🔍 Read the full analysis: Is Moving Away From Claude Expensive? What Meta And Microsoft’s Shift Shows on ThorstenMeyerAI.com
Get the latest gadgets 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 reducing some employees’ use of Anthropic’s Claude tools while directing them to alternatives they already own or use. The reported shift concerns internal use, not an end to Claude access or a verdict on its quality; it shows how existing substitutes can make switching more practical for large companies.
Meta and Microsoft are steering some employees away from Anthropic’s Claude coding tools and models, according to a report by The Information on Oct. 5, shifting internal work toward alternatives the companies already use or own. The reported moves show how large buyers can reduce reliance on one AI provider when they have substitutes ready, but they do not establish that Claude performs worse or that either company is ending its broader relationship with Anthropic.
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 company has directed staff toward its own coding tools: MetaCode, which the source material says has more than 30,000 internal users, and Muse Code, with more than 6,000. Those figures describe reported internal adoption, not independently verified measures of productivity or tool quality.
Microsoft reportedly projected that its internal spending on Anthropic technology would exceed $1 billion a year, covering Claude Code, Claude models in Copilot and Claude Mythos. The report says Microsoft later cut that projection by more than a third and steered employees toward GitHub Copilot and OpenAI models. The source material also says Microsoft continues to spend on Anthropic models for customer-facing Copilot features and that customer spending on Claude through Microsoft platforms is growing.
The reported explanations center on token costs, spending controls and in-house alternatives. Microsoft also reportedly introduced tighter token budgets. One account cited in the source material says some monthly team budgets fell from about $100,000 to about $10,000; that detail is based on a single report and should not be treated as a company-wide policy. Neither company is reported to have said Claude underperformed.
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 Existing Alternatives Change the Math
The business lesson is not simply that companies can change AI vendors. It is that switching is cheaper when a replacement is already deployed. Meta and Microsoft have internal tools and competing products in place, giving them somewhere to redirect work. Most organizations do not have comparable engineering teams or ready-made substitutes, so the same change may involve more than redirecting a software subscription.
A company changing models may need to repeat evaluations, revise prompts and tool integrations, and retrain staff. It may also have to rebuild workflow-specific connections and account for differences in caching, review time and rework. A replacement that costs less per token can still be more expensive if it takes longer to produce an acceptable result or requires more human checking.
The reported Microsoft reduction illustrates the potential scale for a very large buyer: a cut of more than a third from a projected annual spend above $1 billion implies a large possible reduction in that projection. It is not a confirmed saving, however; the report does not establish actual final expenditure or realized savings. Smaller buyers should not assume they would see comparable financial gains, particularly if they must build the replacement workflow themselves.
As an affiliate, we earn on qualifying purchases.
The Report Covers Internal Use
The distinction between employee use and customer-facing products is central to interpreting the report. The reported changes concern how Meta and Microsoft direct their own workers. The source material says Microsoft continues to use Anthropic models in customer-facing Copilot features, and that customer use of Claude via Microsoft platforms is growing. That means a reduction in internal use does not by itself show that Microsoft is withdrawing Claude from its products or that customers are losing access.
Both companies also have strategic reasons to favor alternatives. Meta develops its own models and coding tools; Microsoft owns GitHub Copilot and is a major backer of OpenAI. That does not prove cost was the sole motive, but it means the reported shifts involve companies with existing products and investments that can compete with a supplier’s offerings. The evidence supplied does not establish a broad rejection of Anthropic by enterprise customers.
For other buyers, the practical distinction is between a model that can technically be swapped through an interface and a workflow that can be moved without disrupting work. Maintaining a second provider on real tasks, keeping an internal evaluation set and separating business logic from vendor-specific settings can reduce that gap. These are suggested practices, not steps the report says Meta or Microsoft followed in full.
As an affiliate, we earn on qualifying purchases.
What the Spending Figures Cannot Show
The figures and decisions are based on reported internal activity; the source material does not include direct statements from Meta or Microsoft confirming every detail. It is not clear how the employee counts were measured, what period the Microsoft spending revision covers, or how much of the projected reduction will translate into actual savings.
The reporting also does not provide comparative performance results for Claude and the alternatives on the companies’ internal tasks. It does not quantify the engineering, retraining, evaluation or productivity costs of moving work. Those costs vary by workflow, and no general estimate can be drawn from the reported figures alone. It remains unclear how much Anthropic technology will continue to be used internally, beyond the reported changes, or whether the companies plan further reductions.
As an affiliate, we earn on qualifying purchases.
How Buyers Can Test a Switch
For organizations weighing a move, the next useful step is to compare models on representative work rather than infer performance from the choices of much larger companies. That means tracking cost per accepted result, including review and rework, alongside token spending. A controlled trial can also show whether existing prompts, integrations and staff practices carry over or need substantial changes.
Any further developments in Meta’s and Microsoft’s internal deployments may clarify how much work is moving and whether the reported spending changes become realized savings. Until then, the clearest conclusion is limited: both companies reportedly have alternatives and are using them to reduce some internal reliance on Anthropic technology. The broader customer impact and comparative model performance are not established by the report.
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 Claude access. The source material says Microsoft continues to use Anthropic models in customer-facing Copilot features.
Does the reported shift mean Claude performed worse?
No performance finding is reported. The stated drivers are cost, spending controls and available alternatives; the report does not provide comparative test results.
How much is Microsoft expected to save?
The report says Microsoft cut a projection of more than $1 billion in annual internal spending by more than a third. It does not confirm final spending or realized savings, so the change should not be described as a verified saving.
Why could switching cost more for a smaller company?
Smaller buyers may lack ready-made alternatives and have to spend time on testing, integration changes, staff adjustment and additional review. The report does not quantify these costs or show that every company will face them equally.
Source: ThorstenMeyerAI.com
Halloween Picks
halloween
As an affiliate, we earn on qualifying purchases.
