Initiating Berkshire at Hold: The Operating Engine Compounds Fine — It's the $344B of Idle Cash, the Buyback Drought, and the Succession Clock That Keep Us Neutral

Results vs. Expectations

Berkshire does not guide, does not hold a call, and is only thinly covered on a consensus-EPS basis, so the honest scorecard is year-over-year operating earnings — the after-tax measure Berkshire itself emphasizes, which strips the mark-to-market gyrations of the equity portfolio out of the result. On that basis the quarter was a modest step backward at the headline and a step forward underneath it.

Operating earnings by segment (after-tax)

Revenue / earnings quality

• Operating earnings: Decline is 100% attributable to non-cash FX remeasurement and a prior-year catastrophe base effect. Ex those, operating earnings grew. Durable.
• Underwriting: The −12% is the soft spot. GEICO remained excellent (pre-tax underwriting $1,821M, +2%, combined ratio 83.5%), but BH Primary fell to $63M from $279M on social-inflation-driven adverse development. Worth watching, not yet a thesis problem.
• GAAP net earnings: Dominated by the $4,970M after-tax investment gain and the $3,760M Kraft Heinz impairment. Not a measure of operating performance; we anchor on operating earnings.

Segment Performance

Insurance — Underwriting & GEICO

Underwriting earnings of $1,992M after-tax were down 12% year-over-year, but the composition matters. GEICO continued its multi-year turnaround: pre-tax underwriting profit of $1,821M (up 2%), a combined ratio of 83.5%, and a loss ratio that improved to 71.8% as the rate actions of 2023–24 fully earned in. The expense ratio ticked up to 11.7% as GEICO leaned back into policy-acquisition spend — a deliberate trade, growing policies-in-force again (premiums written +5.2%) now that margins are healthy. The drag came from BH Primary, where underwriting profit collapsed to $63M from $279M on social-inflation-driven adverse development in casualty lines.

Assessment: GEICO is firing on both cylinders for the first time in years — margin and growth simultaneously — and that is the single most important operating story inside Berkshire right now. The BH Primary softness is the kind of casualty-reserve noise that recurs across the industry; we flag it as a watch item, not a structural crack. Net-net, the underwriting franchise is in good health and the float it generates (~$174B, at negative cost) is the engine that funds everything else.

Insurance — Investment Income

Investment income of $3,367M after-tax (+1.4%) is the quiet beneficiary of the cash pile. With $344B parked largely in short-dated Treasury bills yielding ~4–5%, Berkshire is earning real money on its "problem." This line has roughly doubled off its 2021 trough as rates normalized.

Assessment: This is the double-edged sword of the cash hoard. As long as front-end rates stay elevated, the idle cash earns a respectable return and investment income holds up. But it is the lowest-return use of capital Berkshire has, and the moment short rates fall, this line falls with them. The cash is a liquid call option on a market dislocation; the cost of holding it is the gap between a 4–5% T-bill yield and the long-run return on the businesses Berkshire would rather be buying.

Railroad — BNSF

BNSF was the quarter's standout: net earnings of $1,466M, up 19.5%, on roughly flat revenue of $5,726M. The driver was margin — the operating ratio improved to approximately 64.8% from ~68.2% on lower fuel costs, operating-cost discipline, and a lower effective tax rate. Volumes rose 1.4%.

Assessment: A near-20% earnings gain on flat revenue is pure productivity, and it narrows the long-standing margin gap between BNSF and best-in-class peer Union Pacific that has nagged Berkshire watchers for years. New BNSF leadership has been explicit about closing that gap. If the operating-ratio improvement proves durable rather than fuel-driven, BNSF re-rates as a structural earnings-grower rather than a flat-line cyclical — a quietly important development.

Utilities & Energy — BHE

BHE net earnings of $702M rose 7.2%. Critically, there were no new wildfire-litigation accruals in the first half of 2025, versus $251M in the prior-year quarter — the absence of fresh reserves was itself a tailwind. PacifiCorp's cumulative probable wildfire losses stand at roughly $2.75B, of which ~$1.38B remains unpaid.

Assessment: BHE is the segment carrying the most tail risk in the entire enterprise — the PacifiCorp wildfire liabilities are real, open-ended, and have already capped Berkshire's appetite to inject more equity into the utility. The 7% earnings growth is fine; the story is the contingency. Until the wildfire exposure is fully bounded, BHE trades at a discount inside the sum-of-the-parts, and rightly so.

Manufacturing, Service & Retailing

The sprawling collection of industrial, consumer, and retail businesses earned $3,601M after-tax, up 6.5% — broad-based and resilient against a mixed macro backdrop. This bucket spans Precision Castparts, Marmon, Lubrizol, the building-products group, the retail operations, and dozens more.

Assessment: This is the Berkshire that doesn't make headlines and doesn't need to. Mid-single-digit growth across a diversified industrial/consumer base, in an economy that has been anything but uniform, is exactly the steady-compounding profile that justifies owning the conglomerate. No single business is large enough to move the whole, which is the point.

Other — The FX Swing

The "Other" line fell to $32M from $753M, a $721M swing that accounts for the entire year-over-year decline in operating earnings. The culprit is ~$877M of after-tax foreign-exchange losses on Berkshire's substantial yen- and euro-denominated debt, recorded as the dollar moved against those currencies. This is a non-cash remeasurement that reverses as exchange rates move the other way.

Assessment: Do not capitalize this number. FX remeasurement on the funding Berkshire raised to buy the Japanese trading houses (and other non-USD obligations) is a quarterly wildcard that nets to roughly zero over time. It made this quarter look worse than it was; it will make some future quarter look better than it is. We treat it as noise — but we flag that it is large enough to dominate the headline, which is exactly why the segment table, not the operating-earnings total, is where the truth lives.

Key Themes

The Cash Pile Is Now the Story

At $344.1B in cash and Treasury bills — roughly 35% of the company's ~$990B market capitalization — Berkshire's liquidity is no longer a footnote; it is the central fact of the investment case. The pile has been built by the operating engine's cash generation, the steady drip of equity sales, and the absence of any large acquisition or buyback to absorb it.

Assessment: A record cash position is simultaneously Berkshire's greatest strength and its most pointed weakness. It is unmatched optionality — the ability to write a $50B+ check into the next dislocation when nobody else can. But optionality has a carrying cost, and a pile this size earning T-bill yields is a drag on return on equity every quarter it sits there. The bull and the bear are looking at the same $344B; the question is whether you trust it to be deployed brilliantly or fear it sits idle indefinitely.

Zero Buybacks, Again

Berkshire repurchased no stock in Q2 2025 — and none in the first half of the year. The buyback machine that retired tens of billions of dollars of stock in 2020–2021 has been dormant since mid-2024.

Assessment: This is the tell. Berkshire only repurchases when management judges the stock to be below a conservative estimate of intrinsic value, and the absence of buybacks at ~1.5x book is a quiet statement that the people who know the business best do not consider it a screaming bargain here. That is useful information for an outside shareholder: the insider with $344B and a standing authorization is choosing T-bills over his own stock. We take that seriously in setting our own rating.

The Eleven-Quarter Net-Selling Streak

Berkshire was again a net seller of equities, trimming roughly $4.5B more than it bought in the first half — the eleventh consecutive quarter as a net seller. The five largest holdings (American Express, Apple, Bank of America, Coca-Cola, Chevron) now represent about 67% of the equity portfolio, down from 71%, as the concentration eases.

Assessment: Persistent net selling into a rising market, while building cash and declining to buy back stock, is a coherent posture with one interpretation: management sees little that is cheap — including its own shares and the broad equity market. That is a defensive stance dressed as patience. It will look prescient if the market corrects and Berkshire deploys; it will look like opportunity cost if the market keeps compounding and the cash keeps earning 4%.

The Kraft Heinz Writedown

Berkshire took a $3.76B after-tax impairment on its Kraft Heinz stake, writing the holding down to a fair value of $8,408M. It is a long-underwater position dating to the 2015 merger.

Assessment: The writedown changes nothing operationally — it is an accounting acknowledgment of a reality the market priced years ago — but it is a reminder that even Berkshire's "permanent" partnerships can go wrong, and that management will mark them honestly when they do. Reported through GAAP net earnings, it is non-operating; we exclude it from our operating read, but we note the candor.

Capital Allocation & The Cash Question

Everything in the Berkshire thesis at this moment routes through one question: what happens to the $344B? The cash is being generated faster than management is deploying it, and the three traditional outlets are all closed or constrained:

• Acquisitions: No elephant-sized deal has cleared Berkshire's price discipline in years. The universe of targets large enough to move the needle and cheap enough to clear the hurdle is small, and competition from private equity has compressed it further.
• Buybacks: Dormant at current valuation. Management's revealed preference is that ~1.5x book is not cheap enough.
• Equities: A net seller, not buyer — management sees the broad market as expensive.

That leaves T-bills as the residual, by default rather than by choice. The arithmetic is straightforward and slightly uncomfortable: a quarter of the company is invested at ~4–5% while the operating businesses underneath compound at materially higher returns. Every quarter the cash sits idle, blended return on equity is diluted.

Assessment: We are sympathetic to the discipline — refusing to overpay is why Berkshire exists — but discipline and paralysis look identical from the outside until capital moves. The bull case requires faith that the optionality pays off; the bear case requires only patience to watch the drag accumulate. Until we see the cash deployed — into a large acquisition, a re-engaged buyback, or opportunistic equity buying — we cannot underwrite the optionality at full value. This is the crux of our Hold.

Book Value & Valuation

Berkshire shareholders' equity stands at $667,989M. Against the post-print market capitalization of roughly $990B (at the $459.11 close), the stock trades at approximately 1.48x book value. The equity portfolio alone ($267,923M fair value, against a cost basis of $79,398M) carries ~$188B of unrealized gains.

For context: Berkshire spent most of the past decade between roughly 1.3x and 1.5x book. The stock's all-time-high close of $539.80 on May 2 — the session before the succession announcement — implied roughly 1.7x on the then-current book. The de-rating since is, in part, the market repricing the Buffett premium ahead of the handover.

Assessment: At ~1.48x book, Berkshire is neither cheap nor expensive — it is fairly valued for a business of this quality with this balance sheet. That is precisely the problem for a rating: fair value plus a capital-deployment question mark plus a leadership transition is a market-perform setup, not an outperform one. We would get materially more constructive approaching ~1.3x book (where buybacks historically re-engage and the margin of safety widens) or on concrete evidence the cash is moving.

Market Reaction

• Pre-print setup: BRK-B closed at $472.84 on Friday, August 1, up 4.3% year-to-date but trailing the S&P 500's +6.1%. The stock was already ~12% below its May 2 all-time high of $539.80, having de-rated steadily since the succession announcement. It entered the print down 2.5% over the trailing 30 days.
• Reaction session (Monday, August 4): Shares gapped down 1.0% and closed at $459.11, off 2.9% on the day (−$13.73), touching a 52-week intraday low of $455.19. Volume ran 10.9M shares versus a 4.6M 30-day average — 2.3x normal. The S&P 500 rose 1.5% the same session, so Berkshire underperformed the tape by more than four points.

The market read the print the way we do on the headline but not on the substance: it saw a "down" operating quarter and a Kraft Heinz writedown, sold the stock to a fresh 52-week low, and largely ignored the segment-level health underneath. The deeper driver is the one that has weighed on the stock since early May — the discounting of the Buffett premium ahead of the transition. A 2.9% drop on a day the market rose 1.5% is the market continuing to reprice leadership risk, not reacting to FX remeasurement.

Street Perspective

Debate: Is the $344B cash pile optionality or dead weight?

Bull view: The cash is unmatched dry powder — the capacity to act decisively in the next market dislocation when liquidity is scarce and prices are distressed, exactly the environment where Berkshire has historically minted its best returns. It also earns ~4–5% while it waits, so the "cost" of patience is low.

Bear view: A quarter of the market cap earning T-bill yields is a permanent drag on return on equity, and the pile keeps growing because management can't find anything to buy — including its own stock. "Optionality" is the word you use for idle cash when you don't want to call it idle cash.

Our take: Both are right, and the tiebreaker is deployment, which hasn't happened. Until it does, we side with the bear on the rating while respecting the bull on the long-run option value. The cash is worth more than 4% only if it gets spent well; we can't bank that yet.

Debate: How much of the stock is a "Buffett premium" at risk in the handover?

Bull view: The transition is the most telegraphed succession in corporate history. Greg Abel has run the non-insurance operations for years, the culture and decentralized model are institutionalized, and Buffett remains as chairman. There is no operational discontinuity — the premium that's at risk is sentiment, and sentiment mean-reverts.

Bear view: Buffett's capital-allocation judgment — the willingness to do nothing for years and then act enormously — is not transferable, and it is precisely the cash-deployment skill that the $344B most needs. The market is right to take some premium out until Abel demonstrates the same discipline and the same nerve.

Our take: The operating business will not skip a beat under Abel; on that the bulls are correct. But the single hardest, highest-stakes part of the job — deploying a fortress balance sheet at the right moment for the right price — is the part with no track record under the new regime. A modest "show me" discount is warranted, and the stock is paying it.

Debate: Is ~1.5x book the right multiple?

Bull view: For a business with a negative-cost ~$174B float, a $268B marketable portfolio, BNSF and BHE, and the best industrial collection in America, ~1.5x book is undemanding. The look-through earnings power justifies it and then some.

Bear view: ~1.5x is the high end of Berkshire's decade-long range, the cash drag is rising, and management's own refusal to buy back stock here says the multiple isn't cheap. Fair, at best.

Our take: We land with the bear on valuation and the bull on quality, which is the definition of a Hold. Great business, fair price, open capital-allocation question — we want a better entry.

Thesis Scorecard

Overall: Thesis is balanced — an exceptional business and balance sheet offset by a stalled capital-allocation cycle and a leadership-transition discount. Net unchanged at initiation.

Action: Initiate at Hold. Own the quality, respect the fortress, but wait for either a cheaper multiple (~1.3x book) or hard evidence the cash will move before underwriting outperformance.