Power Shortages Threatened Nvidia Revenue Projections
Analysts warn that electricity supply constraints for U.S. data centers could impact future semiconductor earnings.
Updated on Sept. 21, 2026 in Data Centers
Live Poll
Do you believe current energy supply constraints make this a risky time to invest in AI?
VanEck analysts projected a risk that Nvidia could miss earnings estimates due to electricity shortages affecting data center deployments. This analysis evaluates the potential for hardware demand to outpace available U.S. grid capacity through 2027.
Why it matters
The constraint highlights a potential bottleneck where customer inability to secure sufficient electricity for new data centers prevents the operational deployment of chips from Nvidia, AMD, and Broadcom. This risk marks a shift from focusing solely on semiconductor supply chain constraints to broader grid and infrastructure limitations.
VanEck analysts estimate every 1 gigawatt of power shortfall represents approximately $37 billion in lost revenue, with chips from Nvidia, AMD, and Broadcom requiring a collective 30 gigawatts through 2027. This demand exceeds the projected 15 to 25 gigawatts of new U.S. data center capacity expected to come online annually.
The players
Nvidia
A designer of graphics processing units and data center hardware that has exceeded revenue estimates for 16 consecutive quarters.
VanEck
An investment management firm that analyzes industry risks and market impacts for technology and infrastructure sectors.
AMD
A semiconductor company that develops high-performance processors and competes in the data center hardware market.
Broadcom
A semiconductor and infrastructure software company that provides critical components for data center networking.
The details
The risk is calculated based on the energy-intensive nature of high-performance computing clusters required to run modern AI workloads. When data center operators fail to secure sufficient electricity to energize their hardware, the chips remain idle or uninstalled. Consequently, analysts have identified that each gigawatt of unrealized power capacity correlates to an estimated 6% impact on revenue.
Timeline
Through 2027, chips from major manufacturers are expected to require 30 gigawatts of U.S. power.
Nvidia is expected to release its third-quarter earnings in November 2026.
The Tech Race
This risk analysis places current hardware deployment trajectories against the physical limitations of the U.S. electrical grid. It challenges the assumption that rapid data center scaling can match the power-heavy requirements of current semiconductor roadmaps.
This development suggests that future enterprise access to AI computing resources may be dictated by local utility grid capacity rather than hardware supply. Companies looking to integrate or scale high-density data operations should prioritize grid-ready locations to avoid project delays.
The takeaway
The trajectory of high-performance computing is currently tethered to the rate of utility infrastructure development in the United States. Observers should track regional energy availability benchmarks to determine if revenue growth for chip manufacturers begins to moderate in the coming quarters.
What happens next
Investors will monitor Nvidia’s third-quarter earnings report, expected in November 2026, to assess whether these grid constraints have begun to impact actual revenue.
Further reading
For more on infrastructure trends, see the Data Centers section.
Source note: This article includes information reported by Benzinga.
Live Poll
Do you believe current energy supply constraints make this a risky time to invest in AI?
