupt women who push for more than a few hours — all procedures that are more insulating in the case of a lawsuit, Roth said.

OB-GYNs are among the doctors most likely to face a legal threat at least once in their career, a 2023 study found, and a lawsuit filed over the serious harm or death of a baby can result in tens of millions of dollars in jury awards.

"People often will teach young obstetricians that you don't get sued for doing a C-section, you get sued for not doing a C-section," Langen said.

Roth said doing more also satisfies another hospital demand: billing more procedures and garnering more reimbursements.

"Intervening more almost always means a C-section," Roth said, "But less intervention means less money."

In Kelsey Birch's case, the hospital where she delivered disclosed in a 2025 survey that it had an average C-section rate 22% higher than the state average.

Why the US performs so many C-sections

A C-section is major surgery. It's also lucrative.

Providers in the 1970s and 1980s responded to "market shocks" — dips in the birthrate that undercut their profits — by performing more highly reimbursed C-sections, a pioneering 1996 study found. A 2017 study revealed that California women, meanwhile, were 12% more likely to receive a C-section if their doctor was reimbursed just $420 more for the surgery than for a vaginal delivery.

On average, US insurers in 2020 paid $17,103 for a C-section and $11,453 for vaginal birth, according to one 2022 study. Physicians were paid about 16% more, on average, for a C-section than a vaginal delivery, according to a 2015 study. Because a C-section takes much less time, it's more cost-effective for both providers and hospitals.

While researchers in the early 2000s found that pregnant women who are older, obese, or who have other preexisting health issues undergo C-sections at higher rates, more recent studies have since found that controlling for these and many other patient factors doesn't eliminate the wide swings in C-section rates between hospitals or explain the high C-section rate nationwide.

Research has repeatedly shown the connection between profits and surgical deliveries. Women who gave birth at for-profit hospitals were more likely to deliver by C-section; women were more likely to undergo a C-section at hospitals that reap higher profits from each C-section than hospitals where the procedure is less lucrative; and Florida obstetricians increased the number of C-sections they performed after their practices were acquired by management companies that prioritized increasing practice revenue.

Some women are more likely to get the surgery. Black women nationwide undergo C-sections at higher rates than white women. Doctors in hospitals across New Jersey, for example, were much more likely to give Black women unnecessary C-sections to fill empty operating rooms and maximize payouts.

Directly incentivizing doctors to perform more C-sections is illegal, but other incentives like bonus payments for bringing in more money to the practice or mandatory practice revenue thresholds can influence doctors to increase the number of surgeries they perform, said Ambar La Forgia, an associate professor at Haas School of Business at the University of California, Berkley, and the author of the Florida study.

"It's a known fact that C-sections reimburse more highly than a vaginal birth, but they're also easier to do and easier to schedule," said La Forgia, making them attractive procedures to healthcare companies looking to boost their bottom line.

In some cases, overuse becomes criminal. In January 2025, the Department of Justice accused Chesapeake Regional Medical Center in Southeast Virginia of fraudulently billing for unnecessary procedures, including C-sections tied to the hospital's former chief obstetrician, Javaid Perwaiz. Prosecutors said Perwaiz routinely falsified records to justify early labor inductions and C-sections timed to his weekend surgery block. The pattern was so blatant that hospital staff nicknamed the resulting NICU admissions the "Perwaiz special," according to a federal indictment.

The indictment called the fraud "patently obvious" and an "open secret" that was documented repeatedly to hospital leadership and ignored. The hospital turned a blind eye, the indictment said, because Perwaiz's actions reaped a financial reward.

The hospital had been aware of red flags around Perwaiz's history for decades, the indictment charged. Following Perwaiz's federal tax fraud conviction in 1996, another obstetrician at Chesapeake Regional warned the Virginia medical board that "unnecessary gynecologic surgery is a growth industry in Chesapeake," according to the indictment, and that Perwaiz might be incentivized to perform even more surgeries to pay off his fraud debts.

Around the same time, the indictment charged, Chesapeake Regional's president wrote a letter to the medical board supporting the reinstatement of Perwaiz's medical license, citing his value to the hospital's bottom line.

Perwaiz was charged in a separate criminal case in 2019 and convicted in 2021 for performing medically unnecessary surgeries, including C-sections. He was sentenced to 59 years in prison. His lawyers declined to comment on Business Insider's reporting.

Dr. Nick Oberheiden, Chesapeake Regional's attorney, told Business Insider that the hospital denied any wrongdoing. "We are confident that the court and a jury will find, based on the facts, the law, and the evidence, that Dr. Perwaiz was the sole wrongdoer," Oberheiden said.

The case against Chesapeake Regional is ongoing.

C-section rates can be lowered — if hospitals and states make it a priority

The C-section rate at each individual hospital isn't an intractable problem, said White VanGompel, the family medicine doctor. Since hospital policies, administration priorities, and staff culture have so much bearing on increasing C-section rates, she told Business Insider, changing policy, priorities, and culture can also lower them.

In a 2024 study, White VanGompel and co-authors studied hospitals in California and Florida that focused on safely lowering their C-section rate and managed to reduce cesarean births by 5% for at least 18 months.

They found the most successful hospitals had strong leadership support, communication between care teams supportive of vaginal delivery, and a culture that empowered nurses to advocate for their patients without retribution. (Nationwide, hospitals with more nursing staff or that employ midwives are also associated with lower C-section rates.)

California also shows what's possible at the state level. In 2015, the California Maternal Quality Care Collaborative launched a campaign with policy recommendations, financial incentives, and new public transparency requirements — including publicly publishing all hospitals' annual low-risk C-section rates — aimed at lowering low-risk pregnancy C-section rates.

In 2014, California's low-risk pregnancy C-section rate mirrored the country's nationwide: 26%. Five years later, the rate had declined to 22.8% — a 12% drop.

States like Wisconsin, Illinois, and New Jersey have also seen success with initiatives meant to lower C-section rates.

These kinds of state policies and hospital-level interventions aren't just niceties, only implemented when it's convenient for providers and hospitals, White VanGompel said.

Women's health is at stake, White VanGompel said. "This is what actually gets results."

Back in Colorado, Kelsey Birch wrote a complaint letter to the administrator of the hospital where she gave birth. "I now have a scar on my uterus, which could affect any subsequent pregnancies and births," she wrote. She feared her providers had rushed her first with the epidural, and then again as she attempted to give birth vaginally.

"I believe my cesarean could have been avoided."

Have a tip? Contact Hannah Beckler via email at [email protected] or Signal at hbeckler.72. Use a personal email address and a nonwork device; here's our guide to sharing information securely.

Columbus, Ohio, escaped the Rust Belt rut years ago.

Regional economic development officials offered incentives that attracted warehouses, manufacturing plants, and healthcare startups, reviving the economy and generating jobs. By 2018, hundreds of these deals over the previous eight years had created some 150,000 jobs.

Central Ohio now hopes to repeat that success. It's betting big on "Silicon Heartland," a high-tech innovation hub that proponents hope will be flush with high-paying jobs. Economic officials have dangled multimillion-dollar tax subsidy packages before some of the world's biggest technology companies.

The resulting investment, Gov. Mike DeWine promised, "further cements Ohio as the heart of our nation's technology and innovation."

Mostly, they're getting data centers. Central Ohio has become one of America's hottest hubs for these computing warehouses, with companies including Amazon, Google, Meta, and QTS flocking there, lured largely by generous incentives.

The problem: Data centers, which operate largely autonomously, don't produce many lasting full-time jobs.

A Business Insider analysis of construction permits, economic development deals, and company disclosures found that even the largest data centers generally employ fewer than 150 permanent workers, and some have as few as 25. Building those data centers also creates significant numbers of construction jobs, but those are short term, sometimes lasting less than a year — far shorter than the duration of the tax breaks the companies get, which often last a decade or longer.

That means the tax breaks given to developers can amount over time to more than $2 million for every permanent, full-time job at an operational data center, Business Insider's analysis found. That's roughly eight times higher than the $262,000 average per job that watchdog group Good Jobs First found in 18 economic development deals worth at least $50 million awarded in 2023.

The number of jobs doesn't balance the cost, multiple economists and researchers who study tax subsidies told Business Insider — even factoring in the construction and other supporting roles that the tech industry uses to calculate its economic impact. Records show that the workforce on data center projects quickly tapers off, meaning industry estimates often significantly overstate long-term employment benefits.

The costs to the public don't end with tax subsidies. Data centers drive up electricity costs for other ratepayers as utility operators invest billions of dollars in new grid infrastructure to support escalating power demands. That has drawn opposition from other companies including retail giant Walmart, which has said that surging electricity bills are imperiling its expansion in states such as Ohio and Virginia.

Industry advocates argue the deals are worth it.

"Each new data center built in Ohio spurs a significant boost in investment, revenue, and wages that flow to Ohio businesses and workers, stimulating the state's economy," Josh Levi, the president of the Data Center Coalition, an industry advocacy group, wrote in an August 2024 op-ed article published by Cleveland.com. In recent US congressional testimony, he cited an estimate that data centers in Central Ohio supported more than "10,000 construction jobs, 2,000 data center jobs, and hundreds of maintenance and retrofitting jobs last year."

Drilling into the terms of specific economic development deals suggests a more complicated picture.

In 2021, for example, Google entered into a much-celebrated deal with Columbus to construct a data center campus. The city offered a 100% property tax abatement worth an estimated $54 million in tax savings over 15 years.

In exchange, the Google facility promised 20 full-time jobs at the data center, rising to about 40 jobs by 2047.

Pricing the 'Silicon Heartland'

Artificial intelligence is accelerating data center construction that already was growing quickly to power digital services from social media to medical care. In 2025 alone, Meta plans to spend at least $64 billion on facilities and equipment. Google's parent company, Alphabet, plans to spend $75 billion, and Microsoft said it would invest $80 billion.

Tech companies say their investments will supercharge local tax revenues and high-paying jobs will drive economic growth. Even with tax breaks, data centers contributed $162.7 billion in federal, state, and local tax revenue in 2023, according to a February 2025 PwC report prepared for the Data Center Coalition. The industry, the report said, supported 4.7 million jobs directly at data centers or indirectly through their supply chain.

Amazon, the biggest data center operator, calculates that its data centers each year have supported thousands of jobs, including 6,490 in Ohio and 20,700 in Virginia. Matt Hurst, a spokesperson for Amazon Web Services, Amazon's cloud-computing arm, told Business Insider the company was "proud of the good jobs we create, for the trust local communities invest in us, and for the opportunity we have to invest in those communities."

Meta says that its data center operations support 16,000 jobs and $1.2 billion in labor income annually, and that it has backed 440,000 construction jobs over the past decade. Google says its data centers supported 119,000 jobs and contributed $12.6 billion to US gross domestic product in 2023 across its supply chain, including construction. Microsoft's website says its data centers generate "public infrastructure improvements and tax revenue that serve as a catalyst for enhancing the quality of life."

"Our developments generate millions of dollars in tax revenue to support local priorities related to schools, roads, housing, and other critical needs, while also reducing the tax burden on residents," a spokesperson for QTS, which is owned by the investment firm Blackstone, said in a statement.

A Blackstone spokesperson also highlighted the benefits of data center development and said the company was "proud that our investment in QTS provides the digital infrastructure critical to the future of our country and economy."

Competition to score these promised benefits can be a race to the bottom, as developers pit state against state and city against city. New projects cluster in areas that offer the most competitive deals.

To investigate how these incentive deals play out, Business Insider identified areas of data center development and filed requests with all 50 states and Washington, DC, for the air permits that regulate backup generators at every data center. Business Insider compiled records for 1,240 data centers nationwide, the most definitive accounting to date, and requested records of data-center-related economic incentives from municipalities and states.

The largest data centers in Business Insider's analysis — the 322 massive facilities that we estimate consume 40 megawatts of electricity or more each — are heavily concentrated in a few places. Northern Virginia has 214, followed by Arizona's Maricopa County with 16, and Ohio's Columbus region with 9.

Thirty-seven states have tax incentive programs for data center investments. Most exempt developers from sales and use taxes on building materials, machinery, or equipment — resulting in big hits to state coffers. In Virginia, 56 data center projects cost $928 million in abated state sales tax in the 2023 fiscal year alone. Disclosures in Ohio estimate it forfeited nearly $360 million in data-center-related state tax revenue from the 2022 through 2024 fiscal years.

Mason Waldvogel, a spokesperson for the Ohio Department of Development, called the tax incentive program "a strategic tool used to create long-term economic growth by attracting high-value, capital-intensive projects." A spokesperson for the Data Center Coalition said state tax exemptions for data centers were consistent with programs for other capital-intensive industries.

Cities also offer incentives, including breaks on property taxes and reimbursements for building fees. Arizona cities largely don't give property tax abatements but allow the use of precious water resources. Virginia grants access to enormous amounts of electricity and critical infrastructure but requires data centers to pay local property taxes. Indeed, Northern Virginia cities generate up to 31% of their total tax revenue from data centers, funding fire departments, affordable housing, and other services.

In the Columbus region, Business Insider located 19 data center-related deals that, together with state-level abatements, amounted to at least $750 million in forfeited tax revenue for 770 full-time jobs employed at data centers as of December 2023.

The jobs generally pay well, averaging $100,000 a year in Central Ohio, according to company disclosures. At the Google data center in Columbus, salaries range from $74,000 for a data center technician to $162,000 for an operations manager.

Amazon tops the list with seven deals. In one, the northwest Columbus suburb of Dublin agreed to sell Amazon 66 acres, which the city valued at $100,000 an acre, for $1 in total. Amazon agreed to pay the farmers previously leasing the land up to $40,000 total to abandon their soybeans and corn crops and terminate the lease. It told Dublin it expected to hire 25 full-time workers by the end of 2018, a nonbinding projection. In contrast, Amazon projected that it would hire 1,000 Ohioans at a new fulfillment center in Canton several years later — without taking any local property tax abatements or state incentives.

Amazon's Hurst said the company works hard to create every job it projects.

'Let 'em walk'

The deals keep coming, from Batavia, New York, to Meridian, Mississippi.

Nathan M. Jensen, a professor at the University of Texas at Austin who studies regional tax incentive programs, said cities are better off sitting these deals out. Communities throw everything they can at tech companies, yet when the costs of lost tax revenue and escalating electricity prices are factored against what the communities get back in jobs, revenue, and prestige, "there's just no evidence that you're going to benefit from that data center," he said.

If data center developers threaten to walk from cities that refuse to compete for these deals, Jensen's advice is blunt: "Let 'em walk."

Jensen said data centers were shaping up like professional sports stadiums, where cities give millions in tax revenue savings in exchange for temporary construction jobs and minimal economic impact. Construction of data centers generally lasts one to two years, or sometimes longer, and many construction jobs run for only part of that period.

In Virginia, one analysis found that about 80% of jobs from data centers created over a recent two-year period were in construction. And the numbers of such data-center-supported jobs cited in this year's Data Center Coalition report may be misleading, multiple economists and researchers who study incentives told Business Insider.

Timothy Bartik, a senior economist at the W.E. Upjohn Institute for Employment Research, a not-for-profit organization focused on reducing unemployment, said his own study suggests job numbers in the high-tech sector, like data centers, could be less than half of industry estimates.

Microsoft estimated last year that a campus with six data centers that it is building outside Cheyenne, Wyoming, would have 1,005 jobs at peak construction, falling to 335 full-time employees and contractors by the end of next year.

At a construction project in Columbus for the data center operator Cologix, one contractor, Baker Concrete Construction, had 63 people on payroll. Those jobs lasted an average of 6 ½ weeks. Cologix said that overall the site had an average of 146 workers during the project's construction.

Incentive packages often spell out how many jobs a company commits to creating in exchange for its tax breaks. Data center companies generally commit to deliver only the jobs inside their facilities in exchange for their tax breaks — not the construction and other ancillary jobs they say their projects create. Based on what is actually promised in such deals, those jobs can be expensive for local governments. Business Insider identified five deals in Ohio where, as of December 2023, each long-term job in the data centers cost over $1 million in abated taxes over the life of the deal.

An Amazon data center in Hilliard had saved at least $195 million in state and local taxes as of December 2023, according to annual disclosures, driving the price of each job to over $1 million in abated taxes. New Albany, Ohio, garnered 98 jobs at a Meta data center, but forfeited $189.6 million in state and local taxes as of the end of 2023 — making each job worth about $1.9 million in foregone tax revenue.

"We disagree with this way of thinking about the benefits we bring to communities," Amazon's Hurst said, adding that it benefits communities in ways beyond direct job creation, such as spending with local businesses and funding job-training efforts. A Meta spokesperson said it helps communities where it operates through grants and partnerships.

The Data Center Coalition spokesperson said that focusing on jobs inside data centers understates the impact on service providers and suppliers, such as electricians, HVAC manufacturers, and portable sanitation companies.

Companies are still required to make yearly payments to the cities in lieu of property taxes to help ensure minimum contributions to the communities, which Business Insider incorporated into our cost-per-job calculations. Meta, for example, paid $21.8 million in total to New Albany as of December 2022. A spokesperson for New Albany said the payments ensure "data centers contribute meaningfully to the community, even with tax abatements in place."

And tech companies often sweeten the deals by promising to invest in education programs to upskill local workers. Amazon, for example, donated $25,000 and some equipment two years ago to the Tolles Career & Technical Center in Plain City, Ohio, to support the school's IT and cybersecurity training programs, which include a four-week training program for entry-level data center workers. At the nearby Columbus State Community College, the company pledged $50,000 in scholarships for a new data center technician certificate program.

'Unprecedented' electricity use

The ultrapowerful computer chips crammed into data centers consume enormous amounts of power. A 2024 Department of Energy report estimates their electricity use, driven by the AI boom, could soon command as much as 12% of total US electricity use, from just over 4% in 2023.

Data centers are getting breaks on that, too — which residents and other businesses are helping pay for.

From 2020 through last year, Ohio data centers' load on the grid rose sixfold. By 2030, American Electric Power Ohio, the state's largest electricity provider, expects to grow by another 700% to reach 5,000 megawatts, enough to power at least 2 million homes.

If all hookup requests across more than 90 planned data center sites in Ohio are approved, AEP Ohio told regulators, demand could skyrocket to over 30,000 megawatts.

Since 2017, Ohio regulators have authorized multiple 10-year electricity rate subsidies for data center developers, reducing power costs for tech companies in exchange for their promises of new jobs. Other AEP customers have to pay for the shortfall.

Matt Schilling, a spokesperson for the Public Utilities Commission of Ohio, said in an email to Business Insider that while the commission had approved some discounted rates for data centers, it had denied other applications for such arrangements.

At the same time, AEP has proposed spending at least $850 million in new or upgraded grid infrastructure and power plants to serve data centers, and another $350 million in other upgrades to support Central Ohio's extreme demand growth, according to filings. Ratepayers across Ohio foot the bill for this too, as AEP spreads the costs across all customers.

Walmart, one of Ohio's largest employers, said last June that an increasingly expensive electricity bill — owing partly to data centers' demand — imperiled its continued expansion in the state. That warning came in a filing supporting the utility's recent proposition to increase tariffs and regulations on data center customers.

A Data Center Coalition representative warned regulators in 2024 that those proposed tariffs and restrictions in Ohio could "depress the growth of an important emerging industry." The rate case remains ongoing.

Regulators across the US have offered similar deals to subsidize data centers' electricity use, shifting billions of dollars of costs to all ratepayers, including residential customers.

Regulators last year OK'd Georgia Power to construct an estimated $300 million 35-mile high-voltage transmission line and a new substation for a QTS data center near Atlanta. And this year, South Carolina regulators authorized Duke Energy to invest $66.5 million to upgrade a transmission line to serve a new QTS data center. The utilities will recoup their investments by increasing electricity bills for all their customers.

Duke Energy said it follows federal rules in allocating upgrade costs. South Carolina's regulator declined to comment and Georgia Power and that state's regulator didn't respond.

A QTS spokesperson said it pays for all utility infrastructure dedicated to its data centers "to ensure no impact to residential rates."

"Utilities can fund discounts to Big Tech by socializing their costs through electricity prices charged to the public," a 2025 Harvard Law study of regulatory proceedings about utility rates for data centers found. Utilities profit, the study said, by "forcing the public to pay for infrastructure designed to supply a handful of exceedingly wealthy corporations."

Amazon, Microsoft, and Google told Business Insider they were committed to paying their full share for infrastructure serving their power needs. Tech companies and industry advocates say that other factors, such as electric vehicles, also are driving electricity growth and that the transition to renewable power drives up electricity costs.

To estimate the amount of power data centers demand nationwide, Business Insider used data from the air permits issued to data center backup generators. (See here for more on Business Insider's methodology.)

If every data center that's been issued a permit comes online, Business Insider estimates data centers' total electricity use across the country could reach between 149.6 terawatt-hours and 239.3 terawatt-hours a year. Business Insider's low-end estimate is roughly equivalent to the state of Ohio's electricity needs in 2023, and on the high end, is nearly as much power as the entire state of Florida used that same year. A 2024 federal report estimated US data centers' electricity use could reach the high end of Business Insider's estimate by 2026.

A 2024 report to Virginia's legislature found that data centers had historically paid their fair share of transmission upgrade costs but warned their sharply escalating electricity needs "will likely increase system costs for all customers, including non-data center customers."

Last July, Dominion Energy, Virginia's largest utility provider, asked regulators to approve a $23 million grid infrastructure investment billed across ratepayers, a request that is still pending. Regulatory staff said the investment was likely needed just for a single data center customer.

Months later, Dominion disclosed that it would need to roughly double its electricity generation by 2039 primarily to meet meteoric data center demand and new planned renewable energy capacity. Dominion estimates the planned expansion could cost up to $103 billion, increasing residential electricity bills by as much as 50%.

Aaron Ruby, a Dominion spokesperson, told Business Insider that the company had asked regulators to approve additional consumer protections to shield ratepayers from shouldering costs incurred by large customers like data centers. The planned increase in power bills is primarily driven by the utility's transition to carbon-free power generation, as is required by state law, Ruby wrote.

In Virginia, too, Walmart objected.

"Electricity is a significant operating cost for retailers such as Walmart," Lisa Perry, Walmart's director of utility partnerships, told regulators in February 2025, warning that increasing electricity rates would harm Walmart's investment in Virginia.

Andy Farmer, a spokesperson for the Virginia State Corporation Commission, said that data centers affected all the state's utilities, not just Dominion.

Data centers' ballooning power consumption leaves other businesses, residents, and utility regulators in a bind: Either pay to expand capacity for the tech companies, or risk going without enough power to attract other new business.

In Indiana, the River Ridge Property Owners' Association in Clark County told state regulators in 2024 that a single Meta data center project had bled nearly all remaining power from the grid. Meta promised at least 50 high-paying permanent jobs at the site and hundreds of construction jobs, but the community would have no available electricity to attract other prospective companies investing in the area for at least four years.

"It is possible these data centers ultimately restrict, rather than foster, additional economic development," a representative of the Citizens Action Coalition of Indiana, a consumer and environmental advocacy organization, told state regulators. By 2030, the representative said, "just a few" data centers used for applications like AI will use "more electricity than all 6.8 million Hoosiers use at their homes."

Walmart representatives told Ohio regulators last year that data centers' massive electricity use threatened the company's planned rollout of electric vehicle charging locations at its retail locations.

"Growth in data center development is an economic boon for Ohioans," Google representatives told regulators this year, adding that the facilities were "pivotal in establishing the state as a leading technology hub."

Walmart argues that it brings more jobs and other benefits to the local economy — a claim supported by research from AEP Ohio. The utility calculated that each megawatt allocated to traditional commercial and industrial customers like Walmart supported at least 25 jobs.

Every megawatt used by a data center, the utility said, supports less than one job.