Microsoft has joined the growing list of companies cracking down on ‘tokenmaxxing’

Microsoft has imposed limits on staff AI use in a bid to cut ‘tokenmaxxing’ practices among software engineers.

First reported by 404 Media, an internal memo warned staff “need to be aware of how [they] consume tokens” on platforms such as GitHub.

“Tokenmaxxing is now what we are optimizing for. I want all of us focused on maximizing outcomes that move the needle for customers and our business,” wrote EVP Jay Parikh.

“As such, we are updating our internal guidance and managing token spend with the same discipline we apply to every other critical resource."

Under the changes, cheaper models such as OpenAI’s GPT-5.6 will be the default option for engineers moving forward.

GitHub introduced a new usage-based billing scheme for Copilot in April this year due to rising compute and inference costs.

At the time, executives said the move aimed to accommodate the increased use of agents, which are far pricier than reasoning-based models.

Microsoft is still an ‘AI-first’ company

Parikh noted in the memo that the changes don’t represent a change in Microsoft’s approach to internal AI use, adding that the firm is still “AI-first”.

Employees do have certain guidelines to follow, however. An internal Copilot usage guide cited in the memo requires departments to operate with budget targets. Staff also have tools at their disposal to track token usage.

According to 404 Media, the guidelines state: “While there is no target spend value being shared at this time, the data shows that many engineers spend in the range of hundreds of dollars a month to a few thousand dollars in tokens,.”

When asked by ITPro whether 404 Media’s report is true, Microsoft declined to comment.

Tackling rising AI costs

If the reports are correct, Microsoft is the latest in a string of companies to impose AI usage limits amid rising costs. The ‘tokenmaxxing’ trend has taken the industry by storm over the last 12 months, with organizations pushing hard to ramp up AI use.

Companies such as Meta introduced internal leader boards highlighting power users while others have incentivized staff to increase their use of the technology.

The trend has caused serious issues, however, with businesses facing spiraling costs.

Uber, for example, blew through its entire annual AI budget in a matter of months due to staff accelerating their use of AI. Reports from Bloomberg in June revealed the company introduced a $1,500 monthly cap per employee.

In June, Accenture told staff to cut back their use of AI for basic tasks due to what the company described as “soaring token spend”.

“What we’re seeing right now is just rapid escalation in AI token spend,” Justice Kwak, Accenture’s agentic AI strategy lead, reportedly said in a leaked meeting transcript.

Microsoft’s internal limits come as the company actively pushes new internal AI models designed to reduce costs.

As reported by ITPro, CEO Satya Nadella suggested in July that customers should leverage its more economical MAI model range rather than costly frontier models.

FOLLOW US ON SOCIAL MEDIA

Follow ITPro on Google News and add us as a preferred source to keep tabs on all our latest news, analysis, views, and reviews.

You can also follow ITPro on LinkedIn, X, Facebook, and BlueSky.