🔍 Read the full analysis: Meta And Microsoft’s Claude Pullback Puts AI Switching Costs In Focus on ThorstenMeyerAI.com
Get smart everyday buys 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 roughly halved employee use of Claude Code and Microsoft reduced its projected internal Anthropic spending by more than a third. The reported shifts reflect cost controls and available alternatives, not a stated finding that Claude performs worse; both companies reportedly continue to use Anthropic technology in some customer-facing products.
Meta and Microsoft have reportedly reduced or redirected some employee use of Anthropic’s Claude tools, steering staff toward alternatives they own or back, according to a report by The Information on Oct. 5. The reported changes point to cost controls and available substitutes, not a public finding by either company that Claude is inferior, and they highlight the practical costs of switching AI systems.
At Meta, the number of employees using Claude Code reportedly fell from about 60,000 earlier this year to about 30,000. The company has been directing staff toward its own coding products: MetaCode, which the source report says has passed 30,000 internal users, and Muse Code, with more than 6,000. These figures describe internal use, not customer adoption.
Microsoft had reportedly projected more than $1 billion a year in internal Anthropic spending, covering Claude Code, Claude models in Copilot and Claude Mythos. The Information reported that Microsoft has cut that projection by more than a third and is steering employees toward GitHub Copilot and OpenAI models. The source material also says Microsoft continues to use Anthropic models for some customer-facing Copilot features, and that customer spending on Claude through its platforms is growing.
The reported explanations include rising token costs, tighter spending controls and companies’ preference for tools they own or are invested in. The material does not report either company saying Claude performed worse. One reported account says some Microsoft team budgets fell from roughly $100,000 a month to about $10,000; that figure comes from a single account and should not be treated as a company-wide policy.
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.
Why Internal AI Spending Is Shifting
The changes matter because they show that large buyers can respond to AI costs by shifting work among available suppliers, rather than relying on one model or negotiating with a single vendor. Meta and Microsoft have alternatives already in place: Meta’s own coding tools, and Microsoft’s GitHub Copilot alongside OpenAI models. Their choices are therefore evidence of internal substitution at large companies, not proof that every organization can switch as easily.
For other businesses, a lower price per token does not automatically mean a cheaper or better result. A switch can require new evaluations, prompt and tool adjustments, staff training and extra review if output quality changes. Companies need to compare total cost per accepted result, including engineering time and rework, rather than token charges alone. The report does not quantify those costs for either company.
The episode also illustrates a difference between having the option to switch and actually switching. Meta and Microsoft have the resources and existing products to move work internally. Many organizations do not. They may face a lengthy migration or lose productivity while employees adapt, so the leverage enjoyed by the largest buyers may not transfer to smaller customers.
As an affiliate, we earn on qualifying purchases.
Alternatives Shape the Claude Pullback
The reported developments concern the companies’ own employee use and spending forecasts. They do not establish that Meta or Microsoft has ended access to Claude, nor that either company has stopped offering Anthropic models to customers. That distinction matters because internal tool choices are not the same as customer product decisions.
Both companies have commercial reasons to develop or favor alternatives. Meta builds its own AI models and coding tools. Microsoft owns GitHub Copilot and has a major investment in OpenAI. Using an internal or affiliated product can reduce dependence on an outside supplier, but the available material does not establish how much each factor contributed to the reported changes.
The report frames the moves against growing attention to AI spending and usage limits. Its figures are snapshots of reported internal use and projections, rather than independently audited measures of current costs or productivity. The reported spending projection is not the same as confirmed annual expenditure.
“Meta’s employee use of Claude Code reportedly declined from about 60,000 to about 30,000.”
— The Information, in reporting dated Oct. 5
As an affiliate, we earn on qualifying purchases.
What the Report Does Not Establish
The reporting does not provide a detailed breakdown of which workflows moved, how much work remains on Claude, or how the companies measured the effects on software quality and employee productivity. It is also unclear whether the reported usage figures cover the same time periods and definitions at each company.
Neither company is reported as saying the changes were driven by Claude’s performance. The precise role of token prices, budget limits and internal product strategy is not fully established. Microsoft’s projected spending reduction also should not be read as a confirmed reduction in total Anthropic-related spending, particularly as the report says the company continues to use Anthropic models in customer-facing services.
The source material’s estimate that a cut of more than a third from a $1 billion-plus projection would exceed $300 million a year is an arithmetic implication of the reported figures, not a confirmed realized saving. Actual costs, savings and migration expenses remain unclear.
As an affiliate, we earn on qualifying purchases.
Tracking Usage, Costs and Results
The next useful evidence will be whether Meta and Microsoft disclose more about employee adoption, spending and performance across their available tools. Until then, the reported numbers offer a limited view of internal decisions, with no public accounting of migration costs or productivity outcomes.
For organizations making their own AI choices, the practical next step is to test alternatives on representative work before a major migration. Maintaining more than one model option, keeping business logic and tool definitions portable, and using a consistent evaluation set can make a future change less disruptive. Those steps do not guarantee savings or equivalent results; they make the trade-offs easier to measure before a decision is made.
As an affiliate, we earn on qualifying purchases.
Key Questions
Have Meta and Microsoft stopped using Claude?
The reported changes concern some internal employee use and projected spending. The source material says Microsoft continues to use Anthropic models for some customer-facing Copilot features. It does not say either company has ended all access to Claude.
Why are the companies reportedly shifting employees to other tools?
The reported reasons include rising token costs, tighter spending controls and the availability of tools the companies own or back. Neither company is reported to have said Claude performed worse.
How much did Microsoft reduce its projected spending?
The Information reportedly said Microsoft cut a projection of more than $1 billion a year in internal Anthropic spending by more than a third. That is a change to a forecast, not confirmation of actual spending or realized savings.
Can other companies switch AI models as easily?
Not necessarily. Switching may involve retesting workflows, adapting prompts and integrations, training staff and reviewing output quality. Meta and Microsoft already have alternatives deployed, while many organizations may need to build those capabilities first.
Source: ThorstenMeyerAI.com
Halloween Picks
halloween
As an affiliate, we earn on qualifying purchases.
