🔍 Read the full analysis: The Hidden Work Behind Switching From Claude on ThorstenMeyerAI.com
Get tech for your team 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 are directing some employees toward their own or other AI coding tools, reducing projected internal use of Anthropic technology. The reported shift reflects cost controls and available substitutes, not a stated finding that Claude performs worse. For most companies, switching also means work on evaluations, integrations, staff habits and quality checks that does not appear on a model’s price list.
Meta and Microsoft are 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 concern the companies’ internal use and projected spending; they do not establish that Claude performed worse, or that either company is ending access to Anthropic technology.
The Information reported that Meta reduced the number of employees using Claude Code from about 60,000 earlier this year to about 30,000. The report said Meta is directing staff toward its own coding tools: MetaCode, which has more than 30,000 internal users, and Muse Code, with more than 6,000. The figures are reported user counts, not independent measures of tool performance or productivity.
Microsoft had reportedly projected more than $1 billion a year in internal spending on Anthropic technology, including Claude Code, Claude models in Copilot and Claude Mythos. The Information reported that Microsoft cut that projection by more than a third and is steering employees toward GitHub Copilot and OpenAI models. A separate detail in the source account, attributed to a single report, says some monthly team budgets fell from about $100,000 to about $10,000; the scope and applicability of that figure are unclear.
The reported decisions are framed around rising token costs, tighter spending controls and in-house alternatives. Neither company is reported to have said Claude produced inferior results. The report also distinguishes employee use from customer-facing services: Microsoft reportedly continues to use Anthropic models in features for Copilot customers, while customer spending on Claude through Microsoft platforms is said to be growing.
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.
The Engineering Cost of Switching Models
The reported shifts show that large buyers can redirect AI work when they have credible alternatives already in place. But a lower model bill is only one part of the calculation. Companies also need to account for the effort needed to test, adapt and support a different system across real workflows.
That work can include rerunning evaluations on representative tasks; adjusting prompts, tool definitions and agent software; rebuilding integrations with editors and code repositories; and allowing employees time to learn a different tool. Switching can also affect cache behavior and pricing for systems that repeatedly process the same context. These are potential costs described in the source material, not amounts independently confirmed for Meta or Microsoft.
Quality changes may carry a less visible cost. If a replacement handles a company’s tasks less reliably, teams may spend more time reviewing results, fixing work or catching errors. That does not mean a new model will be worse: it means buyers need task-specific testing before they treat token savings as net savings. Meta and Microsoft’s reported scale and existing tools may make the trade-off different from that facing a smaller company.
As an affiliate, we earn on qualifying purchases.
Why Meta and Microsoft Have Alternatives
These are not straightforward examples of ordinary customers abandoning a supplier. Meta develops its own models and coding products; Microsoft owns GitHub Copilot and has a major investment in OpenAI. Both companies have business reasons and technical capacity to promote competing tools among their own staff. Their internal product choices do not by themselves show how outside customers assess Claude.
The reported figures also describe different things. Meta’s count concerns employees using Claude Code and users of its alternatives. Microsoft’s figure is a spending projection, not a confirmed annual bill, and the reported reduction is a change to that projection. The available account does not give a complete breakdown of actual spending, how many workflows moved, or how long the changes took.
That distinction matters when reading headlines that cast the development as a verdict on Claude. The reported evidence supports a narrower conclusion: two companies with substantial internal alternatives are changing how some employees use AI tools, with cost control among the stated drivers. It does not establish that other customers are leaving Anthropic or that Claude has lost access to Microsoft’s customer-facing products.
enterprise AI model evaluation software
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
What the Report Does Not Establish
The source account cites The Information but provides no direct statements from Meta, Microsoft or Anthropic. The companies’ explanations, the precise timing of each change and the methodology behind the user and spending figures are not available here. The extent to which the figures reflect active use, access or projected budgets is also unclear.
It is likewise unknown how much work has actually moved between models, whether the alternatives match Claude’s performance on the tasks involved, and whether the reported budget reductions produced net savings after engineering and review costs. Microsoft’s continued use of Anthropic models for customer-facing Copilot features is reported, but the scale and future direction of that use are not specified. The claims should be treated as reported developments, not as a complete account of either company’s AI strategy.
As an affiliate, we earn on qualifying purchases.
How Buyers Can Prepare to Switch
The next useful evidence would be direct confirmation from the companies about the scope of the internal changes, the spending figures and whether more workflows are being moved. For buyers, the practical next step is to test alternatives on their own representative tasks rather than assume that a model’s published price predicts the cost of switching.
That means maintaining evaluation sets with clear success criteria, tracking review and rework alongside token spending, and keeping prompts and tool definitions in a layer the company controls. Running more than one model family on real work—even at limited volume—can reveal integration needs before a supplier change becomes urgent. The report does not say every business should leave Claude; it highlights that switching is easier when the groundwork already exists.
AI productivity evaluation software
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
Are Meta and Microsoft ending their use of Claude?
The report describes reduced or redirected internal use, not a complete end to access. Microsoft is also reported to keep using Anthropic models in customer-facing Copilot features.
Why are the companies reportedly changing employee use?
The reported drivers are cost controls, rising token costs and available alternatives. Neither company is reported to have said Claude performed worse.
Does this show that Claude is losing customers?
Not on the information provided. The reported changes concern Meta’s and Microsoft’s own employees and spending plans; they do not establish a broader customer exodus.
What makes switching AI models costly?
Teams may need to retest workflows, adapt prompts and integrations, train employees on a new tool, and review output quality. Those costs can sit alongside token charges rather than appear on the model’s price list.
Source: ThorstenMeyerAI.com
Halloween Picks
halloween
As an affiliate, we earn on qualifying purchases.
