🔍 Read the full analysis: The Business Costs Of Switching AI Models After The Claude Pullback 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 internal use of Anthropic’s Claude tools and steering employees toward alternatives they own or back. The reported shift reflects cost controls and available substitutes, not a reported finding that Claude performs worse. For other companies, switching can cost far more than a change in model fees because it requires engineering work, evaluation and staff retraining.
Meta and Microsoft are directing some employees away from Anthropic’s Claude tools and toward alternatives, according to an Oct. 5 report by The Information, in a shift that highlights the often-unpriced costs of changing AI systems. The reported moves concern the companies’ internal use and were driven by spending controls and in-house options; the report does not say either company found Claude inferior.
Meta reportedly reduced the number of employees using Claude Code from about 60,000 earlier this year to about 30,000. The report says the company has steered engineers toward its internal coding tools: MetaCode, which has passed 30,000 internal users, and Muse Code, with more than 6,000. These figures describe reported internal adoption, not customer use.
Microsoft had reportedly projected annual internal spending of more than $1 billion on Anthropic technology, including Claude Code, Claude models in Copilot and Claude Mythos. It has since cut that projection by more than a third and is directing employees toward GitHub Copilot and OpenAI models, according to The Information. The account also says Microsoft continues to spend on Anthropic models for customer-facing Copilot features, while customer spending on Claude through Microsoft platforms is growing.
A separate detail in the report says some Microsoft team budgets fell from around $100,000 to around $10,000 per month. That figure comes from a single account and should not be treated as a company-wide budget change. The reporting does not indicate that either company has ended access to Claude.
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 Fees
The reported decisions show how a buyer with ready alternatives can respond to rising costs by shifting work between providers. They do not show that switching is simple for most companies. Meta and Microsoft have internal tools, competing products and large engineering teams that can support migration. A typical organization may not have those resources or substitutes already in use.
For other buyers, changing models can mean rerunning evaluations, adapting prompts and tool connections, and rebuilding integrations with editors, code repositories and team processes. Employees may also need time to learn a new system. These costs can reduce output during the change, even when the new model’s listed price is lower.
There are costs that are harder to see on a bill. Moving providers can reset cached context used in repeated agent work, changing both performance and cost. If the replacement performs less well on a company’s specific tasks, additional review, rework or mistakes may follow. Those effects depend on the work and the models involved; the report provides no direct measurement of them for Meta or Microsoft.
The financial case therefore depends on more than token prices. The report’s figures imply that a reduction of more than a third from a projected annual spend above $1 billion would represent a substantial potential saving, but it is a change to a projection, not a confirmed realized saving. At smaller budgets, migration work and productivity losses could outweigh savings for a period. Each company would need to measure its own costs and results.
As an affiliate, we earn on qualifying purchases.
Two Buyers With Their Own Alternatives
Meta and Microsoft are not neutral buyers in this reported shift. Meta develops its own models and coding tools. Microsoft owns GitHub Copilot and is a major backer of OpenAI. Steering employees toward products a company owns or supports can reflect strategic and financial incentives alongside supplier costs.
The report is about employee usage and internal spending plans. It does not establish a broad retreat from Anthropic across either company’s customer products. Microsoft’s reported continued use of Anthropic models in customer-facing Copilot features makes that distinction especially relevant. Nor does a change in internal adoption, by itself, establish a comparative judgment about model quality.
“Meta reduced the number of employees using Claude Code from about 60,000 earlier this year to about 30,000.”
— The Information, reporting on Meta
As an affiliate, we earn on qualifying purchases.
Quality And Savings Remain Unmeasured
The reported reasons for the moves are cost, spending controls and internal alternatives. The account does not report that Meta or Microsoft found Claude less capable, and it does not provide comparative test results for the models on employee tasks.
It is not clear how much Anthropic technology the two companies will continue using internally, whether the reported adoption figures will change further, or how much of Microsoft’s revised projection translates into actual savings. The report also does not quantify migration costs, productivity effects or changes in output quality. Those outcomes may differ across teams and workloads.
AI model training and retraining software
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Track Usage, Spend And Work Quality
The next useful evidence will be whether the reported reductions persist and how each company balances internal alternatives with Anthropic models in its products. For Microsoft, actual spending and continued use in customer-facing Copilot features will help clarify the scope of the internal shift. For Meta, later figures on employee adoption of its coding tools would show whether the reported migration continues.
For companies reviewing their own AI budgets, the practical next step is to measure the full cost per accepted result, including tokens, review time and rework. Maintaining a second model option on real tasks, keeping representative evaluations and placing prompts and tool definitions in a reusable layer can make future changes easier to assess. Whether any switch saves money still depends on the organization’s workload, integration costs and results.
As an affiliate, we earn on qualifying purchases.
Key Questions
Are Meta and Microsoft ending their use of Claude?
The report describes reductions or revised internal plans, not a complete end to access. It says Microsoft continues to use Anthropic models in customer-facing Copilot features.
Did the companies say Claude performed worse?
The source material does not report that either company judged Claude to be lower quality. It attributes the reported moves to costs, spending controls and available internal alternatives.
How much could Microsoft save?
The report says Microsoft cut a projected annual internal spending figure of more than $1 billion by over a third. That is a revised projection, not a confirmed tally of realized savings.
Why can switching models cost more than the subscription or token change?
Teams may need to recheck performance, adapt prompts and integrations, retrain users and account for review or rework. The size of those costs depends on the organization and its tasks.
What can a company do before it needs to switch?
It can test more than one model on representative work, keep its own evaluation tasks and make prompts and tool connections reusable. These steps help a company measure a change before moving more work.
Source: ThorstenMeyerAI.com
Halloween Picks
halloween
As an affiliate, we earn on qualifying purchases.
