Missouri's Data Center Gold Rush

Missouri is in the middle of a data center land rush. Billions of dollars in proposed investment from Meta, Nebius, Amazon, Google, and others. One of those projects—code-named “Project Tiger Zou”—is targeting Moberly, the town where I live. I spent several weeks pulling apart what data centers actually deliver to the communities that host them, and the picture is more complicated than the pitch decks suggest. The short version: the capital investment numbers are real, the permanent job numbers are thin, the infrastructure demands are significant, and the towns that do best are the ones that negotiate hard before the ground breaks. I’m not anti-data center. I’m anti-bad deal.

How I Got Here

I live in Moberly, Missouri. Randolph County. About 12,000 people, a railroad town, 45 minutes from everywhere and close to nothing in particular. It’s the kind of place that gets excited when any investment shows up, because investment doesn’t show up often.

So when I started hearing the phrase “data center” in local conversations, and then the code name “Project Tiger Zou” attached to Moberly specifically, I did what I always do when something shows up that I don’t understand. I started reading.

I should be upfront: I knew almost nothing about data centers several weeks ago. I knew they were big buildings full of servers. I knew tech companies were building a lot of them. That was roughly it. What I didn’t know—and what I’ve been trying to figure out since—is how the economics actually work, what the infrastructure demands look like, what the incentive structures are, or what happens to the communities that host them after the ribbon-cutting photos go away.

I’ve gone through a lot of sources at this point: news articles, government reports, state laws, county meeting minutes, cost-benefit studies, research papers, and a lot of local journalism from reporters who’ve been covering this harder than anyone else. What follows is me trying to make sense of all of it.

This is the first thing I’ve written for this site, and it’s also happening right now, in my town. That feels like the right place to start.

The Gold Rush

Missouri is in a data center arms race, and it’s not subtle. In the last two years, the state has attracted or is actively courting tens of billions in proposed data center investment. Here’s what I’ve been able to piece together from public reporting:

So, a note on this table. These numbers come from public reporting: news articles, county commission documents, press releases. They shift. The Nebius project in Independence has been reported at figures ranging from $6.6 billion to over $100 billion depending on the time horizon and what’s being counted. I’m using the more conservative near-term numbers where I can. The point isn’t precision down to the dollar. The point is scale.

Missouri currently has around 48 operating data centers, with another 36 planned or under construction. That’s not a trend. That’s a stampede.

The Incentive Playbook

To understand why these projects keep landing in Missouri, you have to look at the incentive structure, because it’s generous.

Missouri’s Data Center Sales Tax Exemption Program means qualifying data centers don’t pay sales tax on their equipment, electricity, or construction materials. To qualify, you need to invest at least $25 million for a new facility (or $5 million for an expansion), create at least ten jobs, and those jobs have to pay at least 150% of the county average wage. The tax break lasts up to 15 years.

On top of that, cities and counties can offer something called Chapter 100 bonds. Here’s how that works in plain English: the city issues bonds to pay for the project, technically owns the property, and leases it back to the company. Because the city holds the title, no property taxes. The company pays a negotiated amount instead, called “payments in lieu of taxes” or PILOTs, and those payments are usually way less than actual property taxes would be. In the Amazon deal in Montgomery County, the property tax break could be anywhere from $244 million to $982 million. That’s tax revenue the county just… doesn’t collect.

Then there’s SB 4, which Governor Kehoe signed in April 2025. This one’s about electricity. The short version: it’s supposed to make sure regular homeowners don’t get stuck paying for the massive power grid upgrades that data centers need. Whether it actually works that way in practice—I’m still trying to figure that part out.

There’s also a pair of House bills worth watching: HB 3362 and HB 3364, championed by State Representative Colin Wellencamp. Together they’re called the “AI Infrastructure, Grid Integrity and Water Resource Protection Act.” The idea is to create a completely separate utility rate class for large-load consumers—which is legislative language for data centers. If it works the way it’s written, it would build a financial firewall between the tech companies and everyday ratepayers. Whether it survives the lobbying process intact is another question.

Missouri is competing with Nebraska, Ohio, Virginia, and a bunch of other states running the same playbook. The state’s logic makes sense on paper: attract big investment, create construction jobs, get some permanent jobs, and build out infrastructure that helps everyone. The question is whether that’s actually what happens when you look at what these communities end up with.

The Math That Doesn’t Add Up

Here’s the thing I keep coming back to: the capital investment numbers are enormous, and the permanent job numbers are tiny.

Meta’s $1 billion Kansas City data center had roughly 1,500 skilled trade workers on site at peak construction. When the facility went operational in August 2025, it supports about 100 permanent jobs. A billion dollars of investment for 100 ongoing positions.

Amazon’s proposed $8.5 billion complex in Montgomery County projects at least 150 permanent jobs. Nebius in Independence hasn’t disclosed permanent job numbers, but industry ratios suggest it won’t be dramatically different.

The pattern holds everywhere I look. Ohio is the clearest example because they have the most data. As of 2024, Ohio’s data center industry had created approximately 22,300 short-term construction jobs but only about 4,500 permanent operations positions—and that’s across an estimated $40 billion in total statewide investment. The Ohio River Valley Institute straight up called data centers “economic development duds.” I think that’s a little harsh, but the numbers behind it are hard to argue with.

The construction jobs are real, well-paying, and valuable. Construction workers in the data center boom are earning up to 30% more than standard rates, with some pulling six-figure salaries. But they’re temporary. When the building is done, those workers move to the next site. The community is left with the facility and its small permanent staff.

I’ve seen this called the “construction treadmill”—a term that comes out of Prineville, Oregon, where Facebook (now Meta) built their first data center in 2011. Prineville saw a housing cost spike of 45% in five years as construction workers flooded the local market, then experienced the predictable correction when they left. The permanent economic impact was real but modest: franchise fees, some property tax revenue, community grants. Not nothing. But not what the initial investment number suggests.

The jobs argument is weak. I don’t think that means data centers are bad—there’s a real case for infrastructure investment, tax revenue, grid improvements. But the jobs piece specifically? Much thinner than the press releases make it sound. And I think communities should know that before they sign anything.

Moberly’s Hand

So why Moberly? I don’t have the full picture on Project Tiger Zou—the details are still behind NDAs, which is itself a problem I’ll get to—but what I do know is this: the proposal targets over 1,000 acres in Randolph County. To put that in perspective, the entire Moberly Area Industrial Park—the one that’s housed local businesses for decades—is about 206 acres. Tiger Zou wants nearly five times that. It’s not a building. It’s a campus the size of a small town.

I can see why a data center developer would look at Randolph County and get interested.

Power. The Thomas Hill Energy Center sits in Randolph County, and there’s a major high-voltage power line (345kV transmission corridor) running through the area. That means the heavy-duty electrical infrastructure already exists, and nobody has to build it from scratch. For a data center that might need as much electricity as a small city, being close to that kind of power capacity is one of the first things they look for.

Water. Moberly gets most of its water from Sugar Creek Lake, about four miles northwest of town. The city currently has about 3.9 million gallons per day more than it needs. That sounds like plenty until you learn that big data centers using evaporative cooling (basically, they spray water to keep things cool) can burn through over a million gallons a day. One big facility could eat up a huge chunk of that extra water.

And here’s the part that worries me. Sugar Creek Lake is shrinking. The USGS did depth surveys in 2003 and 2018 and found the lake lost 230 acre-feet of storage. That’s basically dirt and sediment filling in the bottom, a little more every year. The lake was built in 1922. It’s over a hundred years old. The water Moberly has to spare today is not the water Moberly will have to spare in twenty years. Any deal that treats today’s numbers like they’re permanent is building on a shrinking foundation.

Fiber. Bluebird Network runs a big regional fiber internet network through the area, the fast data connection data centers need. And the Moberly Area Industrial Park was set up with large water mains and utility connections specifically to attract this kind of project.

Land. This one cuts both ways. Data from similar projects in Nebraska and Georgia shows that when hyperscale developers come shopping, they pay seven to ten times the standard market value per acre for agricultural land. If you’re the farmer who gets that offer, it’s a lottery ticket. But it warps the local land economy. Suddenly every acre around the project is priced for tech money, not farming. Local farmers who want to stay and expand can’t compete with a developer’s checkbook. The land gets more expensive for everyone, even people who have no intention of selling.

And the shock doesn’t stop at farmland. The Randolph County General Plan already identifies a shortage of standard workforce housing—the three-bedroom, $200,000 homes that working families actually need. Drop 100 high-income tech professionals into a housing market that tight and you get a price spike that pushes out the people who’ve lived here for decades. Moberly’s housing stock is thinner than most of the towns where this has already happened.

Moberly has real assets. But they’re finite assets, and the town needs to understand the difference between having something to offer and having something to give away.

What the Neighbors Are Saying

If I only listened to the developers and the economic development pitches, this would be a straightforward story. Big investment, some jobs, tax revenue, everybody wins. But the communities that have been through this process, or are going through it right now, are telling a different story, and I think it’s worth listening to.

Independence is the most active fight in Missouri right now. Nebius, a Dutch AI company that severed ties with Russia-based Yandex in July 2024, is proposing a hyperscale data center on nearly 400 acres off Little Blue Parkway. Residents organized quickly: door-knocking, calling council members, showing up at polling locations during the February primary. They raised concerns about water usage, power demand, proximity to schools and homes, and wildlife impacts. And there’s a detail from Independence that should make anyone in Moberly pay attention: to feed Nebius’s planned 800-megawatt campus, the city is being forced to reopen and expand the shuttered Blue Valley natural gas power plant with 15 massive turbines—just for Phase 1. An 800-megawatt demand is roughly five times what every home, business, and municipal facility in the entire city of Independence uses combined. One data center, and they have to resurrect a dead fossil fuel plant to plug it in.

The city council approved a 90% tax abatement anyway, voting 5-2 on March 3. Opponents are now pursuing a referendum petition, needing 3,700 signatures in 30 days.

St. Charles County saw residents push back against a proposed Google-linked data center after learning that city officials had signed NDAs with the developer, preventing them from disclosing who was behind the project. The city council passed a one-year moratorium on data center applications in August 2025 and is now considering making the ban permanent. Residents’ primary concerns were water contamination and the facility consuming more water and electricity daily than the GM plant in Wentzville.

Montgomery County approved a tax incentive framework for Amazon’s $8.5 billion project near New Florence, but not without pushback. Opponents have filed suit arguing the deal should be voided over transparency issues.

Festus is watching its own version of the same drama play out. KSDK’s reporting revealed that city leaders had been meeting privately with developer CRG for months before the project became public. A grassroots group called Wake Up Jeffco has filed for an LLC to pursue litigation over transparency failures. And KSDK separately exposed an astroturfing campaign, a group called “Missourians for Data and Technology Advancement” running newspaper ads and planting yard signs that looked grassroots but traced back to a Delaware-registered shell entity with no disclosed officers, staff, or donors.

DeForest, Wisconsin, outside Missouri but worth noting,rejected a $12 billion QTS data center proposal in January 2026 after the Wisconsin Department of Administration determined the annexation wasn’t in the public interest, citing concerns about the village’s ability to provide water and sewer services. Residents had organized against the project after public records requests uncovered emails between QTS executives and the village president.

And here’s the thing that keeps bugging me about how these fights get covered: these aren’t anti-tech people. They’re not NIMBYs trying to block progress. They’re people asking straightforward questions: how much water will this use, how much power, will my rates go up, how loud will it be, and why won’t anyone tell us what’s going on? The fact that those questions keep getting treated like problems instead of reasonable asks tells you something about how these deals get done.

What Moberly Needs to Do

I don’t know yet what Project Tiger Zou looks like in detail. But I know enough about the pattern now to know what Moberly should demand before any deal gets a vote. And I want to walk through my reasoning on each one, because I think the “why” matters as much as the “what.”

1. Require cooling systems that don’t drain our lake. The cheap way to cool a data center is to spray water and let it evaporate, which can use over a million gallons a day. Moberly’s water surplus sounds comfortable until a dry year hits and Sugar Creek Lake can’t keep up. There are better cooling systems that recycle the water instead of burning through it. Bluebird Network already uses one at their underground data center in Springfield. The tech exists. It costs more, yes. That’s kind of the point. The developer is making the investment, and the community is providing the water. The water is not unlimited.

2. Make the developer pay for the power upgrades. The state already passed a law (SB 4) that’s supposed to keep regular homeowners from getting stuck with the bill when data centers need bigger power lines and substations. Moberly should lock that in locally too. No deals where the electric company borrows money for upgrades and then passes the cost to everyone’s power bill if the data center doesn’t work out. If the developer needs the power, the developer pays for the power. Simple.

3. If they don’t deliver, they pay us back. Every tax break should come with a “give it back” clause. If the company doesn’t hit their job numbers, investment promises, or wage commitments within 36 months, they repay the tax breaks. Not renegotiate. Not push the timeline. Pay it back. The Montgomery County Amazon deal is projecting $400 million to $1.8 billion in tax revenue, but that depends on Amazon actually building all 17 buildings they’ve talked about. If they build four and stop, the community needs a way to get made whole.

4. No more secret deals. This is the one I feel strongest about. Across Missouri, the same thing keeps happening: developers make city officials sign NDAs, everything gets negotiated behind closed doors, and by the time regular people find out about it, the deal is basically done. St. Charles. Festus. Montgomery County. Same story every time. The secrecy isn’t protecting business secrets. It’s keeping the public from asking questions. Moberly should require that the basics (how much water, how much power, how much land, what kind of cooling) get shared publicly before anything gets voted on. If a developer won’t tell the community what they’re asking for, the community shouldn’t give it to them.

5. Train local people first. Data centers need technicians: networking, electrical, HVAC, security. Without a local talent pipeline, those jobs go to engineers commuting in from Columbia or St. Louis. They do their shift, take their paycheck, and spend it somewhere else. Moberly gets the giant concrete building but none of the actual economic benefit.

This is exactly the problem that RAISE Randolph County was designed to solve. Launched in early 2025, it’s a targeted workforce development initiative that aligns Moberly Area Community College’s curriculum directly with the technical requirements of incoming tech firms. Registered apprenticeships, direct training pipelines, the kind of thing where the local workforce is trained on the exact equipment the company uses before the facility even opens. The idea is sound. But it needs real money and real programs running before construction starts, not after. I’ve seen too many promises about “future workforce investment” that never turn into anything. That’s a press release, not a plan. Put the training commitments in the contract with dates and dollars.

6. Set noise and distance rules now, before anyone asks. Data centers run big diesel backup generators that get tested regularly. Their cooling systems produce a constant low hum. How loud? Environmental studies put it at up to 90 decibels at the source—roughly equivalent to a lawnmower running 24 hours a day or a 737 a mile out from landing. That’s loud enough to be harmful over extended exposure, and it doesn’t stop when you’d like to go to sleep.

And there’s an air quality problem most people don’t associate with “clean tech.” Those diesel backup generators have to be tested regularly to make sure they’ll actually work in an emergency. When they fire up, they can emit 200 to 600 times more nitrogen oxides than a traditional regulated power plant. Long-term NOx exposure is linked to respiratory issues and asthma. The irony of a “cloud computing” facility degrading the air quality in a rural community is not lost on me.

Write the rules into zoning before a specific project shows up: 500-foot minimum from homes, noise limits (60 dB daytime, 50 dB at night (think dishwasher-level, not highway-level)), sound barriers and landscaping berms required, and generator testing restricted to business hours only. None of this is unusual. These are normal industrial zoning rules. The time to set them is now, not in the middle of negotiating a specific deal.

7. Require battery backup instead of diesel generators. This is a straightforward fix for the air quality problem. Battery energy storage systems provide the same backup redundancy that diesel generators do—the servers stay on if the grid drops—but without the massive spikes in nitrogen oxide emissions. If the choice is between a developer saving money on backup power and the surrounding community breathing clean air, I know which side the zoning code should land on.

8. Make the waste heat work for the community. This is the one idea from my research that genuinely excited me. Data center servers run incredibly hot—operational temperatures can range from 77 to 194 degrees Fahrenheit. Most facilities just vent that heat into the atmosphere through massive cooling fans. It’s a staggering waste of energy. But some communities are starting to capture that exhaust heat and pipe it to adjacent agricultural greenhouses or food processing facilities. Given Randolph County’s farming roots, Moberly could pioneer something genuinely interesting here: the data center becomes a free heating engine for year-round indoor farming. It’s the kind of thing that turns a liability into an asset, and it would tie the tech investment back to the agricultural identity the county already has. I’d love to see this written into the deal.

9. Require a conditional use permit, not by-right zoning. This is about who holds the power. Historically, a lot of data centers in Missouri could be built “by right” if the land was zoned for basic industrial use—meaning the city had very little legal ability to attach conditions or demand changes. Cities like Kansas City and Warrensburg have already fixed this by amending their municipal codes to require a conditional use permit (CUP) for any data center development. A CUP means the developer doesn’t have an automatic right to build just because they bought the land. They have to come before the city council, and the city can legally attach specific, binding conditions: closed-loop cooling, noise limits at the property line, landscaping buffers, you name it. The CUP is the lever that makes everything else on this list enforceable. Without it, these are suggestions. With it, they’re requirements.

I’m not saying don’t do the deal. I’m saying do the deal right. The leverage Moberly has is before the agreement is signed. After that, the leverage belongs to whoever is writing the checks.

What I Think (So Far)

I want to be honest about where I am on this, because I don’t think I’ve fully made up my mind.

I’m not anti-data center. The infrastructure investment is real. The construction employment is real, even if it’s temporary. The grid upgrades, if structured correctly, can benefit the broader community. The tax revenue, even at abated rates, is more than zero. And Moberly is a town that could use economic activity that isn’t dependent on the same three or four employers it’s relied on for decades.

But I’m not going to pretend the numbers say something they don’t. A billion-dollar data center that creates 100 permanent jobs isn’t a jobs program. It’s a building project that happens to need some people to run it. There’s nothing wrong with that, but the community needs to see it for what it is, not what the press release says it is.

What I keep coming back to is pattern recognition. Moberly has been courted by big industry before. The town knows what it’s like to have a major employer arrive, change the local economy, and then either scale back or leave.

And there’s a longer-term risk that doesn’t get enough attention: obsolescence. The concrete shells of a data center are built to last 50 to 100 years. But the servers inside them have a life cycle of three to five years before they’re outdated. If AI demand shifts, or if the next generation of computing requires a fundamentally different kind of facility, what’s left? Researchers who studied Prineville, Oregon and West Des Moines, Iowa—towns that welcomed data centers with open arms—found essentially zero secondary economic development 15 years after the initial facilities were built. The data centers didn’t attract other businesses. They didn’t create bustling tech parks. They just existed as massive, silent, isolated islands of concrete. That’s the existential risk of putting all your economic eggs in one basket, and it’s why every deal should include a funded decommissioning plan before the first bulldozer touches dirt.

The question isn’t whether this investment is good or bad in the abstract. The question is whether the city has learned to negotiate from a position of clarity rather than excitement.

I don’t know the answer to that yet. I don’t know what Project Tiger Zou’s specific demands are. I don’t know what Moberly’s negotiating team looks like or what legal counsel they’ve retained. I don’t know whether the community will get a meaningful voice before the deal is structured or after.

What I do know is that the communities in Missouri that have fared worst in this process are the ones that let enthusiasm outrun due diligence. And the ones that have done best—or at least preserved their options—are the ones that asked hard questions early and didn’t accept “trust us” as an answer.

Where This Goes Next

This is the first thing I’ve published on this site, and I picked it because it’s not theoretical. It’s happening in my county, on a timeline I can watch unfold. I’ll keep updating as more information comes out. About Project Tiger Zou specifically, about Missouri’s legislative response, about what the communities that said yes are actually experiencing a year from now.

I should also say: I might change my mind about some of this. If I see data that contradicts what I’ve laid out here, or if the project details turn out to be materially different from the pattern I’ve described, I’ll say so. That’s kind of the point of writing in public. The thinking is visible, which means the corrections have to be visible too.

If you’re in Moberly or Randolph County and you have information I should be looking at, I want to hear it. If you think I’ve gotten something wrong, I especially want to hear that. I’d rather be corrected early than confident and wrong.