🔍 Read the full analysis: What A Move Away From Claude Could Cost AI Teams on ThorstenMeyerAI.com
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
A report by The Information says Meta and Microsoft have reduced or plan to reduce some employees’ use of Anthropic’s Claude tools, directing them toward alternatives they own or already use. The reported shifts are about internal use, costs and competing tools—not evidence that Claude performed worse. Smaller teams may face substantial migration work that the two tech companies could absorb with existing substitutes.
Meta and Microsoft have reportedly steered employees away from some Anthropic tools and toward alternatives, according to The Information on 5 October. The reported changes concern internal use, with cost controls and in-house or affiliated products cited as drivers; they do not establish that Claude performed worse or that the companies have ended access to it.
Meta reportedly 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 tools, including MetaCode, which the source says has passed 30,000 internal users, and Muse Code, which has passed 6,000. The report does not give a full account of how usage was measured or when each figure was recorded.
Microsoft had reportedly projected more than $1 billion a year in internal spending on Anthropic technology, including Claude Code, Claude models used in Copilot and Claude Mythos. The report says Microsoft has cut that projection by more than a third and is steering employees toward GitHub Copilot and OpenAI models. The source also says Microsoft continues to use Anthropic models for customer-facing Copilot features and that customer spending on Claude through Microsoft platforms is growing.
The reported explanation is rising token costs, tighter spending controls and the availability of competing tools, rather than a stated finding about model quality. Microsoft was also reported to have tightened token budgets; one account cited monthly team budgets falling from about $100,000 to $10,000. That figure comes from a single report, and further details about which teams it covers are not provided.
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 Extend Beyond Model Fees
The report highlights that choosing an AI model is not just a comparison of per-token prices. Moving work can require teams to re-run evaluations, adjust prompts and tool definitions, adapt integrations, and help employees rebuild working habits. A model that costs less to run may still increase total costs if it leads to more review, rework or errors on a company’s actual tasks.
Those expenses may be easier for Meta and Microsoft to absorb because they already have alternatives in use and engineering teams capable of supporting them. The source estimates that a cut of more than a third from Microsoft’s projected $1 billion-plus in annual internal spending could amount to more than $300 million a year. That is an implication of the reported projection and reduction, not a confirmed realized saving.
For smaller AI buyers, the balance could be different. A company spending $20,000 a month on a provider might not save enough in fees to cover migration and quality-checking work within a year. The source offers that as an illustrative estimate, not a measured result for a named company. It points to the practical value of keeping a second provider workable, so a buyer can respond to price or policy changes without starting an integration from scratch.
As an affiliate, we earn on qualifying purchases.
Two Buyers With Their Own Alternatives
Meta and Microsoft are not ordinary customers in this story. Meta develops its own models and coding products; Microsoft owns GitHub Copilot and is a major backer of OpenAI. Both have commercial interests in alternatives to Anthropic’s products. That makes employee use of their own or affiliated tools a different situation from an independent company switching vendors.
The report, as summarized in the source material, describes changes to employee usage and projected internal spending. It does not say either company has shut off Claude for all staff or withdrawn Claude-powered features from customers. The internal shifts therefore should not be read as a broad customer exit or a public verdict on Claude’s technical performance.
Companies can also face changing subscription limits, token budgets and provider prices. The source cites a separate SemiAnalysis finding that AI subscription limits can change without clear notice and that list-price reductions can affect subscription value. Those observations are background, not direct evidence about the terms of Meta’s or Microsoft’s Anthropic arrangements.
As an affiliate, we earn on qualifying purchases.
Scale and Savings Remain Unverified
The available account does not provide the companies’ full statements, detailed spending records or consistent measurement periods for the reported usage figures. It is also unclear how much of the projected reduction in Microsoft’s Anthropic spending will become an actual saving, and whether the change affects all teams or only selected internal workloads.
The reported shifts do not establish whether Claude’s quality was compared with alternatives on the same tasks. Nor do they show how migration costs, employee productivity, review time or error rates changed after the tools were switched. The reported budget reductions should not be treated as proof of a general cost-saving result across companies.
As an affiliate, we earn on qualifying purchases.
Measure Before Changing Providers
The next useful evidence would be clearer reporting from Meta and Microsoft on the scope and timing of their internal changes, alongside actual spending rather than projections. For AI teams weighing a move, the practical next step is to test alternatives on representative real tasks, track accepted outputs and review effort, and include integration and training costs in the comparison.
Teams can reduce the work of a future switch by maintaining an evaluation set, keeping business logic and prompts in a layer they control, and running a second model on a limited share of production work. Those measures do not guarantee a cheaper or better replacement; they make it possible to compare providers with evidence before committing to a change.
As an affiliate, we earn on qualifying purchases.
Key Questions
Are Meta and Microsoft ending their use of Claude?
The report describes reductions or changes in internal use, not a complete shutdown. The source says Microsoft continues to use Anthropic models in customer-facing Copilot features.
Did the companies say Claude performed worse?
No such finding is reported in the source material. The stated reasons are cost controls and available alternatives; the report does not establish that either company judged Claude’s quality to be lower.
Why might switching cost more than the new model’s price?
A switch can involve re-testing workflows, changing prompts and integrations, training staff and checking whether output quality has changed. Review and rework can also offset savings from lower token prices.
Are Microsoft’s reported savings confirmed?
No. The report concerns a reduction of more than a third in a projected annual internal spending figure. The source does not confirm the final spending or realized savings.
What can a smaller AI team do before changing providers?
Maintain a small set of representative tasks for evaluation, keep prompts and business logic under the company’s control, and test a second provider on real work. That can expose migration and quality costs before a full switch.
Source: ThorstenMeyerAI.com
Halloween Picks
halloween
As an affiliate, we earn on qualifying purchases.
