Abel's First Print: Why Berkshire's Buyback Says More Than Its AI Restraint
Q1 operating earnings up 18%. Cash at a record $397 billion. The first buyback in two years. The post-Buffett discount is real and, on the May data, mispriced.
The most consequential thing Greg Abel said in Omaha on May 2 was not about artificial intelligence, even though that is the line every wire picked up. It was a sentence he buried in the same breath: capital discipline, he said, is “the rule,” and he intends to act “decisively upon when there is a dislocation, and the price is right.” That is the framing under which his first quarter as Berkshire Hathaway’s chief executive needs to be read. Operating earnings of $11.35 billion, up roughly 18% year-on-year. Cash and Treasury bills at a record $397 billion. A $234 million buyback in March, the first share repurchase by Berkshire since the first half of 2024. The market reaction has been notably underwhelming. The market is, we think, missing the trade.
Two narratives have travelled in parallel since Warren Buffett stepped aside as chief executive on January 1. The first is the obvious one: the post-Buffett discount — a perceived diminution of the moat — would weigh on the stock for years as the new regime found its footing. The second is the more specific concern that Abel, an operator by background rather than an investor, would either lean too cautiously on the cash pile or, in the worst case, swing at a poor target to prove himself. The Q1 print and the annual meeting do not resolve those questions outright, but they significantly tilt the evidence in one direction. Abel is signalling exactly the discipline Buffett spent fifty years modelling, with one additional twist: a willingness to deploy buybacks when the math justifies them, and to refuse to deploy capital when it does not.
The signal in the buyback
The $234 million Berkshire spent on its own shares in March is a modest number against a $397 billion cash position. It is also the first repurchase the company has executed in almost two years, and that is the part that matters. Berkshire’s buyback policy is not a financial-engineering tool. It is a price signal. Repurchases are permitted only when Abel and Buffett judge the stock to be trading below conservatively-estimated intrinsic value, with the company also required to retain at least $30 billion in cash. The decision to resume buying in March, after the stock’s lacklustre 12-month performance, is the new chief executive’s first public assertion that he disagrees with the market’s implied valuation of his own franchise. Coming from a CEO whose tenure is being graded primarily on capital-allocation judgment, that is a non-trivial statement.
Buffett, who remains chairman, was characteristically blunt about the broader environment in his remarks alongside Abel.
“The surrounding environment is not ideal for deploying cash for Berkshire.” — Warren Buffett, Berkshire Hathaway annual meeting, May 2, 2026 (CNBC live coverage)
Read against the cash pile, that one sentence does most of the explanatory work for the period since 2023. Berkshire has been net-selling equities, sitting on bills earning roughly 4.3%, and waiting. The strategic question for shareholders has never been whether Buffett and now Abel can find good companies. It has been whether they can find good companies at acceptable prices. The buyback resumption tells us Abel’s answer for the time being is: Berkshire itself qualifies; the broader market mostly does not.
Abel’s technology framing
The line that made the headlines was Abel’s rejection of AI-for-its-own-sake, and it deserves to be unpacked properly rather than reduced to a soundbite.
“We’re going to be a builder of technology, rather than just a buyer of technology. It has to be additive to our businesses. We’re not going to have AI just to have AI.” — Greg Abel, Berkshire Hathaway annual meeting, May 2, 2026
The framing here is significant for two reasons. First, by positioning Berkshire as a builder rather than a buyer of technology, Abel is implicitly ruling out the kind of multi-billion-dollar hyperscaler-adjacent equity purchases the market has occasionally speculated about and that competing public conglomerates have made. Second, by tying technology deployment explicitly to the operating businesses — BNSF in particular, where Abel discussed AI-driven dispatch and asset-utilisation work — he is reasserting Berkshire’s historical edge as a producer of operating cash flow rather than as a holder of marked-to-market technology positions. In an equity market where the dominant trade has been to buy AI-infrastructure beta, that is a deliberately countercyclical posture.
Ajit Jain, the vice chairman of insurance operations, was even more direct on where AI is and is not, in a comment that has been under-quoted in the post-meeting coverage.
“If AI materializes as anticipated, it will undoubtedly be transformative. We are not close to a point where AI can make decisions on pricing or claims.” — Ajit Jain, Berkshire Hathaway annual meeting, May 2, 2026 (CNBC)
Jain’s caveat is the kind of operator-level distinction that public-market commentary on AI tends to flatten. The most lucrative parts of Berkshire’s insurance business — the underwriting decisions on long-tail commercial risk, the claim adjudications on complex losses — remain irreducibly judgment-heavy. Productivity tooling around routine workflows is real and will compound; full automation of pricing is not on the medium-term horizon. That distinction is exactly the kind of thing a Berkshire-style operator would prioritise. It is also the kind of distinction the implied market valuation of pure-play AI is currently struggling to make.
What the operating numbers actually say
Strip away the AI commentary and the Q1 print is a quietly excellent number. Operating earnings of $11.35 billion, up about 18% from $9.64 billion a year earlier, were carried by an insurance underwriting result that benefited from a benign catastrophe quarter and improving Geico fundamentals, and by BNSF, which has been showing better operating-ratio performance under Katie Farmer. Revenue and headline EPS missed Street expectations on accounting items that have very little to do with the underlying earnings power of the business. The cash balance climbed to $397 billion from $373 billion at year-end, an extension of the trend that has defined the post-2022 period: operating businesses throwing off cash faster than Berkshire can find acceptable places to deploy it.
Greggory Warren, the senior Morningstar analyst who has covered Berkshire for two decades, has been consistent in his framing through the transition: Abel remains Morningstar’s “top pick for the capital allocating chief role,” with the two co-CIOs Todd Combs and Ted Weschler handling the marketable-securities sleeve under him, and Jain on insurance. The credibility of that team structure is what allows Berkshire to hold $397 billion in cash without the market treating it as an admission of defeat. After the May 2 meeting and the Q1 print, the case for that framing is meaningfully stronger than it was on January 1.
Our view
The post-Buffett discount is real but, on the current evidence, mispriced. The market is treating the first quarter under Abel as a low-information print and is implicitly applying a moat-erosion haircut to a franchise that, on the data we have, is behaving exactly as it should: insurance underwriting strong, BNSF improving, the conglomerate intact, capital discipline asserted at the stock level via buybacks, and AI exposure being built through operating productivity rather than purchased through public-equity beta. That is the bull case in three sentences, and Q1 supports it.
Three implications. First, the resumption of buybacks is the floor signal the equity has been missing. Expect repurchase activity to accelerate if the stock continues to trade at the current implied price-to-intrinsic-value discount; Abel has every incentive to use the cash pile this way at these levels because it doubles as a discipline statement and a return generator. Second, the Q1 mix — insurance plus rail carrying the print — is exactly the mix that compounds quietly through a year in which the broader market is wrestling with AI-capex digestion and the front end of the curve. We would size the position for total-return more than for narrative. Third, the right way to read the AI commentary is not as anti-technology but as a public commitment by management to be paid for risk in this segment rather than to pay up for it. In a year in which the dominant equity trade is the opposite, that is a structural alpha source for patient capital. The discount will close. It just will not close on a single quarterly print.
This note is for information and discussion only and does not constitute investment advice or an offer to buy or sell any security. Quotes are sourced from public statements; positions and views are the author’s and may change without notice.
