Phoenix
TSMC's $165B semiconductor bet — the largest foreign investment in US history — in a desert where the Colorado River may breach its compact by 2027. A bet, not a plan.
TSMC is investing $165 billion in Phoenix — the largest foreign direct investment in American history — to build semiconductor fabs that consume roughly 5 million gallons of ultra-pure water per day. The city gets 40% of its water from the Colorado River, which is entering its twenty-fifth year of drought, with Lake Mead at roughly 30% capacity.
This is not a contradiction the city is unaware of. It is a bet that technology will outrun hydrology. Phoenix's metro area of nearly 5 million people grew faster than any large US city over the past two decades, adding population at roughly 1.2% annually in a place that recorded 113 consecutive days above 100°F in 2024 and projects 47 days per year above 110°F by 2050.
Phoenix is building America's semiconductor independence on a $165 billion TSMC investment — in a desert where Lake Mead sits at 30% capacity and the Colorado River compact may breach by 2027.
The semiconductor corridor is enormous: three fabrication plants, two packaging facilities, a major R&D centre, and 2,000 acres of acquired land. Fab 1 is already producing 4nm and 5nm chips. Fab 2 targets 2nm production by 2028. This creates 40,000 construction jobs and locks Phoenix into a high-water-intensity industrial future for decades.
Arizona's Tier 1 water shortage restrictions already cut the state's Colorado River allocation by 512,000 acre-feet — 18% of its share. The Upper Basin snowpack runs at 69% of normal. Water managers warn the compact between seven states and Mexico may breach its structural thresholds within two years.
Phoenix is evolving for one climate regime while the environment shifts to another — the biological equivalent of a species optimising for conditions that no longer exist. The urban heat island adds 5°F on top of global warming. The city's solution is to grow faster, attract more investment, build more infrastructure, and trust that desalination, recycling, and groundwater management will close the gap between what the desert provides and what the semiconductor fabs demand.
The bet may work. But it is a bet, not a plan.