🔍 Read the full analysis: Claude, Meta, And Microsoft: A Closer Look At The Cost Of Switching on ThorstenMeyerAI.com
Get business pricing on office and shipping supplies
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
The Information reported on Oct. 5 that Meta and Microsoft are steering some employees toward their own or other AI tools as they manage costs. The reported changes concern internal use, not an end to Claude access or a broad rejection of Anthropic’s products. The companies’ existing alternatives also mean their switching experience may not reflect the costs most businesses would face.
Meta and Microsoft are directing some employees away from Anthropic’s Claude tools and toward alternatives, according to a report by The Information published Oct. 5. The reported changes reflect efforts to control internal spending and use tools the companies already operate or back; they do not show that either company has ended access to Claude or judged it inferior.
At Meta, the number of employees using Claude Code reportedly fell from about 60,000 earlier this year to about 30,000. The company has been steering staff toward its own coding products: MetaCode, which the source says has more than 30,000 internal users, and Muse Code, with more than 6,000. These figures describe reported employee use; they do not establish how much work has moved or how the tools compare on performance.
Microsoft had reportedly projected annual internal spending of more than $1 billion on Anthropic technology, including Claude Code, Claude models used in Copilot and Claude Mythos. The projection was later cut by more than a third, with employees directed toward GitHub Copilot and OpenAI models, according to the report. Microsoft is also reported to continue spending on Anthropic models for customer-facing Copilot features, while customer spending on Claude through Microsoft platforms is said to be growing.
The account also describes tighter controls on internal AI use. One reported detail, attributed in the source material to a single account, is that some Microsoft teams’ monthly budgets fell from about $100,000 to $10,000. The reporting does not provide a full breakdown of the affected teams, the period covered by that budget change or the companies’ total spending across all AI providers.
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 Switching Takes More Than a New Model
The reported shifts matter because they show how large buyers can respond when AI costs rise: they can redirect work to tools they already have, rather than relying on one provider. But switching is not automatically a simple or cost-free way to save money. The value depends on whether an alternative handles the same tasks well and whether the expense of moving is outweighed by the savings.
For businesses, those expenses can include re-running evaluations, adapting prompts and software integrations, and retraining employees. Coding tools are often embedded in editors, repositories and team practices, so replacing one may disrupt established workflows. A different model can also require more human review or rework if it performs less well on a company’s particular tasks. Those costs may not appear in a token bill, but they affect the overall economics.
There can also be technical changes to account for. AI agents repeatedly use prior context, and changing providers may reset cached information or alter cache pricing. The source material argues that this can affect the cost of agent-based work, but does not supply company-specific figures to quantify the impact. The broader point is that per-token prices alone do not capture the full cost of a model choice.
As an affiliate, we earn on qualifying purchases.
Big Buyers Had Alternatives Ready
The reported changes are about internal employee use, not a complete withdrawal from Anthropic. The source account says Microsoft continues to use Anthropic models for some customer-facing Copilot features. It also says customer use of Claude through Microsoft platforms is growing. The report, as summarized here, does not provide a comparable figure for Meta customer use.
Both companies have strategic alternatives of their own. Meta develops models and internal coding tools; Microsoft owns GitHub Copilot and is a major backer of OpenAI. That makes their decisions different from those of a business that depends on an outside vendor and has not built or deployed a replacement. Their existing tools and engineering capacity can reduce the friction of redirecting some work, though they do not establish that switching is easy for other organizations.
The source material also points to changing AI subscription limits and pricing, citing a separate SemiAnalysis report. Such changes can affect the value of subscriptions, but no specific comparison or baseline is provided here. For buyers, the relevant distinction is between a provider’s listed price and the total cost of getting acceptable work done, including usage limits, integration and review.
enterprise AI tools for developers
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
What the Report Does Not Establish
The available account does not give direct statements from Meta or Microsoft explaining the decisions, nor does it establish that either company cited poor Claude performance. Cost pressures, tighter spending controls and a preference for in-house or affiliated tools are the reported explanations; they should not be treated as proof that Claude was less capable on the affected work.
It is also unclear how much employee work actually shifted, whether the reported user counts refer to active users or broader access, and how the companies measured the replacement tools’ quality. The projected Microsoft spending figure is not the same as confirmed annual expenditure. The reported reduction of more than a third applies to that projection, and the source material does not provide the final spending amount.
There is no basis in the material for calculating what a typical company would save by switching. Costs vary with workload, model performance, existing integrations and the amount of additional review required. The suggestion that a business spending $20,000 a month could face switching costs greater than a year of savings is an assertion in the source material, not a result supported there by a disclosed calculation.
As an affiliate, we earn on qualifying purchases.
Measure Costs Before Moving Work
The next useful evidence would be clearer figures from the companies on internal usage, spending and results, including how they compare Claude with the tools receiving the work. Until those details are available, the reported user counts and spending projection show the direction of the changes, but not their full financial or productivity effects.
For other businesses, the practical test is to measure performance on representative tasks before moving a large workload. Maintaining evaluations, tracking the cost per accepted result and keeping prompts and integrations portable can make a future change easier to assess. Those steps do not guarantee savings or better output; they give buyers a stronger basis for deciding whether a switch makes sense.
As an affiliate, we earn on qualifying purchases.
Key Questions
Have Meta and Microsoft stopped using Claude?
No. The report describes reduced or redirected internal use, not a complete end to Claude access. Microsoft is also reported to continue using Anthropic models in some customer-facing Copilot features.
Why are employees being directed to other AI tools?
The reported reasons include rising costs, tighter spending controls and the availability of tools the companies own or support. The material does not say that either company concluded Claude performs worse.
Does Microsoft spend more than $1 billion a year on Anthropic?
The figure refers to a reported projection of more than $1 billion a year in internal spending, which was later cut by more than a third. It is not presented as confirmed actual spending.
What makes switching AI providers expensive?
Potential costs include re-testing tasks, adapting prompts and integrations, helping staff learn a new tool, and handling any added review or rework. The total depends on the company’s workflows and the alternative’s performance.
Can smaller businesses expect the same savings as Meta or Microsoft?
Not necessarily. The two companies had alternatives already deployed and significant engineering resources. The source material provides no general savings estimate for smaller businesses, whose switching costs may differ.
Source: ThorstenMeyerAI.com
Halloween Picks
halloween
As an affiliate, we earn on qualifying purchases.
