Local fights over data center water and power expose AI policy gap

The rapid rollout of artificial intelligence is upending long-held assumptions about who succeeds in the American economy and at what cost to the planet. Companies racing to build data centers and fiber networks are discovering they cannot rely on the traditional pipeline of recent college graduates. Instead, they are scrambling for electricians, technicians and other skilled tradespeople who can handle the physical infrastructure AI demands.

AT&T chief executive John Stankey has been blunt about the shortfall. The company needs workers who understand electricity and photonics and who can install equipment inside homes and offices. Stankey told interviewers the firm must actively recruit, train and retain these employees because the United States is not producing enough of them. This shift comes as record numbers of students prepare to graduate with four-year degrees that once promised steady entry-level roles now increasingly automated by AI tools.

The same infrastructure boom is reshaping energy markets. A new BloombergNEF analysis projects solar power will overtake coal, oil and natural gas as the largest source of electricity by 2035, driven largely by falling costs. Countries such as Pakistan have already added tens of gigawatts of solar after gas prices spiked. Yet the forecast carries a major caveat: data centers are expected to require an extra terawatt of utility-scale solar plus hundreds of gigawatts of natural gas and coal to guarantee round-the-clock operation. Fossil fuels are projected to supply more than half the additional power these facilities will consume through mid-century.

That outcome reflects the priorities of tech firms and data-center developers more than any broad climate strategy. While long-duration batteries, geothermal and nuclear are competing for a share of the market, gas and coal remain attractive because they run continuously without depending on weather. The result is an energy system that adds record amounts of renewables yet keeps substantial fossil generation online to serve the needs of a handful of large corporate customers.

For parents, the contrast between the natural world and the economic one is becoming harder to explain to children. Nature documentaries that show predators and prey in brutal detail often prompt questions about fairness and survival. When those same lessons appear in conversations about job markets and energy choices, the answers can feel equally stark. Families weighing whether to limit graphic animal footage may soon confront similar choices about the systems powering the devices their kids use every day.

The emerging pattern is one in which the AI economy rewards hands-on skills that colleges have long de-emphasized, while the infrastructure supporting that economy extends the life of older energy sources. Policymakers and companies have yet to reconcile those competing pressures with the broader goals of broad-based opportunity and emissions reductions.