Particle.news
Get it on Google Play
Download on the App Store
5 ARTICLES
·
2w ago

Microsoft and Meta Scale Back Employee Use of Anthropic's Claude

Rising token bills plus data‑leakage risks have led both companies to steer engineers to internal coding assistants.

Image
Microsoft and Meta have each reduced their internal use of Anthropic’s Claude by encouraging employees to use their company’s own, in-house artificial intelligence tools, The Information reported Monday (Oct. 5), citing unnamed sources. Microsoft was on track to spend $1 billion this year on internal use of Anthropic’s technology, but it has since lowered that figure by one-third. The Anthropic tools include Claude Code, Claude models in Microsoft’s Copilot and Claude Mythos, according to the report. A Microsoft spokesperson said in the report that the company has been steering staff to use its GitHub Copilot coding tool, though engineers can still choose to use other models, per the report. Meta has reduced the number of its employees using Anthropic’s Claude Code from about 60,000 earlier this year to 30,000 recently, largely by pushing them to use its Claude Code competitor, Muse Code, and its in-house-only tool MetaCode, according to the report. Meta had been on track to spend billions of dollars this year on internal use of AI tools, with Claude Code being the most popular tool, per the report. Meta did not comment in The Information’s report, and the company did not immediately reply to PYMNTS’ request for comment. The Wall Street Journal reported Monday that companies are finding it increasingly difficult to forecast AI spending. The report cited one recent study that showed that just 11% of almost 400 surveyed businesses could accurately project their AI costs. The report said that AI differs from standard software in that it “behaves more like a human worker,” taking actions and making decisions and mistakes while it’s on the clock. The Ramp AI Index released Aug. 12 found that the top 1% of U.S. businesses spend a median of $7,400 per employee on AI in July. The median company spent just $11.95 per employee, meaning the heaviest spenders have AI budgets more than 600 times larger, per employee, than a typical business and roughly 11 times larger than the next tier of aggressive adopters. The report said Anthropic captured the biggest share of that money, as the company leads overall business adoption, with 43.5% of U.S. businesses paying for its subscriptions or tokens as of July.
Meta Platforms logo (public domain) via Wikimedia Commons
Microsoft logo (public domain) via Wikimedia Commons

Overview

  • Reports published Monday said Microsoft cut its internal spending on Anthropic’s models by more than one‑third and has been redirecting engineers to GitHub Copilot.
  • The same reporting found Meta has reduced Claude Code users from about 60,000 earlier this year to roughly 30,000 and is pushing its Muse Code and internal MetaCode tools.
  • Both pullbacks follow huge, hard‑to‑forecast token use that drove costs up—Microsoft saw per‑engineer token bills of about $500–$2,000 per month and Meta reportedly consumed 60.2 trillion tokens in a 30‑day span.
  • Anthropic remains tightly linked to Microsoft through $31.4 billion in non‑cancelable infrastructure commitments disclosed in its September IPO filing, so the cutbacks do not amount to a full breakup.
  • The moves signal a broader shift: companies are clipping external model use to cap volatile AI bills, protect sensitive training data from model distillation, and accelerate their own in‑house tools, which could squeeze Anthropic’s near‑term enterprise growth.