
Warren Buffett's Berkshire Hathaway has largely stayed out of AI stocks, but its wholly owned subsidiary Berkshire Hathaway Energy is emerging as an indirect AI winner. The company's regulated utilities, including MidAmerican in Iowa and NV Energy in Nevada, are seeing surging electricity demand from AI data centers—now accounting for roughly 8% of peak load in Iowa—and are investing roughly $34 billion(約5.4兆円) to expand generation and transmission. CEO Greg Abel said about half of the energy unit's operations now serve AI-related power needs, allowing Berkshire to profit from the AI boom through the unglamorous but steady business of powering data centers.
Summaries like this, in your inbox every morning.
Sign up free →What happened
Berkshire Hathaway Energy, a wholly owned subsidiary of Warren Buffett's conglomerate, is benefiting from surging electricity demand driven by AI data centers. CEO Greg Abel stated that about half of the energy unit's operations now address AI-related power needs, supported by a roughly $34 billion(約5.4兆円) capital plan to expand generation, storage, and transmission.
Why it matters
Regulated utilities earn steady returns in two ways—selling more electricity as demand rises, and earning a fixed return on capital invested to serve that demand. In Iowa alone, data center clusters now account for roughly 8% of peak electricity load, with management expecting consumption to keep climbing for years. This gives Berkshire indirect AI exposure without the volatility of chip stocks.
What to watch
Utility returns depend on regulators approving rate increases, which is not guaranteed. Berkshire Hathaway Energy also carries real liabilities, including wildfire exposure at PacifiCorp that has cost it dearly. Even a thriving energy unit will not move Berkshire's overall stock as dramatically as a hot chip stock might, making this a slow, steady contributor rather than a high-growth bet.
Warren Buffett built Berkshire Hathaway by avoiding investments he did not understand, a philosophy that has kept the company largely absent from the AI stock boom. However, this defensive stance masks a significant but overlooked source of AI exposure: Berkshire Hathaway Energy, a wholly owned subsidiary that operates regulated utilities across the country.
Berkshire Hathaway Energy owns major utilities including MidAmerican in Iowa, NV Energy in Nevada, and PacifiCorp in the West. These companies are now experiencing unprecedented demand growth driven by AI data centers, which consume enormous amounts of electricity. In Iowa, a cluster of mega data centers now accounts for roughly 8% of peak electricity load, and management expects data center consumption to keep climbing for years. This surge presents a structural opportunity for the utilities that serve them.
The business model of regulated utilities is built on two reinforcing profit mechanisms. First, utilities earn revenue by selling more electricity as demand rises. Second, and more importantly, they earn a regulated return on capital invested to build the infrastructure—generation, storage, and transmission—needed to serve that demand. Berkshire Hathaway Energy is capitalizing on this dynamic through a roughly $34 billion(約5.4兆円) capital plan to expand its infrastructure. Every dollar of approved investment becomes a base on which the company earns steady profits for decades. Notably, CEO Greg Abel told shareholders that about half of the energy unit's operations are now addressing AI-related power needs, a striking statement for a company traditionally associated with power lines and pipelines rather than artificial intelligence.
Yet the article notes important caveats. Utilities grow slowly and require enormous capital expenditures, and their returns depend on regulators approving rate increases—a process never guaranteed. Additionally, Berkshire Hathaway Energy carries real liabilities, particularly wildfire exposure at PacifiCorp that has been costly. Because Berkshire is so vast, even a thriving energy unit will not move the overall stock as dramatically as a hot chip stock might. The takeaway is that while Berkshire Hathaway itself has barely touched AI stocks, it owns a genuine stake in the electricity infrastructure powering the entire AI boom—a slow, steady contributor that reflects the reality that AI's economic beneficiaries extend far beyond the obvious names in semiconductors and software.
Berkshire Hathaway has built its reputation by avoiding industries the company does not fully understand, which has largely kept it on the sidelines of the AI stock frenzy dominated by chipmakers and AI software companies. Yet the subsidiary Berkshire Hathaway Energy reveals an unexpected angle: AI's electricity hunger creates a structural, long-term opportunity in regulated utilities that few observers have highlighted.
The mechanism is straightforward. AI data centers consume enormous amounts of power, and utilities that own generation, storage, and transmission infrastructure can serve that demand while earning regulated returns on capital invested. Unlike hot chip stocks that depend on market sentiment and competition, regulated utilities enjoy government-approved rate structures that lock in margins on approved investments. Greg Abel's statement that roughly half of Berkshire Hathaway Energy's operations now address AI-related power needs signals that this is not a speculative bet but an operational reality shaping the business's capital allocation.
However, the article cautions that utility returns are not risk-free. Rate increases require regulator approval, and Berkshire Hathaway Energy carries material liabilities—PacifiCorp's wildfire exposure has been costly. Moreover, even a thriving energy unit operates at the scale of a slow, steady contributor within a conglomerate as large as Berkshire, unlikely to drive outsized stock gains. The broader lesson is that AI's economic beneficiaries are far more diverse than the public discourse around chip stocks suggests.
AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.
Free · takes 30 seconds · unsubscribe anytime
No comments yet. Be the first to share your thoughts!
Log in to join the discussion





Get curated AI news from 200+ sources delivered daily to your inbox. Free to use.
Get Started FreeFree · takes 30 seconds · unsubscribe anytime
1 minute a day. The AI essentials.
200+ sources · Email / LINE / Slack