It depends on the route. Since 22 october 2026, Claude on Google cloud comes two ways. On the Anthropic managed route, Anthropic operates the service and acts as data processor. On the Google managed route, Google is the processor. So before you approve, ask which route your team picked and who processes your data.
#Key takeaways
- Anthropic made Claude models on Google cloud (Anthropic managed) generally available on 22 october 2026, next to the existing Google managed route.
- The processor is different on each route: Anthropic on its managed route, Google on the Google managed route.
- The processor question and the where-does-it-run question are two separate questions, and both belong in the approval.
- Anthropic describes the Google managed route as the better fit for strict regional residency requirements.
- This is a plain-english guide for buyers. It is not legal advice.
#What changed on 22 october 2026?
Anthropic announced that Claude models on Google cloud (Anthropic managed) is generally available, with us and global endpoints. Anthropic, october 2026 says Anthropic operates this route.
The Google managed offering does not go away. According to the same Anthropic announcement, october 2026, it remains available, with Google as the processor. Your company now has two doors to the same family of models, and they are not identical.
#Why does the data processor matter to a non-technical buyer?
The processor is the company that handles your data on your behalf. That name decides which contract applies, whose security review you read, and who your legal and it teams talk to when something goes wrong.
If your team signs off on a route without asking, the answer gets decided by whoever set up the account. A head of operations or a CIO should know the name before the first real document goes in.
#Does the route change where my data is processed?
It can. The Claude developer docs, undated page read in october 2026 describe residency as two settings: inference geo, which is where the model runs, and workspace geo. Today inference geo offers only us and global, and workspace geo offers only us.
On the Google managed route, the same docs say the inference region follows the endpoint url instead. So the question to put to your it team is simple. Which endpoint are we using, and where does that endpoint run?
We are
Stennir helps you ask the processor and hosting questions, then pick the route that fits your own rules.
We aren't
Stennir is not a law firm and not a cloud reseller. It does not give legal advice and does not sell cloud capacity.
#Which route fits strict hosting rules?
Anthropic's announcement describes the Google managed offering as the better fit for strict regional residency requirements. That is the vendor's own description, so treat it as a starting point and confirm it against your documents.
For companies in the UAE, Saudi Arabia and the wider gcc, hosting rules often come from a regulator, a customer contract or an internal policy. Stennir builds AI where your rules require, in the country, cloud or system that fits. The route should follow the rule, not the other way round.
#What should our it team confirm before we choose?
- Which route is it: Anthropic managed or Google managed?
- Who is named as the data processor on that route?
- Which endpoint is in use, and where does inference run?
- Where is the workspace, and what is stored there?
- Does this match our written hosting rules and our customer contracts?
- Who signs off, and where is that decision recorded?
Each answer should come from the vendor's current documents, not from a chat message. Stennir's ai-for compliance page shows how teams record decisions like this, and the glossary explains terms your it team may use.
#What should a head of operations do next?
Write down the six questions above and send them to your it lead and your cloud contact this week. Then compare the answers with your hosting rules. Stennir's AI consulting helps you pick the route that fits those rules, and the free readiness assessment is a gentle first step. When you are ready, book a 30-minute call.
#Common questions
What does our it team need to confirm before we choose a route?
They need to confirm four things in writing: which route is in use, who is the named data processor on it, which endpoint serves requests, and where inference runs. Then they check those answers against your hosting rules and customer contracts. Get the answers from the vendor's current documents, not from memory.
Who signs off on the processor and hosting choice?
Usually the executive accountable for the process, often a CIO or head of operations, together with legal or compliance and it. The point is that a named person approves the route and the decision is recorded. It should not be settled by whoever created the cloud account.
Does this change where our data is stored or only where it is processed?
The two are separate settings. The Claude developer docs describe inference geo, where the model runs, and workspace geo. Today they list only us and global for inference and only us for workspace. On the Google managed route, the endpoint url sets the inference region.
Does it work in Arabic?
The processor and residency questions do not depend on language. Stennir supports Arabic and English fully, and other languages on request. Whatever route you choose, test it on your own Arabic documents and review the results with the people who use them, before you approve a wider rollout.
What could go wrong if we pick the wrong route?
You could end up with a processor you did not expect, or with inference running somewhere your hosting rules do not allow. Fixing that after documents have gone in is harder than asking first. This is not legal advice, so involve your own counsel for contract questions.
