🔍 Read the full analysis: The Costs To Consider Before Switching From Claude on ThorstenMeyerAI.com
Get audio and creator 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
A report by The Information says Meta and Microsoft have reduced some employees’ use of Anthropic’s Claude tools while steering them toward alternatives they already operate or support. The reported moves are tied to costs and in-house tools, not a stated finding that Claude performs worse. For companies without ready substitutes, switching can bring engineering, evaluation, training and quality costs that may outweigh savings.
Meta and Microsoft are steering some employees away from Anthropic’s Claude and toward alternative coding tools, according to a report by The Information on Oct. 5. The reported shift reflects cost controls and the availability of tools the companies already operate or support; it is not a reported judgment that Claude performs worse, and it does not mean either company has stopped offering Claude to customers.
Meta reportedly reduced the number of employees using Claude Code from about 60,000 earlier this year to about 30,000. The report said staff were being directed toward Meta’s own tools: MetaCode, which has more than 30,000 internal users, and Muse Code, which has more than 6,000. The figures describe reported internal use, not customer adoption or a direct comparison of tool quality.
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 said Microsoft later cut that projection by more than a third and directed employees toward GitHub Copilot and OpenAI models. A separate detail in the source account said some monthly team budgets may have fallen from about $100,000 to about $10,000; that figure is attributed to a single report and has not been established as a company-wide policy.
The reported changes concern employee use and internal budgets. Microsoft is still reported to spend heavily on Anthropic models for customer-facing Copilot features, while customer spending on Claude through Microsoft platforms is said to be growing. The available information does not indicate that access to Claude has been ended.
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 the Model Bill
The reported moves show why a lower model price does not automatically make switching worthwhile. Companies moving real workloads must account for engineering time, evaluation, retraining and the risk of weaker results on their particular tasks. Those costs may not appear in a vendor’s price list, but they can affect output and budgets.
For Meta and Microsoft, the reported alternatives were already deployed: Meta has its own coding tools, while Microsoft has GitHub Copilot and access to OpenAI models. That gives them options most organizations do not have. The source account estimates Microsoft’s reported spending reduction could amount to more than $300 million annually if applied to a projection above $1 billion, but that is a calculation from the reported figures, not a confirmed realized saving.
For a smaller company, the balance could be different. A team spending $20,000 a month might find that rebuilding integrations and workflows costs more than the savings over a year. That is an illustrative possibility, not a measured result for a named company. The practical point is to compare total cost per accepted result, including review and rework, rather than tokens or seats alone.
As an affiliate, we earn on qualifying purchases.
Why These Buyers Have Alternatives
Meta and Microsoft are not typical buyers. Meta develops its own models and coding products. Microsoft owns GitHub Copilot and is a major backer of OpenAI. When companies that build or support competing products shift internal work to those products, the move can reflect vertical integration and budget priorities as well as the economics of a particular supplier.
The reported account says the companies cited rising token costs, tighter spending controls and in-house tools as drivers. It does not report that either company concluded Claude was inferior. That distinction matters: a purchasing decision by a company with ready-made substitutes is not, by itself, evidence that the same tool is a poor fit for other users.
Switching can require teams to rerun evaluations, adapt prompts and tool definitions, and rebuild connections between an agent, an editor, a code repository and local practices. Employees may need time to learn the replacement. For agent workloads, moving providers can also reset cached context or change cache pricing. If the alternative produces weaker results on a company’s tasks, additional review and rework can erase apparent savings. These are potential costs to measure, not confirmed expenses incurred by Meta or Microsoft.
What the Report Does Not Establish
The reported figures do not show how much work employees now complete with each tool, whether the reductions apply across all teams, or how the companies measured productivity and quality after the changes. It is also unclear how much of Microsoft’s revised projection reflects lower usage, tighter budgets, changed forecasts or other factors.
The source material does not identify public statements from Meta or Microsoft confirming every reported figure. Nor does it provide a like-for-like test of Claude against MetaCode, Muse Code, GitHub Copilot or OpenAI models. Quality comparisons remain unconfirmed, as do the switching costs each company incurred and any realized savings.
Measure Before Moving Workloads
For companies considering a change, the next step is to test alternatives on representative work before moving critical workflows. A useful comparison should record task success, human review time, rework, latency and total cost, and should keep the same evaluation criteria across providers. Teams without an established test set may need to build one before they can judge a switch reliably.
Organizations can also reduce the cost of a future change by keeping prompts, business rules and tool definitions in a layer they control, and by running a second model on a limited share of real work. That approach does not guarantee a cheaper or better result, but it creates evidence about trade-offs before a vendor or budget change becomes urgent. Further reporting or company statements may clarify the scope and outcomes of Meta’s and Microsoft’s moves.
Key Questions
Have Meta and Microsoft stopped using Claude?
No such company-wide end is reported. The report concerns reduced internal employee use and spending projections. Microsoft is still reported to use Anthropic models for customer-facing Copilot features.
Did the companies say Claude performed worse?
The source material reports cost controls and a move toward in-house or supported alternatives as the reasons. It does not report either company saying Claude performed worse, and it gives no comparative quality test.
What can make switching models expensive?
Potential costs include rerunning evaluations, adapting prompts and integrations, training employees, rebuilding cached context, and handling more review or rework if the replacement performs differently. The actual cost depends on the workflow and should be measured.
What should a company measure before switching?
Compare total cost per accepted result, not just subscription or token charges. Track task success, review time, rework, integration effort and the time employees need to adapt, using the same representative tasks for each model.
Source: ThorstenMeyerAI.com
Halloween Picks
halloween
As an affiliate, we earn on qualifying purchases.
