AI Provider Trust Registry evidence verified as of 2026-08-19

Registry / DeepSeek via Fireworks AI

DeepSeek via Fireworks AI

developer: DeepSeek platform: Fireworks AI category: cloud distribution

Open-weight DeepSeek models (V3/R1 family, including R1-0528, V3.1 and V3.2) served on Fireworks AI's own inference platform. Fireworks is a US company running on AWS/GCP/Oracle data centers; prompts are processed by Fireworks, not sent to DeepSeek (Hangzhou), which is not a subprocessor. Vendor-trust and data-handling dimensions below describe Fireworks; EU AI Act dimensions describe DeepSeek as the model developer.

Watch-outs 6

The cells where this offering is not a clean public yes. This is what to check before you sign.

  • SOC 2 Type II: Partial SOC 2 certifies Fireworks' platform, not the DeepSeek model specifically.
  • ISO 27001: Partial Platform ISO 27001 only; DeepSeek model not individually covered
  • ISO 42001: Partial ISO 42001 certification applies to the Fireworks platform only; DeepSeek model itself is not individually certified.
  • HIPAA BAA: Yes, sales-gated BAA requires contacting sales and eligibility review
  • Residency: Partial Verified on Fireworks' own docs: self-serve --region pinning to listed EU single regions exists, but single-region quota is only granted at the GLOBAL multi-region level by default and requires contac
  • GPAI Code: No, verified DeepSeek is not present on the EU Commission’s complete signatory list, which is the authoritative source for GPAI Code of Practice signatories, so its absence is a confirmed no.
Vendor trust
SOC 2 Type II Is a SOC 2 Type II report available for this offering?
Partial confidence: high · verified 2026-08-14

Verified the Fireworks trust portal itself lists the SOC 2 Type 2 Report 2025; as a serving-platform certification not covering the DeepSeek model independently, partial is correct.

## SOC 2 Type 2 Report 2025

source

ISO 27001 Is there an ISO/IEC 27001 certification covering this offering?
Partial confidence: high · verified 2026-08-14

Fireworks AI’s Trust Center confirms ISO 27001 for the platform, and no public ISO 27001 evidence was found for DeepSeek itself, so the offering is partially covered.

ISO 27001 verifies that Fireworks has implemented a documented, risk-managed information security program with controls covering access, encryption, incident response, and vendor management.

source

ISO 42001 Is there an ISO/IEC 42001 (AI management system) certification?
Partial confidence: high · verified 2026-08-16

Fireworks Trust Center lists an ISO 42001 Certificate for the platform, and no authoritative DeepSeek page shows the model has its own ISO 42001 certification.

## ISO 42001 Certificate

source

Trust center Is there a maintained trust center / compliance portal?
Yes, public confidence: high · verified 2026-08-15

Fireworks AI maintains a publicly accessible trust center page listing compliance resources and reports, confirming a maintained compliance portal for the DeepSeek offering.

## Compliance ... ### HIPAA Report of Compliance Published January 29, 2026 Added most recent HIPAA Report of Compliance.

source

Data handling
HIPAA BAA Will they sign a HIPAA Business Associate Agreement covering this offering?
Yes, sales-gated confidence: high · verified 2026-08-18

The provider’s own FAQ confirms a BAA can be provided but only after a sales request, matching the definition of yes_sales_gated.

Fireworks AI supports HIPAA-compliant use cases and can provide a Business Associate Agreement (BAA) for eligible customers whose use of the Services involves the processing of Protected Health Information (PHI). To request a BAA, please contact [email protected]

source

GDPR DPA Is there a public DPA with SCCs and a published subprocessor list?
Yes, public confidence: high · verified 2026-08-18

Fireworks AI publishes a public Data Processing Addendum that defines SCCs and also provides a publicly accessible subprocessor list on its trust centre.

"Standard Contractual Clauses" or "SCCs" means the Standard Contractual Clauses annexed to the Commission Implementing Decision (EU) 2021/914.

source

No-training default Is there a public commitment not to train on customer API data by default?
Yes, public confidence: high · verified 2026-08-19

The provider's own privacy notice publicly states they do not train on API inputs without explicit opt‑in, satisfying the commitment.

No AI Training on Your Data: We do not use your prompts, training data, or API inputs to train or improve our AI models without your explicit opt-in.

source

Retention / ZDR Is retention documented, and is zero-data-retention available?
Yes, public confidence: high · verified 2026-08-14

Fireworks publicly documents zero-data-retention for open models and DeepSeek is classified as an open model, confirming retention is documented and zero-data-retention is available.

Fireworks does not log or store prompt or generation data for open models, without explicit user opt‑in.

source

Residency Can data be pinned to a region (especially the EU)?
Partial confidence: high · verified 2026-08-14

Verified on Fireworks' own docs: self-serve --region pinning to listed EU single regions exists, but single-region quota is only granted at the GLOBAL multi-region level by default and requires contacting Fireworks, so the capability is partially self-serve and partially gated. Quote is verbatim on

If you have a specific need for a single region, contact Fireworks to request it. ... For specific single region quota, please contact Fireworks.

source

EU AI Act
GPAI Code Is the model developer on the EC's GPAI Code of Practice signatory list?
No, verified confidence: high · verified 2026-08-19

DeepSeek is not present on the EU Commission’s complete signatory list, which is the authoritative source for GPAI Code of Practice signatories, so its absence is a confirmed no.

## Signatories of the code of practice - AI Studio Delta - Aleph Alpha - Almawave - Amazon - Anthropic - Black Forest Labs - Bria AI - Cohere - Domyn - Dweve - Fastweb - IBM - LINAGORA - Microsoft - Mistral AI - Open Hippo - OpenAI - Pleias - ServiceNow - WRITER In addition, xAI signed up to the Safety and Security Chapter...

source

Art. 53 summary Has the model developer published the Art. 53 training-data summary?
Yes, public confidence: high · verified 2026-08-16

DeepSeek's own page (cdn.deepseek.com/policies/en-US/model-algorithm-disclosure.html) states verbatim "DeepSeek hereby publishes Training Data Summary" and describes the pre-training data categories, so an Art.-53-style training-content summary is publicly available on the developer's site and no_pu

To enhance transparency, DeepSeek hereby publishes Training Data Summary. During the pre-training phase, corpus data is required for training. This phase primarily uses the following two categories of data: Public Data ... Licensed Data ...

source

Spotted an error? Submit a correction with evidence, corrections with a primary source are folded in and credited in the changelog.