The annual DataTribe Challenge uncovers the most promising cybersecurity startups, and gives them an opportunity to pitch their firms and products to an audience of venture and private equity firms and security experts.
DataTribe itself is a seed stage venture foundry focused on early stage cybersecurity firms – startups. Each year the firm holds the DataTribe Challenge – effectively a pitch competition for startups. From these, five finalists are chosen and allowed to deliver pitches during DataTribe’s Cyber Innovation Day conference (this year, set for November 4, 2025).
The foundry element comes from the nature of DataTribe and the common willingness of startups to accept as much help and advice as they can get. “We call ourselves a startup foundry,” explains Leo Scott (MD and chief innovation officer at DataTribe). “What that means is our team will jump in and co-build with the founders to help accelerate growth and de-risk things. Our whole team at DataTribe are all former startup operators. We’re just some crazy people that kept building startups over and over — and then eventually ended up joining DataTribe to help others do the same thing. We enjoy jumping in and getting our hands dirty.” And the startups benefit from their experience.
Each of the five finalists in the Challenge are automatically awarded $5,000 (designed to amply cover expenses). But the real reward is being able to pitch to the Innovation Day conference audience before the ultimate winner is chosen.
This conference is attended by cybersecurity venture and private equity firms, researchers, government officials and defense agencies, and a large sprinkling of CISOs. It is a focused channel for getting attractive cybersecurity startups in front of potential investors – including, of course, DataTribe itself. “Historically, it’s not guaranteed, but historically DataTribe has invested in the winners. But if it’s not DataTribe, there’s a whole audience filled with the top cybersecurity investors in the industry,” said Scott.
The five finalists are now chosen. By definition, they are emerging security vendors to watch – and not surprisingly, this year four of the five directly involve AI. The firms are Ackuity, Cytadel, Evercoast, Starseer, and Tensor Machines.
Tensor Machines is the only finalist not directly focused on AI. This firm provides a hardware security solution that enables printed circuit boards, sensors, and other hardware components to continuously self-certify authenticity while in use. It offers more granularity with fewer development issues than other approaches to the same problem.
Ackuity provides security for AI agents. It collects data from all the interactions occurring inside an agentic system, whether it’s between the user and the agent, between different agents, or the agent and other tools. It builds this telemetry in real time providing real time threat detection and response capabilities.
“They’re doing a kind of pre-SIEM processing, because existing SIEMs don’t really have the context to make sense of this new kind of agent telemetry,” comments Scott. “It can generate events, actions or alerts that can be directly evaluated or sent into a SIEM if you have one.”
Cytadel provides AI-driven autonomous red teaming. It is based in London, UK, and its founder and CEO has a background in GCHQ and was head of red teaming at the Bank of England.
“He has an automated red teaming platform leveraging AI to do decision making to build a full attack path at a level of complexity that historically could only be done by human red teamers. You can validate your defenses, quantify your resistance to ransomware, generate a list of actions that can be taken, and even produce an executive report on weaknesses and possible solutions,” explains Scott.
Evercoast is described as more cyber adjacent than cybersecurity focused. It provides the data used in embodied AI; that is, the AI that controls robots. The firm started by servicing the entertainment industry; by capturing the real movement of an actor and allowing that to be included in a fake CGI scene. “They had multiple cameras and other sensors attached to the subject that would collect the data of movement that could then be added to the movie,” explains Scott.
But as robots and AI have evolved, and mobile robots – bi-pedaled and wheeled – are emerging, the firm is adapting its process to teach robots how to move independently through their AI training rather than just manual control. Evercoast is training robots to collect all the movement data from manually controlled robots or humans doing different tasks and allowing the AI to use that as its own training data. It’s not that simple, but by the end of the process, the robot has clean data from which to learn. “And the better the learning, the more difficult it will be to hijack the robot for nefarious purposes,” adds Scott.
Starseer, the fifth finalist, provides security for new AI models. The rush to adopt AI as rapidly as possible is creating a brand new threat surface that is little understood and frequently ignored. Starseer provides a platform that enables AI system builders to reduce system risk and improve resilience before deployment.
“It treats AI-systems as probabilistic, analyzing the models, the training data, the refinement steps, and the surrounding libraries and system architectures to surface exploitable weaknesses and operational risks – ensuring production systems are safer and more reliable,” explains DataTribe.
The DataTribe challenge is a win-win for all concerned. The most exciting new developments in cybersecurity get to pitch their companies to multiple VC firms on Cyber Innovation Day, while DataTribe itself has had early access to all the finalists. And we shouldn’t dismiss the relevance of glimpsing the future for all the security professionals in the audience.
Related: Former NSA Director Rob Joyce Joins DataTribe as Venture Partner
Related: Frenos Raises $3.88M in Seed Funding for OT Security Assessment Platform
Related: Email Security Startup AegisAI Launches With $13 Million in Funding
Related: Offensive AI Startup Dreadnode Secures $14M to Stress-Test AI Systems
