GitHub Copilot pricing change highlights AI profitability pressures

Microsoft has transitioned GitHub Copilot from a flat-rate subscription model to a token-based pricing structure. This shift marks a significant change from the tool’s original pricing and has sparked widespread discussion across the technology sector regarding the sustainability of current artificial intelligence cost structures. By moving away from flat-rate pricing, Microsoft is charging based on tokens, highlighting the immense operational expenses associated with running AI models.

Observers are describing this transition as the beginning of a “Tokenpocalypse”—a colloquial term for the crisis of high token costs. The AI ecosystem has been heavily subsidized by investor money, making current pricing structures difficult to maintain. For example, ChatGPT Plus originally launched at $20 a month, a flat rate that did not reflect true operational costs. As companies like Uber have demonstrated, scaling often requires aggressive cost-cutting. Uber recently experienced this dynamic internally, blowing through its AI budget faster than expected and subsequently placing caps on usage within the company. AI firms will likely face similar pressures, making price increases and usage restrictions likely as they attempt to control costs. This tension raises fundamental questions about whether AI labs can reduce costs quickly enough to meet customer spending limits.

The regulatory and financial landscape is also shifting rapidly. This week, President Trump signed an executive order to review powerful AI models. At the same time, AI companies like Anthropic are planning to go public. These firms face the challenge of documenting their evolving operational risks in an S-1—an IPO registration statement filed with the SEC. This is particularly difficult given how quickly the market has shifted; for instance, the industry trend of “tokenmaxxing” peaked and fell out of favor within a six-month window due to high costs. Highlighting the difficulty of drafting these risk disclosures, podcast host Kirsten Korosec asked, “How do you even write these risks in, because they are evolving before our eyes?”

Why it matters

Microsoft’s pivot to token-based pricing underscores the fragility of current AI business models, as companies must reconcile high operational costs with market demand before they can achieve profitability.