Vendor comparison

OpenAI vs Anthropic DPA: side-by-side compliance read for buyers

Independent compliance comparison from Janus Compliance. Reviewed by Michael K. Onyekwere, CIPP/E. Last reviewed 2026-05-30. Not legal advice.

Profiles: OpenAI · Anthropic

Share this OpenAI vs Anthropic profile:Share on X Bluesky

TL;DR. Both vendors run protective commercial defaults: no training on customer data through the API and Enterprise tiers, ZDR available on approval. Anthropic edges OpenAI on the API retention default (7 days vs 30) and on the EU/UK transfer documentation depth. OpenAI edges Anthropic on enterprise tier maturity, BAA breadth, and the Azure OpenAI route for buyers already on Microsoft. Consumer-tier defaults at both vendors are not protective and need their own policy.

How the two compare on the eight clauses that decide the risk

This comparison is the buyer's-side read against the eight DPA clauses set out in the DPA for AI vendors hub. Full vendor-by-vendor detail lives in the OpenAI profile and the Anthropic profile.

1. Commercial training default

| | OpenAI | Anthropic | |---|---|---| | API | No training on customer data by default | No training on customer data by default | | Enterprise / Team | No training | No training | | Contract location of commitment | DPA + Service Terms | DPA + Commercial Terms | | How a buyer turns it on | Opt-in via account settings (uncommon) | Opt-in via account settings (uncommon) |

Both vendors are aligned. The clause is in the DPA, not buried in the marketing site, and the default favours the buyer.

2. Consumer training default

| | OpenAI | Anthropic | |---|---|---| | Free / consumer Plus | Opt-out (training default on; user has to disable) | Opt-in (training default off; user has to enable) since 2025-10-08 | | Retention on consumer plans | Conversation history retention up to 30 days by default; longer with memory enabled | Up to 5 years for users who opted in to training |

The consumer-tier default is the area where the two vendors diverge most. Anthropic's flip to opt-in on 2025-10-08 made consumer claude.ai the more protective consumer default, though the 5-year retention horizon for opted-in users is longer than OpenAI's consumer retention. Neither default is good enough to support PHI, special-category GDPR data, or anything subject to a procurement gate.

3. API retention default

| | OpenAI | Anthropic | |---|---|---| | Default API retention | 30 days | 7 days (dropped from 30 on 2025-09-14) | | ZDR available | Yes, approval-gated | Yes, approval-gated | | Time to ZDR approval (typical) | 1-3 weeks | 1-3 weeks | | Carve-outs from ZDR | Abuse and safety monitoring exceptions | Abuse and safety monitoring exceptions |

Anthropic's 7-day default is the strongest commercial-tier retention default in the LLM market at the time of this review. Buyers who need shorter than 7 days still have to go through the ZDR approval process at both vendors.

4. Subprocessor depth

| | OpenAI | Anthropic | |---|---|---| | Published subprocessor list | Yes, updated on change | Yes, updated on change | | Notice cadence on additions | 14-30 days | 30 days | | Objection mechanism | Customer may object; vendor proposes alternative or contract termination right | Customer may object; vendor proposes alternative or contract termination right | | Notable subprocessors at time of review | Microsoft Azure (training and inference); Cloudflare; Stripe; Snowflake | Amazon Web Services; Google Cloud; Cloudflare; Stripe | | Cross-vendor subprocessor relationship | OpenAI runs on Azure infrastructure (Microsoft is the cloud) | Anthropic became a Microsoft 365 Copilot subprocessor on 2026-01-07 |

OpenAI's Azure relationship and Anthropic's appearance as a Copilot subprocessor mean that a buyer using OpenAI and Microsoft 365 Copilot is, in many cases, transacting with the same upstream cloud through different prime contracts. The implications for data-flow diagrams and Transfer Impact Assessments are worth pinning down explicitly.

5. EU/UK transfer mechanism

| | OpenAI | Anthropic | |---|---|---| | EU SCC version | 2021 Module 2 (Controller-to-Processor) | 2021 Module 2 (Controller-to-Processor) | | UK Addendum | Yes | Yes | | Transfer Impact Assessment | Vendor publishes supporting docs; buyer responsible for TIA | Vendor publishes supporting docs; buyer responsible for TIA | | EU Data Boundary equivalent | EU data residency available on certain Enterprise plans | EU data residency available on Enterprise; the Microsoft 365 Copilot route is explicitly out of EU Data Boundary scope as of 2026-01-07 | | Customer-data location commitment | Region-restricted on Azure OpenAI for buyers using that route | Multi-region; primarily US infrastructure for direct API |

Buyers needing strict EU-only processing should pull the most recent regional commitments directly from the sales channel rather than relying on the standard DPA. Both vendors offer the option; both gate it through Enterprise contracts.

6. Security commitments

| | OpenAI | Anthropic | |---|---|---| | SOC 2 Type II | Yes | Yes | | ISO 27001 | Yes | Yes | | ISO 27701 | Yes | In progress at time of review; confirm directly | | ISO 42001 (AI management system) | Pursuing | Pursuing | | HIPAA BAA available | Yes, on Enterprise / API Enterprise / Azure OpenAI | Yes, on Enterprise; sales-channel gated | | GDPR Article 28 commitments | In DPA | In DPA |

Both vendors clear the realistic floor (SOC 2 + ISO 27001) and are pursuing the AI-specific ISO 42001. The HIPAA BAA picture matters for US healthcare buyers — see the HIPAA hub for the procurement workflow.

7. Deletion and contract end

Both vendors commit to deletion or return of customer data at contract end. For buyers who fine-tuned a model on customer data, the practical question is whether deletion extends to the fine-tuned artifact. OpenAI's fine-tuning deletion language is explicit; Anthropic's enterprise tier handles this through bespoke commercial terms — confirm in writing if fine-tuning is in scope.

8. Breach notification

Both vendors commit to notification of a personal-data breach without undue delay, with 72 hours as the realistic floor. Both flow the notification through to the buyer's named DPO contact.

Picking between them

The decision usually turns on two factors that sit outside the DPA itself.

Pick OpenAI when:

The buyer is already on Azure and wants the Azure OpenAI route for region-locked deployment
The buyer needs the breadth of HIPAA-eligible products (Enterprise, Edu, API on Enterprise, Azure OpenAI)
The buyer's app stack is built against the OpenAI API surface

Pick Anthropic when:

The buyer needs the shortest API retention default available without negotiating ZDR
The buyer's procurement gate requires the consumer-tier default to be opt-in for training, not opt-out
The buyer is already using Claude through the Claude API and wants to consolidate

Consider both alongside each other when:

The buyer's deployment is a multi-model pipeline (common in legal, financial-services, and clinical-summarisation work)
The buyer needs a fallback model provider for resilience and the two vendors run on different upstream cloud providers (Azure for OpenAI; AWS and GCP for Anthropic direct)

The consumer-tier exposure is the trap

The single biggest procurement risk with either vendor is not the commercial DPA — both are protective. It is the consumer-tier exposure when staff use the free or personal product alongside the contracted enterprise one.

A staff member who pastes a draft procurement contract into free ChatGPT or free claude.ai is operating under that vendor's consumer default. The enterprise DPA does not apply. The buyer's procurement policy has to address this directly, either by blocking the consumer products at the network or by writing an explicit usage policy that staff have signed.