AI is fueling a massive global surge in data centers

The boom stems from AI's hunger for huge amounts of computing power, electricity, cooling, and connections. Goldman Sachs forecasts data center power demand will climb 50% by 2027 and 165% by 2030.

Investors are piling in: A CBRE survey found 95% of major players plan to boost data center bets, with 41% eyeing $500 million or more in equity this year—up from 30% last year. This shift is turning real estate from flashy offices to "invisible" tech warehouses, as CapitaLand's Kishore Moorjani put it.

Yet costs are sky-high: A big hyperscale center (150-300 megawatts) runs $12 million per megawatt, and AI beasts over 1 gigawatt demand billions. Boston Consulting Group says hyperscalers must pour $1.8 trillion into this from 2024 to 2030.

Banks are buckling under the load, with Moorjani noting the "sheer volume" of builds is straining them. JLL's Stuart Crow asks bluntly: "Is there enough capital at the moment?" Hurdles like soaring construction prices, worker shortages, and red tape aren't helping, as global real estate investment lags forecasts.