- AI’s ongoing operating expenses may soon reach $2 trillion, according to the latest analyst estimates.
The report notes that earnings calls have disclosed capital costs—GPU purchases, power contracts, and land—but have not detailed the recurring spend needed to keep clusters running. Engineer Shashidhar Bhat is highlighted for developing tools that aim to reduce that recurring overhead.
If the $2 trillion figure holds, the recurring bill will dwarf the initial build‑out, pressuring cloud providers to improve efficiency. Bhat’s work could shave costs from power usage and hardware upkeep, offering a modest relief to the widening gap between capital outlay and ongoing expense.
The figure underscores a broader industry blind spot: while headline‑grabbing GPU purchases dominate headlines, the day‑to‑day cost of running AI at scale is already a multi‑trillion‑dollar challenge.