Generated by Codex with GPT-5

This summary covers The Economist’s May 9th, 2026 Business column on Berkshire Hathaway and SoftBank Group, published under the headline Warren or Masa and listed in the contents as Berkshire v SoftBank.

The article presents Berkshire Hathaway and SoftBank as opposite answers to the same question: what should an investment company do when markets look unusually expensive and technology is changing unusually fast? Berkshire, built by Warren Buffett, has accumulated nearly \$400bn in cash because it cannot find enough attractive deals. SoftBank, led by Masayoshi Son, is borrowing heavily to finance a sweeping bet on artificial intelligence. One risks becoming too cautious to justify its structure. The other risks running out of money before its vision pays off.

The comparison works because the companies are not conventional firms. Both mix operating businesses with large investment portfolios. Both were shaped by famous founders whose judgment became central to the brand. Both also attract unusually loyal shareholders. Yet their balance sheets express radically different temperaments.

Berkshire’s Problem Is Excessive Patience

Berkshire’s cash pile is now worth about 40% of its market value, roughly twice its average share over the past two decades. That reserve would look prescient after a major crash, when Berkshire could buy good assets cheaply. Without such a crash, it becomes harder to defend. The company has bought back almost none of its own shares in recent years, apparently judging them expensive, and it has not paid a dividend since 1967.

The article argues that Berkshire has become a more exaggerated version of itself. Buffett made his reputation by avoiding fashionable assets and buying durable businesses with defensible advantages at reasonable prices. That discipline served shareholders well during past bubbles. But it also leaves the company poorly positioned when a long bull market makes bargains scarce. Even Berkshire’s largest technology holding, Apple, is unusual among major tech companies because it is not spending all its cash flow on AI data centres.

This challenge now belongs to Greg Abel, Buffett’s successor as chief executive. Berkshire shares have underperformed the market sharply over the past year. If markets keep rising, Abel may eventually have to return more cash to investors or consider breaking up parts of the conglomerate. The article’s point is not that caution has suddenly become foolish. It is that patience stops looking like a strategy when a company cannot explain what it is waiting to buy.

SoftBank’s Problem Is Relentless Appetite

SoftBank sits at the other extreme. Son has made and lost enormous bets before. The firm’s investment in WeWork became a symbol of the easy-money startup boom, while its purchase of Arm, the British chip designer, proved spectacularly valuable. SoftBank bought Arm for \$31bn in 2016; its stake is now worth far more than that.

Son is again betting on a technological transformation, this time AI. By October, SoftBank is expected to have invested \$65bn in OpenAI, making it the ChatGPT maker’s second-largest external shareholder after Microsoft. It has also committed to spend \$3bn a year on OpenAI products and is buying robotics and data-centre assets. In Son’s view, the AI boom is not a moment for restraint.

The financing is the weakness. SoftBank’s operating businesses do not generate enough cash to cover its commitments. It has sold assets, but some of the holdings left in its portfolio are worth much less than they were a year ago. More debt is the likely answer. A \$40bn bridge loan used to invest in OpenAI matures next March, and the company has already borrowed against stakes in Arm and its Japanese telecoms business. The cost of insuring SoftBank’s debt against default has risen as lenders question how the bills will be paid.

Two Different Succession Tests

The article closes by arguing that only one of these approaches can be vindicated. If AI investment keeps producing extraordinary returns and markets remain elevated, Berkshire’s caution will look increasingly costly. If markets fall, Arm loses value or OpenAI fails to list its shares, SoftBank may struggle to refinance its obligations. Berkshire’s failure mode would be slow: an unwinding, a dividend or a breakup. SoftBank’s would be much more abrupt.

That makes both companies succession stories as well as investment stories. Abel inherits a company so strongly identified with Buffett’s restraint that changing course may be difficult even when restraint becomes excessive. Son remains in charge of SoftBank, but his company has little margin for error because its latest wager depends on continued access to financing.

The takeaway is broader than a contest between two famous investors. Markets need both skepticism and ambition. Buffett’s Berkshire and Son’s SoftBank show what happens when those virtues become concentrated until they start to resemble weaknesses. One has too much cash and too few ideas. The other has more ideas than its balance sheet can comfortably support.