Specific IT protection, but few disclosed rules and tests for AI
Overall rating “limited”. No deduction for controversies. The superscript numbers link directly to the source.
1 · Responsibility
Cadence assigns AI risks to the oversight of the full board, but names neither responsible persons nor an AI policy or approval procedure.
What is documented
- According to the 2026 proxy, the full board oversees risks from technological innovation including AI; the Audit Committee oversees cybersecurity and the agreements with BIS and DOJ1
- The 10-K states as a risk factor that AI can produce flawed or biased content and that Cadence must ensure its responsible use2 3
- The code of conduct and the supplier code offer anonymous reporting channels with a prohibition of retaliation4 5
- Mandatory training on data security and export control is described6
The case against
AI governance not documented
Reports not readable
- The sustainability reports for 2024 and 2025 could not be retrieved; AI figures, AI training and external AI audits are missing7
Sources for Responsibility (7)
- Cadence Design Systems / U.S. SEC: Proxy Statement DEF 14A 2026 · 3 November 2026 · primary source
- Cadence Design Systems (Investor Relations): CDNS FY 2025 Form 10-K (PDF) · 3 November 2026 · secondary source
- Cadence Design Systems / U.S. SEC: Form 10-K for fiscal year 2025 (HTML) · 18 February 2026 · primary source
- Cadence Design Systems: Worldwide Code of Business Conduct (Rev. Feb 2020) · 3 November 2026 · secondary source
- Cadence Design Systems: Supplier Code of Business Conduct (Rev. May 2021) · 3 November 2026 · secondary source
- Cadence Design Systems: Cadence 2023 Environmental, Social and Governance Report · 3 November 2026 · secondary source
- ResponsibilityReports.com: Cadence Design Systems Inc – report archive · 3 November 2026 · secondary source
2 · Safety
Cadence describes its cybersecurity programme in specific terms, but discloses neither AI-specific testing nor usage restrictions for its agents and models.
What is documented
- The 10-K describes a CISO, alignment with NIST, CSA and ISO, external penetration tests, an incident response plan and fixed reporting cycles to the Audit Committee and board4
- SOC 2 Type 1 for cloud products and a zero-trust approach are described as company statements5
- The Maestro models are secured with NVIDIA OpenShell and can be run entirely on the customer's premises6
- Data protection rights and a data protection officer are named8
The case against
AI testing not disclosed
Protection against misuse not described
- Neither usage restrictions nor misuse monitoring nor incident statistics are disclosed for AI agents and models
Sources for Safety (8)
- Cadence Design Systems / Business Wire: Cadence Unleashes ChipStack AI Super Agent, Pioneering a New Frontier in Chip Design and Verification · 10 February 2026 · secondary source
- Cadence Design Systems (Investor Relations): CDNS Q4 2025 Earnings Conference Call Prepared Remarks · 17 February 2026 · secondary source
- CXO Digital Pulse: Cadence Introduces Maestro AI Models for Semiconductor and System Design · 27 October 2026 · secondary source
- board-cybersecurity.com (reproduction of Form 10-K): CADENCE DESIGN SYSTEMS INC 10-K Item 1C. Cybersecurity – 2026-02-18 · 18 February 2026 · primary source
- Cadence Design Systems: Cadence 2023 Environmental, Social and Governance Report · 3 November 2026 · secondary source
- Reuters via The Star: Cadence unveils Nvidia-backed AI models that chipmakers can run in-house · 26 October 2026 · secondary source
- Cadence Design Systems / Nasdaq: Cadence Completes Acquisition of Secure-IC · 30 October 2025 · secondary source
- tostracker.app (archived copy of the Cadence Privacy Policy): CADENCE DESIGN SYSTEMS INC — Privacy Policy (Last Revised 2025-12-23) · 23 December 2025 · secondary source
3 · Transparency
Cadence names AI risks in the 10-K and the use of data for AI in its privacy policy, but discloses neither model cards nor training data nor limits for its own Maestro models.
What is documented
- The privacy policy names the use of data to improve services, including with LLMs and generative AI, and the storage of chat transcripts1
- The 10-K names general AI risks and regulatory frameworks and points to possible flawed AI output6 3
- The Maestro models are partly built on NVIDIA Nemotron and can be retrained by customers4 5
The case against
Own models barely documented
Limits not described
- Product-specific information on limits, error rates or levels of autonomy of the AI tools is missing
- It is unclear when customer data are processed in AI functions
Sources for Transparency (6)
- tostracker.app (archived copy of the Cadence Privacy Policy): CADENCE DESIGN SYSTEMS INC — Privacy Policy (Last Revised 2025-12-23) · 23 December 2025 · secondary source
- Cadence Design Systems / Business Wire: Cadence Unleashes ChipStack AI Super Agent, Pioneering a New Frontier in Chip Design and Verification · 10 February 2026 · secondary source
- Cadence Design Systems (Investor Relations): CDNS FY 2025 Form 10-K (PDF) · 3 November 2026 · secondary source
- Reuters via The Star: Cadence unveils Nvidia-backed AI models that chipmakers can run in-house · 26 October 2026 · secondary source
- CXO Digital Pulse: Cadence Introduces Maestro AI Models for Semiconductor and System Design · 27 October 2026 · secondary source
- Cadence Design Systems / U.S. SEC: Form 10-K for fiscal year 2025 (HTML) · 18 February 2026 · primary source
4 · Why the harm rule does not apply
The harm rule does not apply: within the review period there is no judgment, regulatory decision or fine concerning harm caused by the use of AI.
5 · Controversies (0)
No AI controversy found.
6 · What remains open
- AI-specific governance structures, policies and figures
- Validation and security testing of the agentic AI and its own models
- Handling of confidential customer designs in AI functions
- Content of the Corporate Impact Reports 2024 and 2025
- Usage restrictions and misuse monitoring for AI agents and models
7 · How it was assessed
Cut-off date: 3 November 2026 · 44 searches, 29 fetches.
Primary documents read (7)
- Form 10-K FY2025 (p. 07 HTML, two retrievals; p. 08 PDF, two retrievals; Item 1, Item 1A including AI risk factor and export case)
- Item 1C reproduced via board-cybersecurity.com (p. 09)
- Proxy Statement DEF 14A 2026 (p. 10)
- Form 8-K 2025-07-27 (p. 05), 8-K Annual Meeting 2025 (p. 11), Form 10-Q Q3 2026 (p. 23, two retrievals)
- Plea Agreement via DOJ (p. 03) and as SEC exhibit (p. 06), DOJ case page (p. 04)
- Official findings: DOJ press release (p. 01), BIS press release (p. 02)
- 2023 ESG Report (p. 13), Code of Business Conduct (p. 12), Supplier Code (p. 14), Privacy Policy (p. 15)
Research of 3 November 2026 under Globalance AI Footprint rulebook version 0