AWS Bedrock vs Google Vertex AI for enterprise

Choose the cloud your company already runs on and buys from: both serve Claude and open models under your existing cloud identity, billing and audit trail. Choose Vertex AI if Gemini is your main model or you want a hard spend cap on the service. Choose Bedrock if your stack is on AWS and you want Claude, GPT and open models under IAM.

Google now calls the product Gemini Enterprise Agent Platform (formerly Vertex AI). Both bill per token on demand and sell reserved capacity, and both ration on-demand capacity, so check your quotas against your peak load before you commit.

Last updated 9 November 2026. Facts about AWS Bedrock and Google Vertex AI come only from their own public pages, listed under Sources.

AWS Bedrock and Google Vertex AI side by side

AWS Bedrock compared with Google Vertex AI
AWS BedrockGoogle Vertex AI
ModelsClaude, OpenAI GPT-6, GPT-5.x and gpt-oss, Llama, Mistral, DeepSeek, Grok and others; Google listed with Gemma, Gemini not listedGemini, plus Claude, Mistral, Llama, DeepSeek, GLM, gpt-oss and others as managed models; OpenAI GPT not listed
On-demand pricingPer token on your AWS billPer token on your Google Cloud bill
Reserved capacityProvisioned Throughput per model unit per hour; no-commit, 1-month or 6-month termsProvisioned Throughput in GSUs with 1-week, 1-month, 3-month or 1-year commitments
Capacity limitsTokens per minute per model per Region; increases are not automatic and may be deniedGemini: organization-level throughput tier that grows with spend; Claude: quotas per model and endpoint, raised on the Quotas page
Processing in EuropeEU geographic cross-Region profiles stay in the EU; global profiles may route to any commercial Regioneu multi-region covers EU member states only; the global endpoint may process in any Google Cloud location
RetentionData retention mode set per Region; some models keep traffic up to 30 days for abuse detectionIn-memory caching on by default, deleted within 24 hours, can be turned off per project; Search grounding stores data up to 3 days
TrainingInputs and outputs not used to train base modelsNot used to train or fine-tune models without your prior permission or instruction
IdentityAWS IAM; short-term Bedrock API keysGoogle Cloud IAM; Workforce Identity Federation with Entra ID, Okta or AD FS
BudgetsAWS Budgets alerts, updated up to three times a day; budget actions can apply an IAM policyAlert budgets, plus spend-cap budgets that pause the service until you lift the cap
AuditAWS CloudTrail records Bedrock API callsCloud Audit Logs: Admin Activity always on; Data Access logs off by default
ContractAWS GDPR terms; Standard Contractual Clauses in the AWS Service TermsCloud Data Processing Addendum with EU SCCs for restricted transfers
From each vendor's own pages, checked 9 November 2026.

Where they differ most

The model you build on

Gemini runs on Google's platform and is not on the Bedrock model list, which does list OpenAI's GPT-6 and GPT-5.x. Google lists OpenAI's open-weight gpt-oss but not GPT. Claude is on both.

Capacity

Bedrock sets tokens per minute per model per Region, and increases are not granted automatically. Google says a 429 on Gemini doesn't indicate a fixed quota: standard usage runs on shared resources with a throughput tier that adjusts with spend.

Europe

Bedrock's EU cross-Region profiles keep inference in the EU. Google's `eu` multi-region covers EU member states only and excludes the UK and Switzerland, and newer Claude models there are served from the multi-region rather than single EU regions.

Spend control

AWS Budgets alert after the fact. Google Cloud now offers spend-cap budgets for the platform, which pause the service per project when the cap is reached.

Choose AWS Bedrock or Google Vertex AI

Choose AWS Bedrock if

  • Your workloads, identity and procurement already run on AWS.
  • You want Claude and OpenAI's GPT models on one cloud bill, and Gemini is not a requirement.
  • Your approved per-Region quotas cover your peak load.

Choose Google Vertex AI if

  • Gemini is your main model, or you want Gemini and Claude on one cloud bill.
  • You want a spend cap that stops the service, not only an alert.
  • Your workloads and identity already run on Google Cloud.

A third option for companies

Either way you commit to one cloud's catalog and quotas. Standard Compute serves Claude across several clouds, so one provider's quota or outage doesn't stop your work, and puts GPT, Gemini and open models on the same key.

Cost per step

Each step goes to the cheapest model that can do it and the cheapest provider fast enough to serve it.

Models

Claude, GPT and Gemini alongside open models such as GLM, DeepSeek, MiniMax and Kimi, through one API.

Processing in Europe

A Europe rule per company, team or key: every request, including retries and fallbacks, is served only by approved European inference endpoints, or refused before your prompt is sent.

Zero data retention

Required per company, team or key, and Standard Compute never trains on your prompts or outputs.

Single sign-on

Microsoft Entra ID or any SAML provider, and Claude Desktop and Claude Code sign-in with a Microsoft work account.

Budgets

A fixed monthly budget per person and app that never overruns, and spending caps per key and for the whole company.

Allow-lists

Approved providers per company, team or key, and approved models per key. Fallbacks never route around them.

Audit log

Admin changes with before and after values, and a 12-month export of every request's model, provider and region rule.

Sign our Data Processing Agreement in the dashboard, and see every sub-processor with its purpose and location.

Create a company workspaceor write to contact@standardcompute.com

Questions

Is Claude available on both AWS Bedrock and Google Vertex AI?

Yes. Both list Anthropic's Claude models. Bedrock serves them under per-Region quotas; Vertex serves them per endpoint, with newer models sharing one quota per model lineage per global or multi-region endpoint.

Which keeps data in the EU?

Bedrock's EU geographic cross-Region profiles keep inference within the EU. Google's eu multi-region covers EU member states only. On both, global options may process data outside Europe.

Can I cap spend on Bedrock or Vertex AI?

AWS Budgets send alerts and budget actions can apply an IAM policy. Google Cloud offers spend-cap budgets that pause the platform per project until you lift the cap, alongside alert-only budgets.

What happened to the Vertex AI name?

Google's documentation now calls the product Gemini Enterprise Agent Platform (formerly Vertex AI).

Is there an option that is not tied to one cloud?

Standard Compute is one API for Claude, GPT, Gemini and open models. Claude is served across several clouds, and each step goes to the cheapest model that can do it.

Sources (checked 9 November 2026)

  1. Amazon Bedrock model cardsdocs.aws.amazon.com
  2. Bedrock runtime quotasdocs.aws.amazon.com
  3. Bedrock quotasdocs.aws.amazon.com
  4. Bedrock Provisioned Throughputdocs.aws.amazon.com
  5. Bedrock cross-Region inferencedocs.aws.amazon.com
  6. Bedrock data retentiondocs.aws.amazon.com
  7. Amazon Bedrock FAQsaws.amazon.com
  8. Managing costs with AWS Budgetsdocs.aws.amazon.com
  9. Bedrock API calls in CloudTraildocs.aws.amazon.com
  10. AWS GDPR Centeraws.amazon.com
  11. Google Model Garden pricingcloud.google.com
  12. Google Provisioned Throughputdocs.cloud.google.com
  13. Google standard usage tiers and 429sdocs.cloud.google.com
  14. Claude quotas on Google Clouddocs.cloud.google.com
  15. Google partner modelsdocs.cloud.google.com
  16. Google open modelsdocs.cloud.google.com
  17. Google data residencydocs.cloud.google.com
  18. Google data governancedocs.cloud.google.com
  19. Workforce Identity Federationdocs.cloud.google.com
  20. Google Cloud spend-cap budgetsdocs.cloud.google.com
  21. Vertex AI audit loggingdocs.cloud.google.com
  22. Google Cloud Data Processing Addendumcloud.google.com