In October 2025, nearly 100,000 people watched a deepfake “Jensen Huang” push a crypto scam on YouTube. In fact, for a while, more people watched the deepfake stream than Nvidia’s real GTC keynote. Months earlier, criminals deepfaked an entire management team on video to convince a Hong Kong finance worker to wire $25 million.
These aren’t anomalies. BEC (business email compromise) attacks rose 54% from 2023 to 2024. Criminals now use LLMs to craft personalised phishing at scale, deepfake live video calls, and spin up convincing fake personas in minutes. The playbook has changed—and so must the defence. For example, take a look at this clever phishing email… would you spot it?
The list of cyber crimes goes on… from fake job candidates leveraging their remote positions to steal information to prompt injection for AI agents, it seems like the vectors of attack have become much more diverse in the age of generative AI. Israeli cybersecurity startups, which attract 20% of the global cybersecurity funding, are very much on the case to fight the bad guys.
Israeli Startups Building the Defences
Israel’s security ecosystem is leading the counteroffensive with deepfake forensics, brand impersonation takedowns, and behavioural fraud analytics. Israeli cyber startups ship signal-rich detections (behavioural, liveness, provenance) and rapid takedown mechanics, and not just alerts. Financial services and consumer platforms now treat deepfake resilience as a board-level risk, so the market is actively looking for these solutions.
Autonomous & Agentic Security (Next-Gen SOC)
- BlinkOps – Raised a significant Series B ($50M+) to advance its no-code platform that builds and deploys autonomous software agents to handle tasks like incident response, access control, and patching.
- Tonic Security – Emerged from stealth with a $7M seed round (founded 2024). Their AI platform uses a Data Fabric architecture and Agentic AI to map assets, intelligently prioritize thousands of threats, and rank them by business impact.
- Radiant Security – Focuses on using AI to augment the SOC, specifically to speed up triage, investigation, and response times for alerts, allowing human analysts to focus on complex threats.
Cloud & Runtime Security (CNAPP)
- Sweet Security – Raised a significant Series A ($33M+). Their platform uses runtime intelligence to detect and respond to threats in real-time, pioneering the use of AI to transform early warning signals into a cohesive full-attack story.
- Wiz – continually integrating cutting-edge AI/ML into its platform to analyse vast cloud environments and identify toxic combinations of risks.
Brand, Identity, and GenAI Misuse Protection
- Clarity – Deepfake detection (video/voice/image); $16M seed from 8200 alumni
- Cyabra – Bot/disinformation detection + deepfake capabilities for brands and governments
- ActiveFence – AI moderation platform across text, audio, image, video
- Semperis – Identity resilience for hybrid environments, specialising in protecting and recovering Microsoft Active Directory (a prime hacker target).
- Orchid Security – Identity-first Security Orchestration platform to manage and secure access across the organisation.
Brand & Identity Protection
- Memcyco – They are actively enhancing their real-time protection to specifically combat the growing threat of Generative AI-powered scams and phishing.
- BrandShield – Digital risk protection: phishing, executive impersonation, counterfeit takedown
- Guardio – Browser-level protection against phishing and malicious extensions
Email & Workspace Security
- IronScales – AI-powered email security; Gartner MQ “Visionary” (2024)
- Perception Point (acquired by Fortinet) – Cross-platform threat prevention (email, browser, collaboration apps)
- Cyvore Security – Secures the entire digital workspace, including email, messaging apps, video conferencing, and CRM systems, addressing emerging threats like deepfakes in video calls.
Behavioral & Fraud Analytics
- BioCatch – Behavioural biometrics used by banks to stop authorised push payment scams
- CHEQ – Bot mitigation for marketing funnels and form-fill abuse
- Forter – eCommerce fraud detection platform that uses machine learning to prevent online payment fraud.
- Refine Intelligence – Next-generation Anti-Money Laundering (AML) solution, using AI to detect sophisticated financial crime.
- nSure.ai – Fraud protection and prevention specifically for digital goods and non-physical purchases.
For a more comprehensive list of startups protecting AI Agents, I recommend Insight Partner’s recently published landscape:
What’s Working on Defence
The hackers and scammers only need to succeed once, and the defenders need to succeed every time to stop attacks. In this never ending game of cat and mouse, it’s often the attackers that win. While there’s no silver bullet here are some best practices for protecting your digital property in 2025:
1. Assume AI in the loop
If your playbooks assume human-speed adversaries, you’re behind. Attackers now iterate and personalize at machine scale.
2. Layer hard-to-fake signals
- Behavioural: Cursor patterns, keystroke cadences, mobile tilt (BioCatch)
- Provenance: C2PA cryptographic signatures on media assets
- Liveness: Randomised prompts and challenge-response for audio/video
3. Move verification out of band
Adopt DoD-style challenge phrases and verified callback numbers for any payment or access change. Even multi-party video can be faked.
4. Harden where lures land
Expect polished, multilingual vendor impersonation. Pair modern email security with strict URL sandboxing and browser shields that catch malicious extensions before users do.
5. Monitor social surfaces actively
Deploy brand monitoring and bot detection on owned communities. Platform enforcement is tightening, but your telemetry closes the loop.
Where we’re looking
At Remagine Ventures, we recognise that there are many specialised cybersecurity focused funds in Israel and we shied away from making investments in that space. That being said, as early investors in GenAI startups (we made our first investment in the space in HourOne, text to video avatars in August 2019), we believe that the work now is to instrument, verify, and disrupt and to fund the teams making that possible as these threats increase.
If you’re an Israeli startup building in the space of defence from the dark side of GenAI, we’d love to hear from you!
I'm a former general partner at google ventures, head of Google for Entrepreneurs in Europe and founding head of Campus London, Google's first physical hub for startups.
I'm also the founder of Techbikers, a non-profit bringing together the startup ecosystem on cycling challenges in support of Room to Read. Since inception in 2012 we've built 11 schools and 50 libraries in the developing world.
- Weekly Firgun Newsletter – February 27, 2026 - February 27, 2026
- The Rise of the Personal AI Agent - February 26, 2026
- Pre-Seed Is Stronger than ever, but less startups are getting funded - February 25, 2026
