With wider access to AI, the market began to shift its focus from “use or not” to “how much has been spent and how much has been rewarded”. Chamath Palihapitiya, a technology investor, believes that some companies may soon feel pressure for such expenditure at the profit end.

Management or underestimation of internal consumption

In an interview with CNBC, Palihapitiya stated that many CEOs and CFOs may not know how much AI token was consumed within the company. He referred to “tokenmaxing” as an enterprise that encouraged its employees to use as many AI tools and model interfaces as possible to enhance development and office efficiency.

In his view, the problem might first be revealed in the financial statements. When the return per unit is lower than expected in a quarter, management returns to the cost and finds that AI mobilization costs have accumulated rapidly within the organization.

Enterprises are more focused on returns on inputs

According to Palihapitiya, an increasing number of investors and technology executives are reminding the market that the phase of the simple pursuit of AI may be ending. Businesses had previously viewed more AI use as a positive signal, but the more realistic question now was whether those expenditures actually translated into business returns.

He mentioned in March this year that his own start-up company, 8090, annual AI spending is rising to over $10 million. As a founder of a small start-up company, he described the trend as “very scary” and stated that many companies might be pushing up revenues in the AI industry but not getting a meaningful ROI.

8090 completed $135 million in financing

Palihapitiya founded 8090 in 2024. The company is developing a platform for users to work with AI agents to build enterprise software. In June this year, the company announced the completion of $135 million in financing, which was sponsored by Salesforce.

Apart from him, Alexander Karp, CEO of Palantir, has recently expressed similar views. He indicated to CNBC that there was a current tendency in the United States business community to invest time and money in large-scale token calls, rather than to validate actual business value.

As more technology companies enter the financial season, whether AI costs begin to erode profits may become the next focus of market attention.