There is 'a lot more pushback' on data center construction, Berkshire CEO Greg Abel tells CNBC

Berkshire Hathaway CEO Greg Abel said U.S. data-center construction faces growing community and political pushback, including a New York State moratorium and multiple other states with pending bans/restrictions. Abel noted Berkshire’s involvement is largely tied to power/electricity generation for the sites, with a focus on containing costs and avoiding impacts to other customers’ rates. Mizuho added that investors view data centers as a potential midterm election issue given claims about resource consumption and limited long-term job creation.

Analysis

The market is likely underestimating how quickly local resistance can turn a promised load-growth story into a permit-and-rate-case story. For Berkshire, the issue is not immediate earnings leakage; it is that the utility franchise loses some of the scarcity premium investors were willing to assign to incremental megawatt demand, transmission upgrades, and regulated asset growth. That matters most if hyperscaler demand was meant to offset slower growth elsewhere in the power system.

Second-order winners are regions and operators that can move faster with fewer local veto points: merchant generators, behind-the-meter power providers, and utilities with surplus capacity in friendlier jurisdictions. The losers are the companies whose pipeline depends on multi-year interconnection queues and public goodwill; even a modest delay can push capex into later periods and compress the present value of the backlog. For BRK.B, this is a multiple-risk story more than a near-term earnings risk, because the conglomerate’s utility unit is valued partly on long-duration growth optionality.

The key catalyst window is 1-3 months around state-level policy headlines and midterm-election positioning; the structural test is 6-18 months, when actual signed load agreements either keep compounding or stall. The contrarian view is that the debate is more about site selection than demand destruction: AI buildout probably migrates to lower-friction states rather than disappearing, so the long-run load theme may survive even if the optics worsen. Theses would be falsified if utility load forecasts, data-center interconnection requests, or transmission capex guidance keep inflecting higher despite the political noise.

Key Decisions for Investors

• Do not add to BRK.B on this headline; wait for Berkshire utility commentary and state-level permitting outcomes. The risk/reward is poor for immediate upside because the issue is valuation optionality, not a near-term earnings catalyst.
• Relative-value: short AEP or D against long CEG or VST over 3-6 months if you want to express the view that politically exposed regulated utilities lose the most while contracted/merchant power can reprice faster. Stop out if data-center load forecasts are revised up again.
• Watch XLU as a policy barometer rather than a standalone short: buy a 1-3 month put spread only if additional state restrictions spread beyond New York, because the trade works on headline repricing, not on fundamentals alone.
• Falsifier alert: if Berkshire Energy or peers report rising signed load commitments and no evidence of rate pressure in the next earnings cycle, treat the current pushback as noise and cover any bearish utility exposure.