TechCrunch: Token-Based Payments Reshape AI Market

GitHub Copilot's token-based payment model may prompt AI market pricing changes.

The shift of GitHub Copilot to token-based payments could signal a reevaluation of pricing and limits in the AI product market. This conclusion was drawn by TechCrunch following the Equity podcast.

The discussion was prompted by changes in GitHub Copilot’s pricing. Microsoft transitioned some users from a fixed subscription to a model based on the number of “building blocks” of processed information.

A Reddit user described the situation as a “token-pocalypse” amid the sharp rise in AI tool expenses.

In the podcast, TechCrunch editors Anthony Ha, Kirsten Korosec, and Sean O’Kane linked the potential price reevaluation to the high cost of developing and maintaining AI models. Ha noted that a significant portion of the market is still subsidized by investor capital, meaning some costs may eventually shift to end users. He assessed that companies like Anthropic, as they prepare for an IPO, will need to better explain to investors their path to reducing costs and achieving profitability.

O’Kane cited Uber as an example. He mentioned that the company quickly exhausted its budget for AI tools and subsequently began restricting internal use. He doubted that model developers could reduce technology costs quickly enough to align with customers’ willingness to pay.

Korosec added that pricing approaches in the AI sector are changing too rapidly. In her view, the “token-maxing” trend became noticeable within a few months and has irritated the market.

She also pointed to increased regulation in the US: On June 2, 2026, President Donald Trump signed an order allowing authorities to inspect powerful AI models before public release, as mentioned in the discussion.

In May, Gartner analysts forecasted AI spending to grow to $2.59 trillion in 2026.