OpenAI takes a step back

In a week when OpenAI announced yet another billion-dollar funding round—and we have honestly lost count—it seems that ChatGPT is putting the brakes on one of its key monetisation initiatives.

According to reports, OpenAI has put its in-chat payment project on hold for the time being. The initiative was designed to allow users to complete transactions without ever leaving ChatGPT.

The project was intended to compete with the likes of Amazon and Google Shopping by creating a conversational ecommerce experience. In other words, OpenAI is stepping back from one of its key monetisation ideas.

The official explanation points to limited user adoption and the challenges of managing inventory, fraud, taxes and other operational complexities. But it is difficult to believe that these are the real obstacles for a technology giant like OpenAI.

We suspect the real reasons may lie elsewhere and could come down to two very different factors:

A) The first is that OpenAI is unlikely to position itself as a direct competitor to Amazon, particularly given that Jeff Bezos himself is among the investors backing the US company’s technology.

B) The second possibility is a disconnect between the idea and the commercial reality. If OpenAI starts directing sales through a single-answer model before establishing a clear advertising framework, transactions are likely to be distributed very unevenly. That, in turn, could lead brands and retailers to question whether it makes sense to provide their content and participate in the ecosystem if they receive little or no value in return.

In other words, OpenAI needs to prove that it can become a commercially viable business. To maintain investor confidence and continue attracting capital, it must demonstrate that it has real, scalable revenue models—not just technological potential.

But OpenAI needs to take things one step at a time. Until it clearly defines how and why brands appear in its answers, why some are prioritised over others, and who has access to all the data generated, it is difficult to see how it can move on to controlling the transaction itself.

On top of all this, ChatGPT is now approaching one billion users, yet it still accounts for only around 5% of global searches—despite the widespread perception that it has already become mainstream. At the same time, its agreements with the US government, particularly in the defence sector, appear to be triggering a growing backlash among a significant share of users.

Over the past two weeks, the number of users uninstalling the app has surged by 295%, compared with the usual rate of around 9%. Many of those users appear to be moving to CLAUDE, developed by Anthropic, which has publicly opposed the Trump administration’s approach to the use of AI.

AI companies ultimately need to prove that they can build solid, sustainable business models—not ones driven solely by short-lived expectations of future growth. At the same time, they must address ethics openly and transparently, particularly when it comes to how training data is sourced and used, and how the technology’s processing capabilities are deployed.