Security Startups funded by Y Combinator (YC) 2026

DeepMark is the authentication layer for AI voice. When an AI agent speaks, we embed a compact, machine-readable ID directly into the audio, not in metadata, which gets stripped the moment a file is touched. The ID resolves to a trust record, so a bank, a call center, or the person on the line can verify in real time who the agent is.

Verdict Machine uses AI to understand each financial institution's complete on-chain environment, assess its cyber risk and generate tailored controls. The same live model powers transaction prevention, continuous cyber diligence, security posture and cost-efficient wallet screening. Before execution, Verdict Machine evaluates proposed transactions against the institution's active policies. In our first weeks, we started POCs with asset managers and digital-asset infrastructure providers whose systems move billions of dollars every week for financial institutions. We're now in active evaluations with digital-asset teams at some of the world's largest banks. The founding team launched and led blockchain security at Check Point (NASDAQ: CHKP).

Trident is the standard for modern day security. its an agentic pentester that chains vulnerabilities together like a real attacker mapping verified attack paths across AWS, Azure, GCP, web apps, and APIs. Run AI attacks and get fixes. Get started today: tridentsecurity.io

AI apps such as ChatGPT and Claude have become part of everyone's daily workflow, but they introduce new security risks that traditional security tools weren't built to handle. Security teams at companies from startups to Fortune 500s, are struggling to keep up. Traceforce gives a company’s security team visibility of how AI apps are being used across company’s devices, and the ability to detect and prevent security breaches and unsafe actions as early as possible.

Fabraix builds state-of-the-art AI red-teaming agents that continuously detect security vulnerabilities in customer-facing AI. Our product, Nyx, has already found vulnerabilities in agents at dozens of Fortune 500 companies. On AgentHarm, the leading benchmark for offensive AI security, Nyx achieved a 78% attack success rate, compared with 67% for GPT-5.6 Sol. AI agents can change more often than teams can test them. A new model, prompt, tool, permission, or data source can change what an agent does, even when the application code stays the same. And AI is also increasing how much software companies produce and how often it changes. We built Nyx to automate the work required to red-team AI agents. It is multi modal by design and interacts with the target indirectly through controlled replicas of SaaS products and websites that we maintain, placing malicious payloads in webpages, documents, files, messages, and tool outputs to test how an agent behaves under different conditions. Nyx often finds its first vulnerability within minutes or hours rather than days or weeks. Because the work is fully automated, companies can repeat the test with every change and at a much lower cost than what a 7-figure red-teaming (and point-in-time) engagement would require. We're also the team behind ACE (Adversarial Cost to Exploit), a benchmark that measures AI security in terms of how much it costs attackers to break an AI system; providing a game-theoretic framework to understand how motivated a rational attacker would be in exploiting the system.

NebuSec is founded by the world’s #1 hacking team. We were the first team to build an nginx RCE and the first to root Android 17. Our co-founders include members of the world’s No. 1 CTF team, DEF CON winners, Black Hat speakers, and PhD in cybersecurity. We have won $400K in bug bounties by exploiting the Linux kernel and Chrome browser. “Audited by NebuSec” is a mark of serious security credibility: a signal to customers, investors, and partners that your product has been reviewed by the world’s best hackers.

Tolmo is an agent-driven security platform that secures every layer — code, CI, cloud applications at the speed of AI changes. Codebases are growing faster than ever, at a rate humans can’t keep up with. Tolmo deploys fleets of AI security agents on a live production knowledge graph that connects data from cloud services, containers, databases, and third-party providers (Observability, Security, etc.). Tolmo agents triage, verify impact, and remediate security issues autonomously, escalating to humans only when truly necessary.

RASPIRE is the first line of defense for applications against AI-powered fraud attacks, API abuse, and runtime threats. Unlike code scanners and pentesting tools that identify vulnerabilities before release, RASPIRE operates like a security expert running inside applications, monitoring, detecting, and blocking attacks 24/7 in real time. All without writing a single line of code.

Clawvisor lets AI agents use apps like Gmail, Slack, and Google Drive without going rogue or ever seeing your credentials. You approve tasks once. Clawvisor enforces them on every request.

Crosslayer Labs detects impersonation attacks on websites and APIs. We provide "outside-in" monitoring of customer infrastructure, comprehensively discovering and monitoring all Internet dependencies, including DNS, BGP, TLS certs, Javascript. We correlate signals from different layers of the Internet stack to identify attack patterns. Our alerts are actionable and contain root cause analysis/remediation recommendations. Our founding team is world renowned for previously inventing the MPIC standard adopted by all the major CAs including Google, Apple, Amazon. The MPIC standard is securing over 700 million websites via the issuance of over 3 billion digital certificates, and is now a global internet standard that protects every single HTTPS connection.

BeeSafe AI is building the frontier model layer for social engineering defense. Cyberattacks have moved from code to conversation: attackers now use AI to carry out long-term, trust-based manipulation across SMS, voice, email, and social platforms, often inside encrypted channels where defenders cannot see, detect, or stop them. BeeSafe changes that by deploying undercover AI agents that engage attackers directly, expose hidden infrastructure, and convert live adversarial conversations into actionable intelligence. Each engagement adds to BeeSafe's proprietary dataset of real attacker behavior and trains foundational models that continually learn about counter-persuasion, deception, and manipulation in the wild. BeeSafe is starting in financial crime, where banks already value our data against Authorized Push Payment (APP) fraud and money laundering, but the opportunity is far broader. The same data and models that uncover mule networks today can protect consumers, secure enterprises, and strengthen governments against the next generation of AI-powered social engineering attacks tomorrow.

Parameter builds AI agents that run continuous penetration tests against your apps and infrastructure. Instead of a once-a-year penetration test, Parameter's agents works 24/7 to find and verify critical vulnerabilities so you can prevent them before attackers. Formerly Hex Security

We provide real-time video intelligence that identifies escalating incidents as they unfold. Protent detects early escalation patterns and predictive threat signals, giving security teams the immediate context they need to prioritize alerts and intervene before situations spiral out of control.

Lexius turns existing security cameras into an AI security guard that detects, documents, and alerts incidents like shoplifting and slip-and-falls. Millions of locations already have cameras but no intelligence. Replacing them costs $50K to $100K per site. Lexius upgrades every camera instantly, no rip and replace, no new hardware. Real-world camera systems are messy, fragmented, and decades old, and AI breaks on low-quality CCTV. The hard part isn’t the model, it’s making unreliable systems work at scale with accuracy high enough for real decisions. Lexius starts with security and liability, and becomes the operating system for physical spaces.

Agentic Fabriq helps companies securely manage what AI agents can access and do across business tools, data, and teams. We give each agent an identity, scoped permissions, approval flows, and audit logs, so employees and customer-facing workflows can use agents without relying on shared API keys, overbroad service accounts, or one-off integrations.

Kestrel is the AI automation layer for platform engineering teams. Platform teams use Kestrel to automate incident response, cloud provisioning, CI/CD, security, and developer requests across their stack. Kestrel turns natural-language prompts into deterministic, production-ready workflows with 40+ integrations, 200+ pre-built actions, and support for custom HTTP and webhook actions. Developer-first, with a CLI, Python SDK, and MCP server to manage workflows from where platform teams already work.

Veria Labs is developing AI agents that hack better than the best human hackers. Founded by members of the #1 US competitive hacking team, we've already found critical bugs in AI tools, operating systems, fintech and billion dollar crypto exchanges. Now we're scaling that expertise through AI-native security tooling that operates faster than human researchers while finding deeper, more complex vulnerabilities that traditional tools miss.

The agentic era is here, but its foundation is dangerously insecure. A growing number of high-profile "AI agent gone rogue" security incidents make headlines each week. As enterprise adoption of agentic software has far outpaced the development of applicable cybersecurity solutions, CISOs have loudly voiced a growing need for verifiable, fine-grained security controls on complex multi-agent systems. That's why we built Multifactor. Multifactor offers best-in-class authentication, authorization, and auditing technologies for agentic systems, allowing users to securely and revocably share online accounts with agents while maintaining fine-grained permissions and a detailed event history. We are doing for AI security what Cloudflare did for web security, Okta did for identity security, and Wiz did for cloud security: defining the category. A world-class "account manager" that allows agents to securely use online accounts on your behalf is just the start—our early enterprise partners include companies using AI agents to manage $7.5+ billion manufacturing supply chains, where the consequences for lapses in security are severe. Our patented AI security technology seamlessly integrates with existing agentic software to thwart attacks like prompt injection, so CISOs and end users alike can breathe a sigh of relief, without slowing the pace of AI adoption.

ContextFort gives security teams visibility and controls for AI browser agents. It’s a Chrome extension that: Detects when an AI agent takes control of the browser Records every action during the session (pages visited, clicks, text entry) Provides controls to prevent risky actions and cross-site data leakage

We emulate the attacks real adversaries run today: help-desk vishing, deepfaked executives, MFA fatigue, and the payloads that follow. We run them against your people and the controls around them to reveal what is actually exploitable, not what looks compliant. Then we close the gap, retest until the fix holds, and give your leadership the evidence that risk actually went down. From initial access to remediation.

Alter is a zero-trust identity and access control platform purpose-built for AI agents. It wraps every tool call in strong authentication, fine-grained authorization, and real-time guardrails, so agents can move fast without breaking things. Each request is verified at the parameter level, authorized against granular policies, executed with least-privilege access, and fully audited in real time. Unsafe actions, whether it’s a rogue DROP TABLE or a payment above policy limits, are blocked before they touch production. Behind the scenes, Alter manages credentials, issuing ephemeral, scope-narrowed access for every interaction, then rotating or expiring it in seconds. The result: no long-lived secrets, no blind spots, and no surprises in audit. With Alter, teams can move fast on AI agent initiatives while staying fully compliant with SOC 2, HIPAA, GDPR, and internal security standards. A CISO-ready dashboard delivers real-time visibility, detailed audit logs, and compliance-ready controls, removing silos, eliminating excessive permissions, and providing complete oversight of every agent workflow.

Probo acts like your compliance team. You tell us how you work, and Probo does the rest: 1. Set the right requirements - not overkill 2. Handle docs, risks, controls, evidence 3. Run the audit 4. Keep everything up to date afterward Founders shouldn’t lose weeks to this stuff. Even one day is too much.

Casco performs autonomous security testing for your web apps, APIs, infrastructure, and AI systems. Human supervision optionally available. Use Casco for year-round security and get a penetration test report for your SOC2 and ISO27001 compliance certifications. Trusted by 100+ companies from enterprises to fast-moving startups, such as Gusto, CrewAI, Novig, and Accrual. Hack yourself at: https://casco.com --- We already secure software deployed in 60% of Fortune 500 companies and are default alive. [1] [1] http://www.paulgraham.com/aord.html

MindFort is building fully autonomous AI agents powered by their in-house models that can find, exploit, and patch complex vulnerabilities in web applications at massive scale. MindFort agents use the same principles as human penetration testers to understand their target and figure out the best ways to work together to exploit it. We can do fully autonomous AI web pentesting and bring offensive security to teams out numbered. While security teams were already behind with legacy tools and processes, AI coding is make the problem worse. AI generates code far faster than human teams can secure it; MindFort can match that speed and scale. Our agents will become critical as organizations scale up development with AI.

Cotool is an agentic security platform that eliminates manual and repetitive work for security teams. Today’s cybersecurity teams remain bogged down by alert fatigue, context switching, and documentation overhead. Security teams don't need more dashboards with "AI-powered insights" — they need their time back. Cotool empowers teams to build security their way. Our platform provides an AI co-pilot that automatically assembles context across tools, a no-code agent builder that transforms successful investigations into reusable automations, and automated report generation, providing comprehensive documentation in seconds. Unlike solutions that force teams to adapt to vendor workflows, Cotool puts the power of AI directly in practitioners' hands, letting them shape tools around their unique processes. Early users have already cut time spent on investigation and detection engineering by 70%.

Golf gives enterprises visibility into shadow AI and control over AI agents and MCP. See what's connected, enforce policies, and generate audit trails - all from one place.

Tinfoil makes it easy to make your AI workloads secure and provably private. You get the privacy of on-prem deployments, while running on the cloud. It's like end-to-end encrypted messaging but for your AI applications. We’ve built a full-stack platform on top of the latest NVIDIA GPUs offering confidential computing capabilities, meaning that you don’t have to trade off performance for privacy. We integrate a suite of recent advances in secure hardware technologies, in particular NVIDIA’s confidential compute mode available on Hopper and Blackwell. When combined with Tinfoil’s software stack, companies can prove their security claims like "we can't see and don’t log your queries.” Tinfoil guarantees that all data always stays private and cannot be accessed by anyone other than the end user -- not even by Tinfoil or the cloud provider it’s processed on. Everything can be made end-to-end encrypted and private with Tinfoil. The founding team combines deep academic and industry experience in security and Internet protocols. Tanya was previously a systems engineer at Cloudflare, where she built Internet security protocols used by billions, and contributed to the Workers AI platform. Jules and Sacha hold PhDs from MIT, where they worked on secure hardware, cryptography, and privacy technologies. Jules has also worked at NVIDIA on their confidential computing team. Tinfoil was born from our personal frustration with the false choice between access to powerful AI and the massive data privacy implications of deeply integrated AI. AI is a universal tool that becomes part of our personal lives and business workflows. In the process, it needs access to all personal, private, and proprietary data. Right now, the only solutions are data processing agreements (i.e., "pinky promises"), band-aids like PII redaction that don't work in practice, or AI locally/on-prem, which doesn't scale. Tinfoil promises to unlock significantly deeper AI adoption and integrations by making it easy to obtain true privacy, just as TLS on the Internet enabled e-commerce to flourish by securing credit cards on the network. Today we offer several self-serve products: - A consumer-facing chatbot that keeps your chats end-to-end encrypted and provides zero-access inference with powerful open-source models like Kimi, Gemma, and GLM. You can try it out for free: https://chat.tinfoil.sh - A developer-friendly API that allows you to build AI applications where all user data is kept private and we as the inference provider cannot see any of the prompts. Our SDKs make sure that all relevant security measures are checked for you before even sending a single byte of data. Learn more:https://docs.tinfoil.sh - An advanced confidential computing platform we call Tinfoil Containers that allows you to secure any Docker application. We make it easy to deploy your own logic and applications inside secure enclaves and prove to the world what code is running inside. This enables attested workloads and publicly verifiable privacy policies for sensitive applications. Learn more: https://tinfoil.sh/containers

Candor is agentic data loss prevention (DLP) and insider risk management. Employees share sensitive data with third parties and email source code to personal inboxes every day. For fast-moving companies, protecting these crown jewels and customer trust is core to closing deals. Legacy DLP tools can cross off compliance checkboxes but flood security teams with false positives. Candor's AI agents understand both the data movement and the person behind it, giving lean security teams broad coverage without the noise.

SubImage builds maps of your infrastructure to give you visibility on what assets you have and how they relate to each other. This enables any security team to deeply understand their environments and reduce risk. SubImage is a managed offering built around the Cartography open-source project, and is founded by members of Cartography's original team.

We built Gecko for teams that want to build secure code quickly without wasting time on security tools that don’t deliver results, or relying on one-time human pentests that quickly become outdated. Gecko uses AI to understand how your application should work, simulates relevant attacks to find critical vulnerabilities, and then verifies these vulnerabilities by exploiting them. It also helps you understand the risk of these vulnerabilities and applies a working fix to keep your code secure.

Winfunc (winfunc.com) is an AI Hacker that autonomously finds, verifies, and patches security vulnerabilities in codebases, just like a human security engineer would. Winfunc can find business logic errors with context-aware scanning and automate the full auditing cycle to generate reports with zero false positives. Winfunc's agent has found vulnerabilities in Google, Supabase, Sentry, Cal, Gumroad, Bun, Hoppscotch, etc.

Clearly AI automates security and privacy reviews with AI. We performed hundreds of reviews at Amazon and Moveworks and now we're building the product we wish we had. Our users complete reviews in minutes instead of days. Emily and Joe are technical co-founders and senior engineers. We met in 2019 while working on an Alexa security review.

ZeroPath is a developer tool that autonomously detects, verifies, and submits fixes for vulnerabilities in your code. Engineers can use ZeroPath to find security problems that they might only otherwise catch in pentests or from bug bounty researchers.

Tracecat helps AI-native security teams build agents and automate work. Open source and community driven at it's core: https://github.com/TracecatHQ/tracecat

Shiboleth automates consumer lending compliance for financial institutions using AI. Banks and fintechs use us to save months of manual work by automating audits and drafting reports for the governments. LLMs allow whole categories of manual compliance processes to be automated in ways that weren’t possible until recently. We are excited to automate back-office operations and enhance consumer protection in one of the most litigious industries.

Superagent helps AI-native developers trust what they merge, control what their agents do, and show users it's safe. We work in your coding agent and on GitHub: - Trust the change: checks on every pull request, before merge - Control it where it runs: boundaries around what your agents can see, say, and do - Break it before it ships: red teaming against what you actually run Free for open source.

Rimward helps solar asset managers detect intrusions using the cameras already installed across their sites. It flags camera failures and coverage gaps, and gathers video evidence from multiple cameras to help their team or monitoring provider filter out false alarms and decide when to act.

CompliantLLM detects every data leak into third-party GenAI tools used in your company by monitoring employee interactions. Security and tech leaders get visibiliity and control over the data flowing into various Gen AI apps used in their organizations.

Metalware develops advanced firmware security solutions for critical infrastructure, protecting industries like aerospace, defense, automotive, telecom, and healthcare from cyber threats. Our product is an automated, intelligent binary fuzzer that enables organizations to efficiently discover and remediate security weaknesses in hardware products before deployment.

Corgea is an autonomous application security platform that finds and validates vulnerabilities, then generates safe code fixes developers can apply in their existing workflow. In production customers like Zapier and Yageo, see ~20% more true positives and ~90% fewer false positives (vs their prior tooling), reducing triage and speeding remediation without adding headcount.

Proof of Human is an API that invisibly detects bots and verifies human identity. Founded by Princeton cognitive scientists, Proof of Human leverages world-class research with industry-beating benchmarks so platforms can verify their human users friction-free.

Variance builds AI agents for fraud, identity, and investigations for Fortune 500s, Marketplaces, and regulated financial services.

Replace legacy scanners and manual offensive security processes with AI agents that discover, test, and remediate directly in your engineering workflows.

Infisical is the all-in-one security infrastructure platform for secrets, certificates, and privileged access management. Infisical stores your secrets and injects them back into your development workflow, CI/CD pipelines, and cloud infrastructure. It comes with secrets versioning, point-in-time recovery, audit logging, automatic secret rotation capabilities, and more.

90% of security breaches are caused by a simple employee mistake. Your employee clicks a bad link, and gives up their password. They download ransomware from an email and run it. They leak your intellectual property and sensitive customer data to ChatGPT. Repacket allows your employees to enjoy the internet without risk - instantly blocking threats like phishing and sensitive data leaks. This gives your employees the ability to make that mistake without impacting your company. Repacket is built for smaller companies, who may not have dedicated IT or security people to run the show. We install in 60 seconds and give your employee instant protection where they need it most. Founded by cybersecurity leaders from SpaceX, AWS, and Google - we're building the easy button for cybersecurity.

Matano is a modern SIEM, built for cloud-first security teams. It replaces traditional SIEM databases like Splunk or Elastic with a cybersecurity platform built on top of a cost-effective Security Data Lake.