Skip to content

Claude in Microsoft Copilot: What Changes for Users Outside the US

Your next creation starts with a sentence. — free credits on sign-up Bring my idea to life

Microsoft Copilot is no longer tied to a single model provider. According to Crypto Briefing, the rollout of Anthropic's Claude into Copilot began on 24 September 2025. Users can now pick Claude variants, including Sonnet 4 and Opus 4.1, alongside OpenAI models. This is an expansion, not a replacement. For users outside the US, the most visible difference is not raw capability but access: outside the EU, EFTA and the UK, Claude is enabled by default for most commercial users; inside those three areas, an administrator must turn it on manually.

A model choice that varies by region

The source says Claude has become the default model for many commercial customers, without specifying which ones or how that status is determined. That gap matters: a user cannot infer from their subscription alone that they are getting Claude by default. The verifiable point is elsewhere. Where activation is automatic, the model slips into the workflow without a prior decision; where it is manual, an administrator opens or closes the door.

According to the source, this asymmetry stems from local data-boundary rules. Claude's availability in Europe is restricted to comply with them. In other words, compliance does not merely govern storage: it shapes the product experience itself. The same Copilot can offer different options depending on the account's region.

What it means for an organisation

For a business, a second model provider inside an already-deployed tool mainly creates decisions to document:

  • Governance: who authorises activation, and on what criteria?
  • Data handling: where do prompts go, and with what contractual guarantees?
  • Logging: can you tell which model produced which answer?
  • Training: do teams know the underlying model may vary?

The source makes exactly this point: more options mean more decisions, governance and data-handling reviews. For teams in the EU, EFTA or the UK, the practical question is simple: is manual activation a temporary hurdle or a lasting choice? The source does not say, and assuming an answer would be unwise.

Why Microsoft is diversifying its AI

Claude's integration fits a broader strategy. Microsoft is also developing its own models, called MAI, and continues to use OpenAI technology. Diversification here is not a switch but an accumulation of suppliers.

The reported financial commitments point the same way. On 18 November 2025, Anthropic reportedly pledged to secure $30 billion in Azure compute capacity. Microsoft reportedly committed $5 billion to Anthropic, and NVIDIA $10 billion. The source does not detail the terms, timeline or exact scope of these commitments: the figures should be read as reported announcements, not verified flows.

Another element shows the depth of the relationship: on 9 March 2026, Microsoft launched Copilot Cowork, an agentic tool built on Anthropic technology. The source does not say whether it will be available in all regions, or under what conditions.

An ambiguous signal on cost

The source reports that Microsoft cut its projected internal spending on Claude by more than a third. The claim is interesting but thin: it comes with no primary source and no absolute figures. It suggests an extra provider does not necessarily mean an extra bill at the same level, but it says nothing conclusive about what customers pay.

For Anthropic, the deal is above all a distribution win: Copilot is embedded in Microsoft tools that businesses already pay for. For Microsoft, it is a way to reduce reliance on a single partner while keeping control of its own model.

Key takeaways

  • Claude has been available in Copilot since 24 September 2025, alongside OpenAI models — an expansion, not a replacement.
  • Outside the EU, EFTA and the UK, Claude is on by default for most commercial users; inside those areas, an administrator must enable it.
  • Data-boundary rules fragment the experience by region.
  • Microsoft now combines OpenAI, Anthropic and its own MAI models.
  • The reported figures ($30bn Azure, $5bn Microsoft, $10bn NVIDIA) lack detail on their terms.

For a user outside the US, the first useful step is not technical: check with your administrator whether Claude is enabled, and under which data-handling rules. The rest — costs, regional scope, Copilot Cowork conditions — remains partly uncertain at this stage.

Source

Anthropic pledges $30B in Azure compute as Claude moves into Copilot

Related reading