Google warns AI growth outpaces US energy supply

• Google’s president says the United States may not be expanding its energy capacity fast enough to support the rapid rise of artificial intelligence.
• She argues that large‑scale data centers require far more electricity than current infrastructure can reliably deliver.
• Her comments highlight growing concerns about the energy demands of next‑generation AI systems.

Google’s President and Chief Investment Officer Ruth Porat has warned that the United States may not be developing new energy resources quickly enough to support the accelerating growth of artificial intelligence. Speaking at the CERAWeek conference in Houston, she emphasized the scale of electricity required to operate and expand the company’s AI‑driven data centers. These facilities rely on vast amounts of power to train and run advanced models, and their consumption is rising sharply as AI adoption increases. Porat noted that the company is concerned the country is not moving “full throttle on energy,” underscoring the urgency of the issue.

Her remarks come at a time when AI infrastructure is expanding across multiple industries. Companies are racing to build larger data centers, deploy more powerful chips and support increasingly complex workloads. These developments require significant electrical capacity, often exceeding what local grids can provide without upgrades. Energy availability has therefore become a strategic factor in determining where new AI facilities can be built.

The challenge is not limited to Google, as other technology companies face similar constraints. AI models require continuous access to high‑performance computing resources, which in turn depend on stable and abundant electricity. Grid operators are already seeing rising demand from data centers, electric vehicles and industrial electrification. This convergence is placing pressure on energy planners to accelerate investment in generation, transmission and storage.

Growing strain on the energy system

Porat’s comments reflect a broader debate about the readiness of the U.S. energy system for the next wave of digital infrastructure. Many regions are experiencing delays in connecting new power projects to the grid, slowing the pace of renewable energy deployment. These bottlenecks complicate efforts to meet the needs of AI data centers, which often require long‑term, high‑capacity power contracts. Some states are exploring new regulatory frameworks to speed up approvals and encourage investment in clean energy.

Electricity demand from data centers is projected to grow significantly over the next decade. Analysts expect AI‑related computing to become one of the fastest‑growing sources of power consumption in the United States. This trend is prompting utilities to reassess their long‑term planning assumptions. Several companies have already announced major grid expansion projects to accommodate rising digital demand.

The energy intensity of AI workloads is also driving interest in alternative power sources. Some technology firms are exploring small modular nuclear reactors as a potential solution for long‑term, high‑capacity power. Others are investing in large‑scale renewable energy projects paired with battery storage. These strategies aim to secure reliable electricity while reducing environmental impact.

Industry and policy implications

Porat’s remarks at CERAWeek highlight the intersection of technology policy and energy planning. AI development is increasingly tied to national competitiveness, making energy availability a strategic concern. Policymakers are being urged to consider how permitting, infrastructure investment and regulatory frameworks affect the pace of digital innovation. Industry leaders argue that without sufficient power, the United States risks falling behind in the global AI race.

The discussion also raises questions about sustainability. AI systems require substantial energy, and their carbon footprint depends heavily on the sources powering data centers. Companies are under pressure to balance performance with environmental responsibility. Many have committed to long‑term renewable energy goals, but achieving them will require significant grid modernization.

CERAWeek has become a key venue for conversations about the future of energy and technology. This year’s event brought together executives, policymakers and researchers to discuss how emerging technologies will shape global energy demand. AI was a central theme, reflecting its growing influence across sectors. Porat’s comments added urgency to an already complex debate about how to power the next generation of digital infrastructure.

Recent studies estimate that global data center electricity consumption could double by 2030, driven largely by AI workloads. Some researchers warn that without major efficiency improvements, AI could become one of the largest industrial energy consumers worldwide. This trend is prompting renewed interest in energy‑efficient chip design and advanced cooling technologies, which may play a crucial role in managing future demand.