🔍 Read the full analysis: Meta And Microsoft Pulled Back From Claude. What’s The Cost Of Following? on ThorstenMeyerAI.com
Get monitors, keyboards and dev gear delivered free — and shop member deals
- Fast, free delivery on millions of items
- Access to Prime Big Deal Days deals on October 6–7
- Prime Video, Amazon Music and more included
TL;DR
The Information reported on Oct. 5 that Meta and Microsoft have reduced some employees’ use of Anthropic’s Claude tools and directed work toward alternatives. The reported shifts concern internal use, not a complete break with Anthropic, and the companies’ stated reasons center on cost, spending controls and available substitutes rather than reported quality problems. Their ability to switch also reflects engineering resources most companies may not have.
Meta and Microsoft have been steering some employees away from Anthropic’s Claude tools and toward alternatives, according to a report by The Information on Oct. 5. The reported changes involve internal use and spending—not a full withdrawal from Anthropic—and show how large technology buyers can shift AI work when costs rise and substitutes are already available.
At Meta, use of Claude Code reportedly fell from about 60,000 employees to about 30,000 earlier this year. The company has directed engineers toward its own coding tools, with MetaCode reported to have more than 30,000 internal users and Muse Code more than 6,000. These figures describe reported internal adoption, not the number of employees who have stopped using Claude altogether.
Microsoft had reportedly projected annual internal spending of more than $1 billion on Anthropic technology, including Claude Code, Claude models used in Copilot and Claude Mythos. The Information reported that Microsoft cut that projection by more than a third and has steered employees toward GitHub Copilot and OpenAI models. The source material also says Microsoft continues to use Anthropic models in customer-facing Copilot features and that customer spending on Claude through Microsoft platforms is growing.
The reported explanations are rising token costs and tighter spending controls, alongside the availability of tools the companies own or support. The report does not establish that either company shifted internal work because Claude performed worse. Microsoft’s reported spending figure is a projection, not a confirmed final bill; a further account cited in the source material says some team budgets fell from around $100,000 to around $10,000 a month, but that detail is attributed to a single report.
Meta and Microsoft pulled back from Claude. Here’s what switching actually costs.
The Information reports both companies steering their own employees away from Claude. Read as a verdict on Claude, it misleads. Read as a demonstration of switching — and who can afford it — it’s the most useful enterprise-AI signal this month.
Staff steered to GitHub Copilot and OpenAI models; stricter token budgets. One unconfirmed report: some team budgets ~$100k → ~$10k/month.
Microsoft reportedly still spends heavily on Claude for customer-facing Copilot — and that spending is reported to be growing.
Reported drivers: rising token costs and owned alternatives. Neither company is reported to have called Claude worse.
Meta builds coding tools; Microsoft owns Copilot and backs OpenAI. This is ordinary vertical integration.
Keep a second vendor live on real work.
A few hundred tasks with pass criteria.
Logic, prompts, tools in your layer.
Tokens are the cheap half.
Know what you’d rebuild.
On the evidence reported, Meta and Microsoft didn’t reject Claude. They brought spending in-house where they could and kept buying where they couldn’t — Microsoft remains a large Anthropic customer for the products it sells. The signal is the mechanism: the most sophisticated buyers treat models as interchangeable suppliers behind a layer they control.Meta could halve its Claude usage because it had built somewhere else to go. Build somewhere else to go.
Switching Costs Shape AI Buying
The report matters because it puts attention on a cost that can sit beyond model prices: the work required to move an organization’s workflows. Meta and Microsoft have alternatives already deployed. Their internal tools, staff and existing arrangements with other AI providers give them options that many buyers lack.
For other companies, a change in model can mean retesting workflows, adapting prompts and tool integrations, and retraining staff. A switch can also disrupt accumulated context or caching arrangements. If the replacement produces weaker results on a company’s own tasks, additional human review and rework may erase some of the savings. Those costs are not established in the report as specific losses at Meta or Microsoft; they are practical risks for other organizations to measure.
The figures also show why scale changes the calculation. If Microsoft’s reported annual projection exceeded $1 billion, a reduction of more than a third would imply a very large potential saving against that projection. But that is not proof of realized savings, and it should not be treated as a directly comparable estimate for a smaller buyer. For a company spending tens of thousands of dollars a month, the engineering and productivity costs of switching could outweigh near-term price reductions.
A practical implication is to make alternatives usable before they are urgently needed. Testing more than one model family, maintaining representative evaluation tasks and keeping business logic separate from a particular provider can make a future switch easier. That does not establish that a multi-provider setup is right for every company; it highlights the value of knowing how different options perform on the work that matters.
As an affiliate, we earn on qualifying purchases.
Two Buyers With Their Own Options
Meta and Microsoft are not neutral buyers in this market. Meta develops its own models and coding tools, while Microsoft owns GitHub Copilot and is OpenAI’s largest backer, according to the supplied source material. Their reported decisions sit within broader efforts to use products they control or have a stake in, as well as to manage spending.
That context limits what can be concluded from the employee-use figures. A reduction in internal use is not the same as ending a supplier relationship, and Microsoft’s reported continuing use of Anthropic models for customer-facing Copilot features points to a distinction between how a company equips its own staff and what it offers customers. The report, as described in the source material, does not say either firm has stopped providing customers access to Claude through its platforms.
The source material also refers to recent findings from SemiAnalysis about AI subscription limits and the value of subscriptions changing. That is separate reporting and does not explain Meta’s or Microsoft’s decisions. Taken together, the developments are reminders that AI service costs and terms can change, but the available information does not establish a single market-wide cause or outcome.
As an affiliate, we earn on qualifying purchases.
What the Report Does Not Settle
The Information’s account is the basis for the reported employee counts and spending shift; the supplied material does not include direct statements from Meta, Microsoft or Anthropic confirming each figure. It is not clear how the companies counted active users, how long the changes took, or whether the numbers reflect regular use, access or another measure.
It also remains unclear how much either company has actually saved, whether internal output or quality changed after the shifts, and how the reported decisions affect Anthropic’s revenue. Microsoft’s figure is described as a spending projection, while the separate monthly team-budget detail comes from one account. No specific quality comparison or measured productivity result is provided in the source material.
As an affiliate, we earn on qualifying purchases.
Watch Spending and Usage
The next useful indicators are whether the reported internal usage patterns persist, whether Microsoft revises its spending plans again, and whether either company provides more detail about how it evaluates its models and tools. Customer-facing use is a separate measure from employee adoption and will need to be tracked separately as new reporting emerges.
For companies weighing similar changes, the immediate question is not simply which model has the lower listed price. They will need to compare performance on representative tasks, engineering and migration effort, review time, and total cost. Until Meta, Microsoft or Anthropic provides more detail, the reported shifts are evidence of internal reallocation, not a settled verdict on Claude’s quality or the economics of switching for every buyer.
AI development tools for businesses
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
Did Meta and Microsoft stop using Claude?
No full stop is reported. The account describes reduced or redirected internal use. Microsoft reportedly continues using Anthropic models in customer-facing Copilot features, and customer use through Microsoft platforms is said to be growing.
Why are the companies steering employees toward alternatives?
The reported drivers are rising token costs, tighter spending controls and existing alternatives. The supplied material does not report that either company said Claude performed worse.
How much did Microsoft reportedly cut its Anthropic spending projection?
The Information reportedly said Microsoft cut a projection of more than $1 billion a year in internal spending by more than a third. This is a revised projection, not a confirmed amount saved or a final spending total.
What could make switching AI models expensive?
Organizations may need to retest workflows, adapt prompts and integrations, retrain users, and account for changes in review and rework. The actual cost depends on the company’s systems and tasks; the report does not provide a measured switching bill for Meta or Microsoft.
Source: ThorstenMeyerAI.com
Halloween Picks
halloween
As an affiliate, we earn on qualifying purchases.
