Data centers forecast to use 20% of US power by 2035 — analysts estimate usage will rocket to 194 gigawatts, 83% more than forecast seven months ago

Summary

Data centers will consume 20% of US power by 2035, a significant increase from today. This surge is driven by AI workloads, with projections for demand and grid capacity shortfalls. Lead times for transformers are also a growing concern.

Key Takeaways

1. U.S. data centers are on track to consume about 20% of the nation's electricity by 2035, up from 5.9% today, according to a new BloombergNEF forecast.
2. The research firm raised its 2035 demand projection to 194 GW, an 83% increase on the forecast it issued in December, and calculated that even a record pace of grid connections sustained every year for a decade would leave a 19 GW supply shortfall.
3. Nearly half of the projected capacity will serve AI training and inference workloads, and the U.S. is expected to host 64% of the world's AI chips by power demand in 2033.

Statement Breakdown

• Claimed Facts: 60% of statements the article presents as facts
• Opinions: 30% of statements classified as editorial or subjective
• Claims: 10% of statements surfaced for additional reader evaluation

Credibility & Bias Reasoning

Credibility assessment: The article relies on forecasts from reputable research firms like BloombergNEF and EPRI. It clearly attributes data and projections to these sources, enhancing its credibility. However, the speculative nature of future AI development introduces an element of uncertainty.

Bias assessment: Technological Optimism. The article focuses on the rapid growth and future projections of data centers, particularly driven by AI. While presenting data, it leans towards highlighting the expansion and potential challenges of this growth, framing it as an inevitable technological advancement.

Note: This article presents projections based on current trends and analyst forecasts. Future developments in AI and energy infrastructure could significantly alter these outcomes.

Credibility flag: Future-focused, data-driven

Claimed Facts (8)

• This is a direct projection presented as a factual outcome based on a specific forecast.
• This presents specific numerical data and a calculated outcome from the research firm.
• This provides historical data points of previous forecasts, illustrating the trend of increasing projections.
• This states a specific allocation of projected capacity and a future expectation regarding AI chip demand.
• This presents a forecast from another independent entity, providing corroborating data on global trends.
• This provides a historical benchmark for grid connections.
• This provides specific regional projections for electricity consumption by data centers.
• This presents a direct causal link between data center demand and increased power costs, as reported by a market monitor.

Opinions (4)

• The phrase 'moving in the same direction' is an interpretation of trends rather than a direct factual statement.
• The phrase 'Close to half' is an estimation and projection, not a definitively proven fact.
• The term 'speculative' is an assessment of the nature of the projects, reflecting an opinion on their certainty.
• This is a metaphorical statement used to illustrate the scale of energy consumption, rather than a precise, quantifiable fact.

Claims (5)

• This is a projection of a shortfall based on specific assumptions about 'record pace' and sustained connections, which are hypothetical.
• This is a highly specific future projection about global market share, which is inherently uncertain and difficult to verify.
• This is a projection of a shortfall based on a hypothetical 'pace' and a 'base case,' making it speculative.
• While attributed to an analyst, the statement is a broad generalization and an analogy rather than a precise, verifiable calculation.
• Attributing a specific percentage increase in costs directly to one factor can be an oversimplification, as power costs are influenced by multiple variables.

Key Sources

• BloombergNEF — Research Firm
• Gartner — Research Firm
• TechCrunch — Technology News Outlet
• PJM's independent market monitor — Independent Market Monitor
• Lloyd Arnold — BNEF Analyst