Opinion: The Smart Money Is Buying on the Low Side

The most useful thing said in the markets this week came from a chief executive explaining why he is not trying to get every last dollar. Asked by Jamie Dimon why Anthropic is going public now, Dario Amodei said the company buys computing capacity on the low side of its forecasts and wants extra cash as a buffer. He would rather set modest expectations and beat them.

That is unusual coming from a company with an order book about six times oversubscribed. It is also good advice for nearly everyone else doing a deal right now.

Restraint as a strategy

Look at how the better transactions were built. Millicom agreed to buy Entel Peru for about five times earnings before interest, taxes, depreciation and amortization, and it will pay a quarter of the price a year and a half from now. That gives the buyer time to prove the business before the last check clears. UniCredit raised $3.5 billion in dollars to get a head start on next year, while buyers are still lining up. Rowan Digital Infrastructure arranged $1.5 billion of credit for land, power and equipment before it needs to break ground, rather than scrambling for money when a tenant calls.

None of that is glamorous. All of it reduces the chance of being forced to sell, borrow or build at the wrong moment.

Where the discipline is missing

The contrast came on the trading floor. Modine rose 28% in a day on reports of talks with Samsung that have produced no price, no terms and no confirmation. Traders have already paid a premium that the buyer has not offered. If Samsung walks, or bids less than the market expects, the stock has a long way to fall.

Private investors are also paying for momentum. LangChain doubled its valuation in a year to $3 billion. It has real customers and a strong position with developers. It also competes in a field that grows more crowded every month.

The lesson from Amodei is not that ambition is wrong. It is that leaving something on the table is a form of insurance. Next Wednesday will show whether investors reward him for it.