United States Grid May Fail to Support AI Growth
The nation's aging power infrastructure threatens to create a deficit for data centers by 2030.
Updated on Oct. 15, 2026 in Data Centers
Live Poll
Is the U.S. power grid currently capable of handling the energy demands of future technology?
Flex CEO Revathi Advaithi warned that current United States energy infrastructure is insufficient to support the scaling of artificial intelligence. A 2025 Berkeley study projected that data centers could consume 11.8% of the nation's total electricity by 2030.
Why it matters
The rapid expansion of AI models is outstripping existing grid capacity, which has seen little innovation since the Edison era. This infrastructure deficit threatens to limit the development of new semiconductor technology, which currently requires a five-year lead time for creation.
A 2025 study from Berkeley estimates that data center electricity consumption will reach 11.8% of total U.S. power demand by 2030. This growth is compounded by long-term cycles, where semiconductor development alone spans five years, far faster than current grid upgrades.
The players
Revathi Advaithi
The CEO of Flex, a global manufacturing company specializing in supply chain management and electronics assembly for the data center and semiconductor sectors.
The details
The United States power grid relies on legacy transmission and distribution systems that have not undergone significant innovation since the Edison era. Expansion is currently throttled by safety standards and regulatory approval processes that require decades to navigate. Upgrading this infrastructure requires fundamental changes to transmission line capacity and next-generation distribution models.
Timeline
2025: A Berkeley study was published regarding data center energy demand.
2030: Data centers are projected to account for 11.8% of U.S. electricity usage.
Next decade: The U.S. faces a projected power gap and potential energy deficit.
The Tech Race
This bottleneck marks a departure from traditional infrastructure scaling, where energy supply reliably tracked industrial hardware cycles. As identified in the 2025 Berkeley study, the disconnect between AI hardware demand and grid innovation now threatens to stall sector-wide progress.
The potential power gap may cause systemic delays for developers and firms relying on AI compute resources as grid capacity becomes a limiting factor. Organizations will likely face increased scrutiny regarding their energy procurement strategies over the next decade.
The takeaway
The United States faces an energy deficit as power grid development continues to lag behind rapid artificial intelligence advancements. Watch for future regulatory policy changes that aim to accelerate approval processes for power transmission projects.
Further reading
For broader context on current hardware-to-infrastructure dynamics, see Data Centers.
Source note: This article includes information reported by Fortune.
Live Poll
Is the U.S. power grid currently capable of handling the energy demands of future technology?
