America Must Rapidly Build Physical AI Infrastructure or Lose the Race to China
Source: "AI race depends on data centers, power grid and digital infrastructure | Fox News." September 18, 2026. www.foxnews.com
The Gist
The authors argue that America's biggest challenge in the AI race isn't inventing better technology—it's building enough power plants, electrical grids, and data centers fast enough to run that technology. They warn that China is racing ahead on this physical buildout while the US gets bogged down in permitting delays and grid bottlenecks, and they suggest military veterans could help fill the skilled labor shortage needed to build this infrastructure.
Conclusion
The United States must urgently accelerate the buildout of power generation, transmission, data centers, and skilled workforce (the 'Fifth Utility') to win the AI race, or it will lose technological leadership to China regardless of software innovation.
Premises
- Winning the AI race depends not just on developing the best software or algorithms, but on building the physical infrastructure (power, transmission, data centers) needed to run AI at scale.
- Permitting delays, grid bottlenecks, and supply-chain constraints are currently slowing infrastructure buildout in the US at precisely the moment AI demand for electricity is surging.
- China is aggressively expanding power generation, transmission, manufacturing, and computing capacity because it understands technological leadership requires physical capacity.
- While Washington debates AI governance, competitors like China are focused on building deployment infrastructure.
- Data centers should be understood as part of a critical 'Fifth Utility' underpinning banking, healthcare, transportation, commerce, emergency communications, and national defense—not merely commercial buildings.
- Building this infrastructure requires a skilled workforce (electricians, welders, high-voltage technicians) and the ~200,000 veterans leaving military service annually are well-suited to fill this gap.
- America has historically demonstrated the capacity to turn innovation into physical infrastructure (highways, ports, power plants), showing this is achievable again if prioritized.
Assumptions
- China's infrastructure expansion will translate directly into AI capability leadership if the US does not match it.
- The primary bottleneck to US AI competitiveness is physical infrastructure rather than software innovation, talent, or capital allocation.
- Community concerns about data centers (power, water, rates) can be adequately addressed without significantly slowing infrastructure development.
- Faster permitting and reduced regulatory friction will not create unacceptable risks to communities or the environment.
- Veterans transitioning to civilian careers represent an underutilized and readily available workforce solution for this specific industry.
- The 'AI race' framing (US vs. China as a zero-sum competition) is the correct lens for evaluating infrastructure policy.