The AI Infrastructure Race Has Begun: What Economic Development Leaders Must Ask Now
September 28, 2025 | Zack Huhn + AI
The AI arms race isn’t about algorithms anymore. It’s about infrastructure.
With Nvidia’s latest $100 billion infrastructure commitment to OpenAI, alongside parallel moves from hyperscalers, telcos, and regional power companies, the world is witnessing the dawn of a new industrial buildout — one that rivals the scale of highways, grids, or railroads in its transformative potential.
This is not just about tech. It’s about economic development, energy, education, workforce strategy, and national competitiveness. The question now is: are policy and development leaders ready to respond?
Here are the key questions we should be asking — now, not later.
1. Can Our Infrastructure Handle AI’s Demands?
AI infrastructure isn’t plug-and-play. It requires vast amounts of power, water, land, cooling, bandwidth, and supply chain reliability.
Is our electric grid up to the task? The energy intensity of modern AI workloads is staggering — and rising.
Do we have the land, water, and permitting capacity for the data centers, chip fabs, and edge nodes these models require?
What about connectivity? Fiber, low-latency networks, and backhaul will separate winners from watchers.
This is a moment to treat AI infrastructure like a utility. Because it is one.
2. What’s Our Strategic Bet — and Who Are We Building For?
The stakes are too high for generic plans.
Are we targeting training or inference? Hyperscale or edge? Public or private workloads?
Are we clustering AI with existing industry strengths — like healthcare, logistics, defense, or agriculture — to drive outsized ROI?
Are our public-private partnerships structured for long-term value, not just short-term headlines?
Infrastructure is destiny. The regions that align industrial policy with AI’s physical footprint will win.
3. Who Gets to Play — and Who Gets Left Out?
AI has the potential to widen the gap between “have” and “have not” regions.
Are we equipping non-metro or historically underinvested areas to compete?
Are local startups and SMEs being priced out of access to compute and infrastructure?
What are we doing to ensure broad-based job creation, not just elite talent acquisition?
This can’t be another story of digital deserts. It must be a story of shared digital opportunity.
4. What’s the Environmental Cost — and Who Pays It?
The AI era could set sustainability back a generation — or propel it forward.
Data centers are power-hungry and water-thirsty. Are we mandating low-carbon, renewable sources?
How do we measure compute per watt or inference per kilowatt hour?
Can we build standards that ensure climate alignment even as we scale?
If AI eats the planet to save it, we’ve failed.
5. Are We Regulating Infrastructure — or Just the Algorithms?
While public debate focuses on AI model behavior, the physical layer remains underexamined.
Who governs access to compute?
Should hyperscalers be required to offer fair access to SMEs, researchers, and governments?
What transparency or accountability is required of the infrastructure owners, not just the model developers?
Ignore this layer, and you risk regulating the icing while the cake bakes in secret.
6. What Are Our Metrics — and Are We Even Measuring the Right Things?
We can’t manage what we don’t measure.
Where is compute capacity actually located?
Who’s using it — and who’s locked out?
What’s the job creation per megawatt? The startup output per petaflop?
Good policy starts with good baselines. Right now, we’re flying blind.
7. Are We Building for Resilience — or Vulnerability?
If the U.S. has all the models but no fabs… if your city has the data but no grid… if your startup has ideas but no access to GPUs — you’re vulnerable.
We must ask:
How diversified are our supply chains?
What backup exists for critical infrastructure?
Can we withstand cyberattacks, climate shocks, or geopolitical ruptures?
AI infrastructure must be treated as critical infrastructure. Because it already is.
The Window Is Now
AI infrastructure is being sited, built, and locked in right now. Tax incentives are being written. Power deals are being signed. Land is being rezoned. Permits are being fast-tracked.
The regions that act early and think holistically will be positioned not just to host AI infrastructure — but to shape the future built atop it.
As Chair of the Enterprise Technology Association, I’ve seen firsthand how economic development, workforce, and digital transformation leaders can align — but only if they know what’s at stake.
This is your moment. Don’t let it pass while others pave the road ahead.
