Anthropic Digital Audit

Enterprise Technology Stack & Vendor Relationships

• Anthropic’s Trust Center publicly identifies AWS, Google Cloud Platform, Microsoft Azure, and Cloudflare as subprocessors or infrastructure providers, with disclosed functions including cloud infrastructure and traffic routing ¹.
• Anthropic distributes Claude through Google Cloud Vertex AI, and Google Cloud announced Claude 3 availability on Vertex AI on March 14, 2024 ².
• AWS states that Anthropic’s models are available through both Amazon Bedrock and the separate “Claude Platform on AWS” offering ³.
• Anthropic states that U.S. government agencies can access Claude through partnerships with AWS, Google Cloud, and Palantir ⁴.
• Anthropic’s Palo Alto Networks case study states that Palo Alto Networks uses Claude on Google Cloud Vertex AI for internal developer-productivity workflows, which evidences Anthropic as a supplier to an Israeli-founded cybersecurity company rather than a purchaser of that company’s tools ⁵.
• No public evidence identified of verified Anthropic licensing, subscription, or integration relationships in which Anthropic is a customer of Check Point, Wiz, SentinelOne, CyberArk, NICE, Verint, Claroty, or Palo Alto Networks ⁶¹⁵.

Surveillance, Biometrics & Retail Technology

• No public evidence identified.
• No public evidence identified that Anthropic uses Israeli-origin facial recognition, biometric identification, gait analysis, retail analytics, predictive policing, sentiment-monitoring, social-media monitoring, or workforce-surveillance tools ⁶¹⁷⁸⁹.
• No public evidence identified that those technology categories reach Anthropic indirectly through named third-party enterprise suites or managed services in the reviewed public record ⁶¹.

Cloud Infrastructure, Data Residency & Sovereign Cloud Participation

• Anthropic’s supported countries page lists both Israel and Palestine as supported locations for Claude availability and/or commercial API access ¹⁰.
• Anthropic states that for its commercial products customer data storage is U.S.-only by default, while processing may occur in select countries in the U.S., Europe, Asia, and Australia unless otherwise agreed ¹¹.
• Anthropic states that it uses multiple cloud providers for processing and directs customers to its subprocessor disclosures for further detail ¹¹¹.
• Anthropic states that for Claude for Work it processes commercial-customer data as a processor on the customer’s behalf ¹¹.
• Anthropic’s Trust Center identifies AWS and Google Cloud Platform as worldwide cloud-infrastructure subprocessors for all products ¹.
• AWS states that for Claude Platform on AWS, customer data and metadata are transferred to Anthropic for processing, while Claude on Amazon Bedrock keeps customer data within AWS infrastructure ³.
• AWS announced Amazon Bedrock in the Israel (Tel Aviv) Region on September 29, 2025, and AWS regional model documentation lists Anthropic models in Bedrock regional availability tables ¹²¹³.
• Those AWS materials evidence Anthropic model availability through infrastructure available in Israel, but they do not establish an Anthropic-operated data center in Israel ¹²¹³.
• Anthropic has not publicly identified an Anthropic-operated or Anthropic-leased data center in Israel in the reviewed materials ¹¹¹¹⁴.
• No public evidence identified that Anthropic is itself a named participant in Project Nimbus or another Israeli sovereign-cloud procurement, although Anthropic distributes through Google Cloud and AWS, which TIME identified as the prime Nimbus contractors ²¹⁵.
• No public evidence identified that Anthropic markets or contracts services specifically to provide digital sovereignty, residency, or resilience for Israeli state institutions or military bodies ⁴¹⁶³².

Defence, Intelligence & Security Sector Technology Relationships

• Anthropic states that its U.S. Department of Defense agreement carries a $200 million ceiling and that Claude is integrated into defense workflows on classified networks with partners including Palantir ¹⁷.
• Anthropic states that U.S. government customers can access Claude through AWS, Google Cloud, and Palantir while maintaining control over their data ⁴.
• Anthropic announced Claude Gov models on June 6, 2025 as models built exclusively for U.S. national security customers and stated that they are already deployed by agencies operating at the highest level of U.S. national security ¹⁶.
• Anthropic states that Claude Gov use cases include strategic planning, operational support, intelligence analysis, and threat assessment ¹⁶.
• Palantir stated in its February 5, 2024 Q4 2023 Business Update that it agreed to a strategic partnership with the Israeli Ministry of Defense to supply Palantir technology supporting Israel’s war effort ¹⁸.
• Palantir stated in its May 22, 2025 response to the UN Special Rapporteur that its work in Israel spans civil, defense, and intelligence-related government mandates ¹⁹.
• The reviewed public record therefore establishes an Anthropic relationship with Palantir in classified and national-security deployments and a separate Palantir record of defense and intelligence work in Israel ^1741819.
• No reviewed public source in this memo shows Anthropic itself contracting with the Israeli Ministry of Defense, the IDF, or Israeli intelligence agencies ^1741819.
• No public evidence identified of Anthropic developing, selling, licensing, or maintaining offensive cyber capabilities, zero-day exploit tools, or digital weapons systems for Israeli state actors ¹⁷⁴¹⁶⁷.

AI, Algorithmic & Autonomous Systems

• Anthropic publicly markets and supplies AI systems for U.S. national-security use cases including intelligence analysis, operational support, and threat assessment through Claude Gov and related public-sector offerings ⁴¹⁶.
• No public evidence identified that Anthropic has verified provision of AI or machine-learning systems to Israeli state, military, or security bodies ⁴¹⁶²⁷.
• No public evidence identified that Anthropic models have been publicly reported as trained on, or granted access to, civilian population data, intercepted communications, or surveillance-derived datasets originating from Israel or the occupied Palestinian territories ¹¹¹⁶¹⁷.
• No public evidence identified that Anthropic provided autonomous target-generation, automated threat-detection, or autonomous tracking systems to Israeli forces ¹⁷⁴¹⁶⁷.

Technology Ecosystem & R&D Footprint

• Anthropic’s current jobs page shows roles in the United States, United Kingdom, Canada, and Switzerland, and no Israel-based roles were identified on the reviewed page ¹⁴.
• No public evidence identified of an Anthropic-operated R&D center, engineering office, accelerator, or innovation lab in Israel ⁶¹⁴.
• No public evidence identified of completed Anthropic acquisitions of Israeli-origin technology companies in the reviewed record ⁶.
• No public evidence identified of Anthropic patent portfolios, licensing arrangements, or co-development agreements with Technion, Hebrew University, or the Weizmann Institute in the reviewed record ⁶.
• Anthropic’s company page lists directors and Long-Term Benefit Trust members, but the reviewed corporate disclosures do not tie board composition itself to Israeli state institutions or Israeli technology firms ⁶.
• Anthropic’s Economic Index materials identify Israel among the highest countries on the company’s AI Usage Index and state that Israel accounted for 1.1 percent of total global Claude.ai use in the cited dataset, which evidences user adoption rather than state contracting ²⁰.

Civil Society Scrutiny & Regulatory History

• The research memo’s document-scoped pre-scan reported no Anthropic mention in the locally stored copies of UN document A/HRC/59/23, the 2023 OHCHR settlement-database update, and the 2025 OHCHR settlement-database update ⁷⁸⁹.
• No public evidence identified that Anthropic has been the subject of an organized BDS campaign specifically tied to technology provision to Israel or to settlements ⁷⁸⁹.
• No public evidence identified of regulatory inquiries, legal challenges, export-control actions, or sanctions-related investigations involving Anthropic technology sales or services to Israeli state entities ⁷⁸⁹.
• No public evidence identified from reviewed sources of pre-2020 Anthropic-Israel technology relationships because Anthropic’s company materials state the company was founded in 2021 ⁶.