Tip-offs
•
Each week, we surface promising startups & founders through data signals.
This week, we’re spotlighting AI safety and the rise of the guardian agents →
📊 Model Risk Index
Even the top models crack under pressure
PropensityBench, a benchmark by ScaleAI, evaluates how likely AI models are to select harmful or misaligned actions when placed under intense operational pressure across 5,874 simulated, high-stakes scenarios, spanning biosecurity, cybersecurity, chemical security, and self-proliferation. It shows what models choose to do when given a choice between a safe path and a harmful shortcut.
The results? The safest model fails 10.5% of the time, and the riskiest fails 66.5%.
🛡️ AI Security & Governance Startups
High-growth startups building the security, compliance, and governance layer for enterprise AI. Ranked by website visit growth over the past 3 months.
PromptArmor | 3603% growth in 3 months
AI risk management platform helping enterprise security teams assess and monitor third-party vendor AI risk, with a focus on prompt injection and data exfiltration threats. Raised a $3.2M Seed round in Apr 2024.
San Francisco, US
Delve | 156% growth in 3 months
AI-powered compliance automation platform that eliminates compliance busywork for startups and enterprises across SOC 2, HIPAA, GDPR, and ISO 27001. Raised a $32M Series A in Jul 2025 from Insight Partners.
San Francisco, US
Xygeni | 146% growth in 3 months
All-in-one application security platform covering SAST, SCA, secrets detection, CI/CD security, and software supply chain protection for modern development teams. Raised a $4.4M Seed round in Jun 2023 from Investing Profit Wisely.
Madrid, Spain
Nexos.ai | 79% growth in 3 months
Enterprise AI platform providing a unified workspace and API gateway for deploying and governing hundreds of AI models, with built-in Shadow AI controls and compliance oversight. Raised a $35M Series A in Sep 2025 from Evantic Capital and Index Ventures.
Vilnius, Lithuania
Portkey.ai | 78% growth in 3 months
LLMOps and AI gateway platform helping engineering teams reliably build and scale AI applications, with features spanning model routing, observability, cost controls, and network-level guardrails across 200+ LLM providers. Raised a $15M Series A in Feb 2026 from Elevation Capital.
San Francisco, US
🏌️♂️ AI Security & Governance Founders
Serial founders with exits building the infra layer for AI safety, governance, and LLM security, from LLM red-teaming to agent security and enterprise AI governance.
Leonid (Leo) Feinberg | 2x Exits, 3x Founder
Previously co-founded AcceloWeb (web acceleration, acquired by Limelight Networks) & CloudEndure (cloud migration, acquired by AWS). Now co-founder & CEO of Verax AI, helping enterprises gain real-time visibility and control over LLM-based products in production.
London, UK
Michael D'Angelo | 1x Exit, 3x Founder
Co-founded Matroid (computer vision platform, Series A) & Arthena (acquired by Masterworks). Now co-founder & CTO of Promptfoo, the leading open-source platform for LLM evals, automated red-teaming & vulnerability detection.
New York, US
Ariel Dan | 1x Exit, 2x Founder
Co-founded Porticor Cloud Security (cloud encryption & key management, acquired by Intuit). Ex-CEO of Cloudify (acquired by Dell). Now co-founder & CEO of Rogue Security, the end-to-end platform for AI agent security covering discovery, red teaming, guardrails & monitoring.
Tel Aviv, Israel
Pankaj Thapa | 1x Exit, 3x Founder
Previously co-founded Mobiliya (mobile security engineering, acquired by QuEST Global) and founded FireSense.ai (GenAI legal & maritime models). Now co-founder & CEO of Mirror Security, building a comprehensive AI security platform.
Mountain View, US
Omri Iluz | 1x Exit, 2x Founder
Previously co-founded PerimeterX (behavior-based bot & web threat protection, merged with HUMAN Security). Now co-founder & CEO of Lumia Security, an AI usage control platform providing visibility, governance & real-time guardrails across every layer of AI adoption.
New York, US
🤝 Carried Interest
Tweets of the week
