Anthropic

Security and trust center evidence

Anthropic, PBC is an American artificial intelligence (AI) public benefit corporation headquartered in San Francisco, California

anthropic.com AI infrastructureLast verified 3 Nov 2026

Request the report Are you Anthropic?

HIPAAVendor-stated

SOC 2Verified

FedRAMPVendor-stated

ISO 27001Vendor-stated

Cyber EssentialsVendor-stated

ISO 42001Vendor-stated

SOC 3Verified

Anthropic has 2 registry-verified rows and 5 vendor-stated rows in the CertReports index, last verified 3 Nov 2026. The strongest row is HIPAA: Vendor displays a HIPAA badge on its trust centre (claim; BAA availability not yet captured). This page is independent of the vendor’s own trust centre: dates, sources and caveats come from CertReports captures.

Reviewer brief

Anthropic has a SOC 2 Type II report, evidenced by a SOC 3 report issued 12 Nov 2025 for the period ending 30 Sep 2025, confirmed as of 29 Sep 2026. The vendor states on its trust centre a HIPAA badge as of 21 Sep 2026, though BAA availability is not confirmed in the evidence. Anthropic also states FedRAMP High on its trust centre as of 21 Sep 2026; this is a vendor statement, not a registry listing. It states ISO/IEC 27001:2022, ISO/IEC 42001:2023 and UK Cyber Essentials on its trust centre, all as of 21 Sep 2026. Only the SOC 2 and SOC 3 evidence is independently audited; the remaining frameworks rely on Anthropic's own trust centre statements dated 21 Sep 2026.

SOC 2 Type II: evidenced by SOC 3 report, issued 12 Nov 2025, period ending 30 Sep 2025, confirmed as of 29 Sep 2026.
FedRAMP High: vendor-stated on trust centre as of 21 Sep 2026 (not a registry listing in these rows).
HIPAA badge, ISO/IEC 27001:2022, ISO/IEC 42001:2023 and UK Cyber Essentials: vendor-stated on trust centre as of 21 Sep 2026; HIPAA BAA availability not confirmed.

Facts only, each dated; nothing here is inferred, scored or advised. Brief generated 29 Sep 2026; it is regenerated whenever a row changes.

Among ai infrastructure vendors

2verified rows

Category median 1, across 35 indexed ai infrastructure vendors. Anthropic has more verified rows than 89 percent of them.

generative ailarge language modelsllm api

Compare with similar vendors

Pick your own comparison

Compliance grid

No public evidence yet for GDPR, PCI DSS, CSA STAR, ISO 27701. This does not mean the vendor lacks them; it means nothing public was found at the last check.

Legal artefacts

Public trust centre documents

Public · [Anthropic] 2025 Type 2 SOC 3 Report.pdf; [Anthropic] ISO 27001 Certificate (2025).pdf; [Anthropic] ISO 42001 Certificate (2025).pdf; [Anthropic Ireland Limited] Cyber Essentials Certificate (2025) · as of 19 Sep 2026

Open

Subprocessor list

Public · 20 subprocessors listed on the trust centre · as of 21 Sep 2026

Open

Public · as of 21 Sep 2026

Open

AI and data use policy

Public · Listed on the trust centre as "Anthropic Frontier Compliance Framework" · as of 21 Sep 2026

Open

Subprocessors (20)

Amazon Web ServicesCloud Infrastructure · Worldwide
AL
Arkose LabsFraud and abuse detection · United States
BO
BoldrUser support · Canada
BS
Brave SearchWeb Search · United States
CloudflareTraffic Routing (CDN) · Worldwide (Local to Customer)
ElevenLabsText to speech · United States
Functional Software, dba SentryError Handling, User Support · United States
Google Cloud PlatformCloud infrastructure · Worldwide
IntercomUser support · United States
IT
IterableEmail communications · United States
MA
Microsoft AzureCloud Infrastructure · Worldwide
NU
NutunUser support · South Africa
PF
Palantir Federal Cloud Service (PFCS)FedRAMP Cloud · United States
PersonaFraud and abuse detection, identity verification · United States
SiftFraud and abuse detection · United States
StripeBilling · United States
TU
TurboPufferWeb Search · United States
TwilioAnalytics, email/SMS communications · United States
WO
WorkOSSecurity, Single Sign-On · United States
YO
YotiFraud and abuse detection, identity verification · United Kingdom

Security profile

Pen test cadence: annual
Trains on customer data: no

Displayed as a security profile from public statements, not as attestations, as of 21 Sep 2026.

Change history

25 Sep 2026SOC 2 state changedState moved from Vendor-stated to Verified.
25 Sep 2026SOC 3 evidence addedA SOC 3 row entered the index with state Verified.
22 Sep 2026AI training policy changedTraining on customer data moved from unknown to no.
19 Sep 2026First indexed by CertReports6 evidence rows across 6 frameworks entered the index.

Similar vendors with evidence

Related by product tags and the AI infrastructure category, ranked by shared tags, description similarity and overlapping evidence. Never by popularity.