Selling your data exhaust to AI labs is capital, not revenue: price it as a moat or sell it once

Selling data to the AI labs may be the most lucrative trade in business right now. That is the reason to be careful about building a company around it.

Yoni Rechtman at 99d argues that data sales are better seen as a source of capital than as a revenue line. Buyers have very large budgets today, but he points out that this may not last, and that what they want will probably change several times. The data a lab wants this year may be worth little next year, and a company that has bent its roadmap to one customer's appetite is exposed when the appetite moves.

Two of his points are worth holding on to. The first is the accounting. Selling data converts an asset that never sat on a balance sheet into cash, and he notes that accounting standards have not caught up. The second is that data does not run out when it is sold. Once a buyer has a copy, it keeps it. He asks when selling your exhaust means selling your moat, and answers that it depends on the business. His example is that a legal AI company should never sell its workflow data to the model makers, because that data is its source of margin.

For Australian organisations this is a live question, not a start-up curiosity. Telcos, retailers, insurers and industrials hold operational data that buyers will pay for. Some of it is incidental to how they win. Some of it is exactly why they are hard to copy.

My own view is that the test is simple to state and hard to apply honestly. If the data is a by-product of running the business, selling it once can fund the real strategy, provided you keep the strategy and do not start reshaping it around one buyer's needs. If the data is what makes you hard to copy, it should be priced as if it were the moat, because that is what the buyer is getting. In that case the answer may be no, however good the offer looks.

A buyer will eventually put the question to you, and it is better to have decided which kind of data you hold before that conversation starts.

Source: Yoni Rechtman, "Data beyond brokerage", 99d (9 October 2026), https://lnkd.in/gKSN9tCr

#DataStrategy #Strategy #AustralianBusiness