🔍 Read the full analysis: What Businesses Should Know About The Cost Of Switching From Claude on ThorstenMeyerAI.com
Get everyday essentials 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
A report by The Information says Meta and Microsoft have reduced some employees’ use of Anthropic’s Claude tools while directing them to alternatives they already operate or support. The reported shift is about internal spending and available substitutes, not a confirmed judgment that Claude performs worse. For most businesses, changing models can involve substantial costs beyond the price of tokens, including evaluations, integration, retraining and review.
Meta and Microsoft have reportedly reduced some employees’ use of Anthropic’s Claude tools and steered them toward alternatives they already operate or support, according to a report by The Information on Oct. 5. The development offers a case study in the cost of changing AI providers: large companies can shift work when they have substitutes ready, while other businesses may face engineering, evaluation and productivity costs that do not appear in a model’s listed price.
The Information reported that Meta cut Claude Code users from about 60,000 earlier this year to about 30,000. The account says Meta has directed employees toward its own coding tools, MetaCode, which it reportedly has more than 30,000 internal users, and Muse Code, which has more than 6,000. These are reported internal-use figures, not a count of customers or public users.
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 report says the company subsequently reduced that projection by more than a third and 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 reasons for the internal shift are cost, tighter spending controls and in-house alternatives, not a stated finding that Claude performs worse. The companies are also positioned to favor products they build or back: Meta develops its own AI tools, while Microsoft owns GitHub Copilot and is a major backer of OpenAI. The reported moves do not amount to a general end to Claude access or a public withdrawal from Anthropic.
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 Beyond Model Prices
For businesses, the central issue is not simply whether one model charges less per token. A switch can require teams to retest workflows, revise prompts and rebuild integrations. Coding assistants, for example, are often fitted to a company’s editor, repositories and working practices. A replacement may need time and engineering effort before it reaches comparable usefulness.
There can also be costs after deployment. Employees must learn a different tool, and managers need to check whether changed outputs increase review time, rework or errors. A lower API bill does not necessarily mean a lower cost per accepted result. The right comparison depends on the task and on the quality of the work that reaches production.
The report’s account of Microsoft’s projected spending puts the possible savings at a large-company scale: a reduction of more than a third from a projection above $1 billion a year would imply a substantial amount. That is not a confirmed realized saving. For smaller buyers, the economics could look different: a company spending $20,000 a month may spend more on switching than it saves, depending on how much adaptation its workflows require.
As an affiliate, we earn on qualifying purchases.
Why Big Buyers Had Alternatives
Meta and Microsoft are not typical buyers choosing among vendors from scratch. According to the report, both had alternatives already in use: Meta’s internal coding tools, and Microsoft’s GitHub Copilot and access to OpenAI models. That existing capacity makes it more practical to redirect employee work than it would be for a company whose applications and staff depend on one provider.
The report also describes tighter token controls at Microsoft. One account put some monthly team budgets at about $10,000, down from around $100,000. That figure comes from a single report and should not be treated as a company-wide policy or independently established average. It illustrates how spending limits, alongside model choice, can shape internal use.
These developments concern the companies’ own employees and budgets. They should not be read as evidence that external customers have stopped using Claude or that Anthropic’s models have failed a performance test. Internal purchasing decisions and customer-facing product choices are distinct, and the reported continued use of Claude in Microsoft products underscores that distinction.
“Meta reduced the number of employees using Claude Code from about 60,000 earlier this year to about 30,000.”
— The Information
enterprise AI integration software
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Usage Figures and Savings Remain Fluid
The reported figures do not establish how much Meta or Microsoft ultimately spent, what savings they realized, or whether internal usage will rise or fall again. The Microsoft spending figure was a projection, and the reduction was reported as more than a third; the material does not provide a final annual spend. The reported team-budget example is based on one account.
It is also unclear how the companies measured productivity and output quality across tools, or how much work shifted to each alternative. Neither company is reported to have said Claude performed worse. No detailed public comparison of model performance, switching costs or results was included in the source material.
For other businesses, the cost of changing providers depends on their workflows, existing integrations and ability to measure results. There is no universal switching-cost figure, and the reported experience of two technology companies cannot by itself predict the outcome for a smaller buyer.
As an affiliate, we earn on qualifying purchases.
Measure Before Changing Providers
Businesses weighing a change should first establish how their current tools perform on representative work. That means tracking cost per accepted result, not only token use, and recording review effort, rework and quality alongside direct charges. A repeatable evaluation set can make comparisons more useful than informal employee impressions.
Companies can also limit the disruption of a future switch by testing a second model on real tasks and keeping prompts, tool definitions and business logic in systems they control. Such steps do not guarantee that a replacement will be cheaper or better, but they can reveal integration and training needs before a large-scale change.
The next clear evidence will be updated usage, spending or product disclosures from Meta, Microsoft or Anthropic, if provided. Until then, the reported moves show that major buyers are redirecting some internal work, but do not establish a broad retreat from Claude or settle whether switching would pay off for other organizations.
As an affiliate, we earn on qualifying purchases.
Key Questions
Have Meta and Microsoft stopped using Claude?
No such company-wide stop is established. The report concerns reduced or redirected internal use. The source material says Microsoft continues to use Anthropic models in customer-facing Copilot features.
Why did the companies reportedly shift some employees away from Claude?
The reported drivers are rising costs, tighter spending controls and available in-house or supported alternatives. Neither company is reported to have said Claude performed worse.
What costs can a business face when switching AI models?
Potential costs include retesting workflows, adapting prompts and integrations, training staff, and handling changes in review time, rework or output quality. Cache behavior and pricing may also affect some agent workloads.
Does switching to a cheaper model guarantee lower overall costs?
No. A lower token price may be offset by engineering work, productivity losses or additional review. Businesses need to compare total cost and quality on their own tasks.
What should a company do before considering a switch?
Build a set of representative tasks with clear pass criteria, then compare providers on cost, accepted output, review effort and reliability. Testing a second provider on a limited share of real work can expose migration needs early.
Source: ThorstenMeyerAI.com
Fall Picks
fall essentials
As an affiliate, we earn on qualifying purchases.
