Greg Abel Says Berkshire Will Serve the Hyperscalers Only If Its Other Customers Don't Pay for It

Berkshire Hathaway CEO Greg Abel said Berkshire will serve AI/hyperscaler data centers only if they deliver a net benefit without raising rates for other utility customers. U.S. utilities after-tax earnings rose 38% YoY to $597 million in Q2, while utility margin expanded 8% YoY, supported by higher retail volumes (+3% through June). The company is investing heavily—$10.6B in first-half capex with $6.7B tied to energy—and is forecasting about $8.6B more for the remainder of 2026, though Abel flagged increasing community pushback risk.

Analysis

This is more important for BRK.B/BRKA as a compounding utility story than as a near-term trading headline. The mechanism is straightforward: hyperscale load is effectively a long-duration growth annuity for regulated rate base, but only if the utility can keep the load from being politicized into a broader customer-subsidy fight. That makes Berkshire’s edge less about absolute demand and more about regulatory credibility; the same AI buildout that stresses other utilities may actually widen Berkshire’s moat if it keeps winning approvals on “incremental load, incremental capex” terms.

The second-order winner is the utility complex with transmission, generation, and siting optionality in politically receptive states. The losers are utilities that need to absorb data-center load without clean cost recovery, because they risk margin dilution, delayed rate cases, or community backlash that can freeze interconnects. Over the next 1-3 months, the key catalyst is not load growth itself but whether management commentary and state filings confirm a larger 2026-27 capex pipeline; over 6-18 months, the real driver is whether rate base growth can outrun the earnings drag from financing costs and construction lag.

The market may be underweighting the concentration risk embedded in “AI load” enthusiasm: if a few hyperscalers dominate incremental demand, bargaining power stays with the customer and regulators can force concessions if public pushback rises. The thesis is falsified if community resistance turns into a formal moratorium in a Berkshire-served state, or if upcoming utility results show the 38% earnings cadence was materially helped by one-offs rather than recurring load-driven margin expansion. For Berkshire equity, this is still constructive, but the upside is likely steadier compounding rather than a rerating unless investors start valuing BHE as a faster-growth regulated asset base.

Key Decisions for Investors

• Long BRK.B / short XLU over the next 3-6 months: Berkshire has a clearer path to rate-base growth from AI load with less subsidy risk than the average utility; target is modest relative outperformance rather than absolute squeeze.
• Accumulate BRK.B on weakness if the stock sells off on any general market rotation away from mega-cap quality; the utility story is a 12-24 month compounding driver, but near-term price should still be anchored by insurance/cash flow.
• Watch AEP, DUK, and NEE as read-throughs on data-center monetization and regulatory friction; if they guide to slower interconnects or higher capex with no commensurate earnings uplift, that supports a relative long BRK.B stance.
• Set a catalyst alert on Berkshire utility filings and state PUC commentary in the next 1-2 quarters; if rate recovery language weakens or capex is deferred, fade the utility-growth thesis and reduce BRK.B exposure.
• No aggressive options trade here unless there is a sharper selloff: the opportunity is better expressed as a patient relative-value position than as a binary event-driven call.