AI’s Cell Tower Moment

We want the AI answer, but not always the data center that makes it possible.

By Mustafa Tameez
JUNE 23, 2026

In the 1990s, people wanted the cell phone, but not the cell tower.

Today, we want the AI answer, but not the data center.

The tension feels familiar to me.

When I started out in advertising sales, work changed. For the first time, work was not just something that happened behind a desk. It was on the go. It was in the car. It was between meetings. It was on the phone.

From a contact standpoint, that mattered. If you were trying to reach clients, close business, respond quickly, or keep a relationship moving, being reachable was part of the job.

First, I had a pager. Then I added a Motorola cell phone and carried both.

The cell phone cost about 25 cents a minute, so you did not casually hand out your cell number the way we do today. The system was simple: if you needed me, you paged me. If I was not near a landline, I called you back from my cell phone.

Then came the fat extended battery. The phone was already big, but the bigger battery made it feel like you were carrying a small brick. It was bulky. It was expensive. It was inconvenient.

And it still felt like the future.

After I moved to Texas, I was one of the first people I knew to get a cell phone from a company called Aerial Communications. They had an offer that felt unbelievable at the time: something like 1,000 minutes for $50. Compared with paying by the minute, that was a big, big savings.

But there was one problem.

The phone did not work in a lot of places.

So for a while, I carried two cell phones: my Aerial phone and my Cellular One phone. Aerial was cheaper, but the coverage was spotty at best. Cellular One had much better coverage, but it was expensive.

I learned about cell towers the hard way.

The difference was not just the phone in my hand. The difference was the system behind it. One company had coverage in the places I needed to be. The other had a better price, but could not always connect me.

There was the tradeoff.

And it became one of the controversies of the time.

If you are old enough, you probably remember Verizon’s famous “Can you hear me now?” commercials. The ads worked because they captured a real frustration. People did not just want a cell phone. They wanted the call to go through.

Verizon understood this perfectly.

The reason you could hear the person on the other end was not just the handset. It was the network behind the handset. Verizon was selling coverage, reliability, field testing, and the tower infrastructure that made the phone useful.

The commercial was funny and memorable, but it was really selling something most people never thought about until their phone stopped working.

It was selling the system.

I remember watching those Verizon commercials with envy. At the time, I was using the lower-cost carrier — the company that later became part of the T-Mobile story — and the savings were real. But the coverage was not the same. Verizon could ask “Can you hear me now?” because it had more of what mattered in more places.

The question was not really about the phone.

It was about everything behind the phone.

People wanted cell phones. They wanted pagers. They wanted good coverage. They wanted phones that worked everywhere. But they did not want a cell tower near their home, their office, their child’s school, their church, or anywhere they could see it.

People worried the towers were ugly. They worried about property values. They worried about health. Some feared the towers could cause cancer. Others simply did not want a large metal structure changing the look and feel of their neighborhood.

Looking back, I do not think people were wrong to ask questions. Communities should ask questions when new infrastructure shows up.

But the tension was clear.

Everyone wanted the benefit of the technology. Very few people wanted to host the physical system that made the technology work.

The data center debate brings me back to that same tension.

Most of us experience AI as something on a screen. I type a question. It answers. It helps write, summarize, search, edit, design, translate, listen, and now even coach in real time.

It feels weightless because that is how we encounter it.

But somewhere, a machine is doing the work.

Behind every AI answer is a physical system: land, electricity, water, fiber, steel, concrete, cooling equipment, and people. The answer on your screen depends on a building full of servers somewhere else.

We call that building a data center.

And right now, data centers are where the promise of AI is running into the reality of communities.

The controversy is not hard to understand.

When a data center lands near your home, your power grid, your water supply, your roads, your school district, and your tax base, people start asking questions.

They should.

A recent Gallup survey found that roughly seven in ten Americans oppose building an AI data center in their local area.

Seven in ten should get everyone’s attention.

People worry about electricity demand, water use, noise, construction traffic, land use, utility bills, and whether the economic benefits really stay in the community.

I understand those concerns. If a major facility were being built near my neighborhood, I would want to know the same things: who pays, who benefits, and what happens if the promises do not hold up.

I do not read that as people rejecting technology.

I read it as people realizing, sometimes for the first time, that the technology they like has to live somewhere.

Full disclosure: my firm has worked with Facebook, including work connected to the launch of its digital wallet. So I have seen how technology companies think about adoption, scale, access, and trust. My experience with Facebook does not make me dismissive of community concerns. It makes me more convinced that companies have to answer them directly.

The cell tower era taught me something basic: infrastructure has to earn trust.

Cell phones became normal because the value became obvious, the technology improved, the coverage expanded, the devices got better, and people adjusted. But adjustment did not happen by magic. It happened because the infrastructure became part of everyday life.

Data centers are entering that same stage now.

AI is moving from something we use on a screen to something that requires power plants, transmission lines, water systems, cooling equipment, construction workers, electricians, land-use approvals, and local political permission.

Meta’s recent announcement fits directly into this.

Meta, Facebook’s parent company, announced America’s Workforce Academy, a cost-free skilled-trades training program tied to the data-center buildout. The company says it is making an initial $115 million investment and launching the program in Louisiana, Ohio, Indiana, and Texas — with Houston among the first pilot cities. Meta says graduates will receive industry-recognized credentials and a guaranteed job.

For workers, the value is obvious.

For companies, there is another lesson.

For companies building this new AI infrastructure, this is not just a communications issue. It is a permission-to-build issue.

Data centers require power, water, land, workers, permits, and public confidence. A company may have the capital, the technology, the engineers, and the land option. But if the community does not trust the bargain, the project can still run into delays, higher costs, political backlash, and a longer path to approval.

Workforce training is more than a nice announcement.

Done right, it is part of the public bargain.

It says to a community: this project is not only going to use your power, your water, your roads, and your local infrastructure. It is also going to train your people, hire your workers, and create a pathway into the economy being built around AI.

Most of the AI jobs conversation has focused on what might disappear.

Writers. Coders. Analysts. Call-center workers. Young professionals just starting out.

But AI also creates demand for work that looks very different: electricians, HVAC workers, fiber crews, construction managers, utility planners, and power engineers.

AI needs people in hard hats as much as people writing code.

It does not erase the controversy. It makes the bargain more honest.

A data center can bring real construction jobs. It can create permanent technical jobs. It can expand the tax base. But communities deserve to know which jobs are temporary, which are permanent, whether local workers will actually get them, and whether the training pipeline is real.

A construction boom is not the same thing as a long-term employment base.

A tax incentive is not the same thing as shared prosperity.

A ribbon cutting is not the same thing as public trust.

The companies that win this next phase will not simply be the ones that build the fastest. They will be the ones that understand public concerns early, answer them honestly, and do the harder work of building trust before opposition hardens.

The first concern I would want answered is power.

AI uses enormous amounts of electricity because the computing work is enormous. The Department of Energy has estimated that data centers consumed about 4.4 percent of total U.S. electricity in 2023 and could consume between 6.7 percent and 12 percent by 2028.

Those numbers matter because electricity is not imaginary.

Someone has to generate it. Someone has to transmit it. Someone has to pay for new grid capacity. And someone has to decide whether that power comes from natural gas, renewables, nuclear, batteries, or some combination of all of them.

A homeowner hears that a data center is coming and wonders whether the electric bill will go up, whether the grid will become less reliable, whether a company is getting a special deal, or whether a new gas plant will be built nearby to serve the project.

These are kitchen-table questions, not abstract ones.

The second concern is water.

Servers produce heat. Almost all the electricity that goes into computing eventually becomes heat. If that heat is not removed, equipment slows down, fails, or shuts off.

Data centers need cooling because they are trying to keep expensive machines alive.

Cooling creates tradeoffs. Air cooling can use more electricity. Evaporative cooling can use more water. Liquid cooling can be more efficient for advanced AI chips, but it requires different design, higher capital costs, and more technical complexity.

The better questions are practical ones.

How much water will this facility use? From what source? Will it use drinking water or reclaimed water? What happens during a drought? What happens when several large users arrive in the same region at once?

Engineers can solve many of these problems.

But the public should not have to find out the plan after the deal is done.

The issue I think companies underestimate most is speed.

AI is moving faster than local government.

A company can plan a multibillion-dollar data-center project. A utility can negotiate new power arrangements. A state can offer incentives. A city can update zoning. A county can discuss infrastructure.

Residents often find out late, after the most important decisions already feel settled.

Opposition hardens in that gap.

People may like AI. They may use it every day. They may want the jobs. They may want America to lead China. They may want Texas to win the next wave of economic development.

But they do not want to be treated like an afterthought.

Companies and public officials need to slow down long enough to be honest.

The old model was simple: announce the project, praise the investment, promise jobs, minimize concerns, and move forward.

The model is breaking.

The better approach is pretty basic: tell people how much power it will take, where the water comes from, what the real job numbers are, what incentives are being offered, and who pays if the grid needs upgrading.

I do not see that as anti-growth. It is how you keep growth from turning into backlash.

The other side of the story is also true: data centers are becoming one of the major engines of the American economy.

McKinsey has estimated that global data-center demand could require $6.7 trillion in capital spending by 2030, with most of that tied to AI workloads. The International Energy Agency projects global data-center electricity consumption could roughly double by 2030.

This is not a niche real-estate sector anymore.

It is a new industrial stack: chips, land, power, water, fiber, cooling, construction, cybersecurity, workforce training, local permits, state incentives, and national security.

Every mayor, county judge, governor, utility board, workforce board, and economic development group needs to understand it.

Data centers are not just places where technology companies store information.

They are where the AI economy becomes physical.

And when an economy becomes physical, politics follows.

There is one more part of this that frustrates me.

As Americans, when we are struggling through complicated issues like this, we do not always get to work through them on our own.

Our open society is one of our greatest strengths. We argue in public. We organize in public. We criticize powerful companies in public. We question public officials in public. A free country is supposed to work that way.

But our adversaries understand that openness, too.

They know that if Americans are already anxious about something important, one way to weaken us is not to invent the anxiety from scratch. It is to amplify it, sharpen it, and turn our own public debates against us.

I read the reports about foreign influence around data centers through that lens.

OpenAI reported that it disrupted accounts linked to a PRC-origin influence operation that generated social media comments and images claiming AI data-center buildouts were increasing electricity prices for average families. OpenAI called that cluster the “Data Center Bandwagon” campaign.

That should concern us.

AI infrastructure is now part of national power. If foreign actors can make Americans distrust the systems we need to compete, they will try.

But we should be careful here.

The existence of foreign amplification does not make local concerns illegitimate. It means we have to protect two things at once: the honesty of our public debate and the right of communities to ask hard questions before major infrastructure is built around them.

That is not easy.

But it is the work of an open society.

Some people look at all of this and ask: why not move the data centers somewhere else?

Into the ocean. Into space.

It sounds futuristic, but it is not science fiction.

Microsoft tested underwater data centers through Project Natick, placing servers in a sealed module off the coast of Scotland. Space-based data centers are also being discussed because they could use solar power and avoid some land-use fights.

But moving servers does not move the physics away.

In the ocean, you still need cables, maintenance, permits, energy, and a plan for what happens when something breaks.

In space, servers still produce heat. Cooling can be harder because there is no air to carry heat away. You also have radiation, launch costs, debris, insurance, and repair.

These ideas are fascinating. They may become part of the future. But they do not change the core lesson.

AI may feel like magic on the front end.

On the back end, it is a machine that needs power, cooling, land, workers, and public permission.

The data-center controversy is not a reason to stop building. It is a reason to build better.

America will need data centers. Texas will need data centers. The AI economy will need power, cooling, land, and skilled workers. Communities will need jobs, tax base, infrastructure, and opportunity.

But communities also need confidence.

A good data-center project should strengthen the grid, protect local ratepayers, use water responsibly, train local workers, create real opportunity, and tell the truth before the deal is done.

Maybe the simplest way to say it is this: the answer on the screen depends on a system most of us never see.

In the 1990s, I learned that with cell phones.

Today, we are learning it again with AI.

The question is not whether the system will be built.

The question is whether communities will trust the way it is built.

Frequently Asked Questions

Sources and Further Reading

1. Gallup — Americans Oppose AI Data Centers in Their Area: Useful public opinion context on local opposition to AI data centers. https://news.gallup.com/poll/709772/americans-oppose-data-centers-area.aspx
2. S. Department of Energy — Electricity Demand Growth Resource Hub: Context on U.S. data center electricity demand and projected growth. https://www.energy.gov/policy/electricity-demand-growth-resource-hub
3. International Energy Agency — Energy Demand from AI: Global context on data center electricity demand and AI-related energy growth. https://www.iea.org/reports/energy-and-ai/energy-demand-from-ai
4. McKinsey — The Cost of Compute: A $7 Trillion Race to Scale Data Centers: Context on projected global capital spending needed for data center growth. https://www.mckinsey.com/industries/technology-media-and-telecommunications/our-insights/the-cost-of-compute-a-7-trillion-dollar-race-to-scale-data-centers
5. Meta — America’s Workforce Academy: Context on Meta’s skilled-trades training program connected to AI infrastructure. https://about.fb.com/news/2026/06/americas-workforce-academy-free-skilled-trade-training/
6. OpenAI — PRC-linked Influence Operations Targeting AI Debates in the U.S.: Context on foreign influence activity connected to AI data center narratives. https://openai.com/index/prc-linked-influence-operations-ai-debates/
7. Microsoft Research — Project Natick: Context on Microsoft’s underwater data center research. https://natick.research.microsoft.com/

VP’s Take

VP’s Take: AI infrastructure is becoming a public trust issue as much as a technology issue. As data centers reshape conversations around power, water, jobs, and community benefit, organizations need clear research, honest messaging, and stakeholder engagement before concern turns into opposition.

– Sabiha Gire, VP Client Services

Related OS Services

Public Affairs

Major infrastructure projects need more than investment announcements. OS helps clients communicate with public officials, stakeholders, and communities so complex projects can move forward with clarity and trust.

Research & Strategy

Before communities support a project, organizations need to understand what people value, fear, and expect. OS helps clients use research, polling, and sentiment analysis to build strategies grounded in real public insight.

Crisis Communication

When public concern grows quickly, organizations need a steady, credible response. OS helps clients prepare messages, engage stakeholders, and protect trust during high-stakes public conversations.

Education & Workforce

As AI infrastructure increases demand for skilled trades, workforce communication becomes part of the public bargain. OS helps education and workforce organizations connect programs, employers, and communities through clear, authentic messaging.