Unhappy that there’s now a limit on how many AI credits they can burn on Microsoft’s GitHub Copilot, developers are threatening to stop eating at the all-you-can-eat AI buffet. But they could just think more carefully about how to use their allowances.

“This is a staggering shift from a ‘predictable subscription’ to a ‘stressful meter-based’ service that hinders my productivity rather than helping it,” wrote one developer on GitHub's user forum.

They said they were paying $39 per month for the Copilot Pro+ plan, but burned through around 8% of their monthly AI Credits allocation in two hours under the new system, which went into effect on Monday.

What’s the actual cost-benefit ratio?

Already in April, when Microsoft announced the change, the tech giant said it was moving from monthly billing to usage-based billing because GitHub Copilot is “not the same product it was a year ago.”

“It has evolved from an in-editor assistant into an agentic platform capable of running long, multi-step coding sessions, using the latest models, and iterating across entire repositories,” said the company.

“Agentic usage is becoming the default, and it brings significantly higher compute and inference demands.”

This isn’t exactly surprising. Until recently, AI companies have been happy to shoulder the costs of providing relatively cheap access to their models, and hundreds of millions of users flocked into their ecosystems.

Have thoughts about this topic? Others do, too. Join them in the discussion.

Soon, though, after seeing that software engineers – also encouraged by the executives – like to run multiple AI coding agents at the same time, AI firms saw all this as simply getting too expensive and unsustainable – and had to act.

Anthropic has also repeatedly tinkered with Claude's code rate limits, and in March, began imposing tighter limits during peak hours.

Businesses, already racking up huge AI costs, are beginning to see that these changes might start killing them.

According to Axios, one company was charged $500 million for one month of Anthropic’s Claude after the firm failed to set usage limits.

Seeing that their employees now simply default to automating tasks they dislike instead of doing something truly valuable, companies are naturally questioning the cost-benefit ratio of using AI.

AI is becoming too expensive

Still, individual devs are furious. One wrote: “Please take this as a formal notice: unless GitHub reverts to a more sustainable billing model or significantly increases the credit allocation for Pro+ users to match professional workflows, I will be canceling my subscription immediately.”

“Woke up to the new billing UI this morning. Figured I’d test it out on some actual work – just needed Claude 4.8 to help fix a couple things on a site I'm editing,” a Reddit user posted.

“It gave some pretty mediocre suggestions. Didn’t really solve the problem – I still had to do most of the work myself. Then I checked the actual usage page, 1,180 credits used. That’s 16% of my monthly Pro+ allowance. Gone. For basically nothing.”

Andrew Orlowski, a tech journalist at The Daily Telegraph, chose to ask this in his weekly column on Monday: “When AI is more expensive than people, why replace the people?”

Ironically, not that long ago, CEOs around Silicon Valley were actually encouraging employees to burn as many tokens as possible. Now, though, they are finding it hard to justify the cost of AI initiatives because the output isn’t meeting expectations.

Uber president and COO Andrew Macdonald recently told the Rapid Response podcast that the company cannot yet draw a clear connection between rising Claude Code token consumption and the additional useful consumer features it has shipped.

No wonder Andrew Orlowski, a tech journalist at The Daily Telegraph, chose to ask this in his weekly column on Monday: “When AI is more expensive than people, why replace the people?”

---

Unlock more exclusive Cybernews content on YouTube.