Berkshire Hathaway Inc. BRK-B

Key data

The business

Berkshire is three things bolted together. First, businesses it owns outright: the BNSF railroad, Berkshire Hathaway Energy, GEICO and the reinsurance units, and a wall of manufacturing and retail names from See's Candies to Precision Castparts. Second, a stock portfolio worth more than $300 billion, led by Apple, American Express and Coca-Cola. Third, the largest cash reserve any company holds. The engine underneath is insurance float: money paid in by policyholders and held before claims come due, invested for Berkshire's own account at no cost as long as underwriting roughly breaks even. That float is the moat, cheap leverage no rival can copy. Sales and service from the operating businesses now make up 57 of every 100 dollars of revenue.

The numbers

Revenue climbed on the operating side while insurance sat still.

Second-quarter revenue rose 10%, carried by sales and service up 15%, while insurance premiums earned barely moved, up 1.3%, the mark of a soft market. The net earnings column is mostly noise: first-half 2026 profit of $35.8 billion includes $11.4 billion of after-tax investment gains, and first-half 2025's $17.0 billion carried a small investment loss and a $3.76 billion Kraft Heinz writedown. Strip both out and operating earnings went from about $20.8 billion to about $24.3 billion, up 17%. That is the figure that matters.

The 2022 loss was not a bad year of operations, it was the stock portfolio marked down. Read the annual net line and you learn almost nothing about the business, which is the whole reason book value, not earnings, is Berkshire's scoreboard.

Underwriting slipped. All-in costs ran about 90% of premiums against 89% a year earlier, and the pretax underwriting gain fell about 14% to roughly $2.2 billion. Interest and dividend income dropped 2.4% to $5.9 billion even with a larger cash hoard, because T-bill rates rolled lower and the float's recent tailwind is fading.

The scoreboard is book value per share. At 1.5 times book the market pays about $1.09 trillion for roughly $727 billion of net worth, or about $337 a share in book. Return on equity runs about 12%, so book compounds around 10% to 12% before anyone does anything clever. At 1.5 times, the top of its twelve-year range of 1.3 to 1.5, you pay full price for that compounding, and Berkshire agrees with you: buybacks fell to $235 million in the first quarter of 2026, down sharply from $2.9 billion for full-year 2024 and $27 billion in 2021. What this memo believes and the tape does not is that the price fairly captures the quality but pays nothing for the cash as a coiled option, and the print that settles it is a large acquisition or a resumed buyback.

Management

Greg Abel became chief executive on January 1, 2026, with Buffett staying on as chairman. The record to read is capital allocation, and it is speaking. Buybacks stopped at 1.5 times book, which is discipline, not weakness, because when the one buyer who knows the intrinsic value best will not buy the stock, that is the market's answer on price. No elephant acquisition has landed despite the large cash position sitting idle at T-bill yields. The insider signal is small and mixed: Ajit Jain, the vice chair over insurance, sold $7.5 million of stock in September 2025, plan status not disclosed, while director Michael O'Sullivan bought about $431,000 across two 2026 purchases. Pay is nominal by design. The real test for Abel is whether that cash gets deployed or quietly compounds at bond rates.

How it fails or surprises you

Cash drag. The large cash position earns about 4% in T-bills while equities compound faster, so every quarter Abel does not deploy it, book grows at bond rates. Buybacks have already halted. Watch for a resumed buyback or an acquisition above $20 billion. Absent both, this is a bond fund priced at 1.5 times book.

Underwriting turn or mega-catastrophe. Premiums are flat, up 1.3%, in a softening market, and a single hurricane or quake outside the models could wipe out a year of underwriting profit. All-in costs already crept toward 90% of premiums. Watch next quarter's underwriting result and catastrophe losses.

The cash as a loaded gun (right tail). A market crash hands Berkshire the one thing no rival holds, a large cash reserve to buy distressed businesses and securities at Buffett prices. The market pays nothing for that option today. Watch for a dislocation met by Berkshire spending $50 billion or more in a single quarter.

Closing thoughts

The quality is mostly in the price already. At 1.5 times book, the top of its range and above the level where Berkshire will buy its own shares, the durable engine of float, fortress balance sheet and diversified earnings is recognized by everyone. Nothing in the next four quarters settles the real question, which is whether the cash gets put to work at a discount before it drags. The fatter tail is the upside one, but it needs a market break you cannot schedule. The left-tail risk here is not permanent loss, it is years of book compounding at bond rates while you wait, having paid a full price. What is at risk if the cash sits is opportunity, and what it is worth if the dislocation comes is a decade of returns booked in a single year.

The bet is still that Berkshire keeps growing its net worth faster than the market, now under Abel, with the insurers underwriting at a profit and the cash spent well. It breaks if underwriting turns loss-making while the cash stays parked. The one pair to watch is the underwriting result against premium growth, and the dollars of cash deployed each quarter. If book value per share stops compounding above about 10% for two years running with the cash still idle, the read is wrong.