Leaders Fund

Oasis Security raised $120M Series B to secure the way enterprises deploy and manage AI agents, and we're proud to be part of it alongside Craft Ventures. Agents are already inside your enterprise, and most security teams aren't set up to govern them. Since OpenClaw and Opus 4.6 dropped, agentic AI stopped being a concept and moved into production. They operate autonomously, are always-on, and move at machine speed. The challenge is that the security infrastructure most enterprises run was built for a completely different era. It was built to manage human access, defined by known identities, predictable behavior, and relatively static permissions. AI agents don't fit that model. Every one of those access events is a potential risk, and most organizations have no systematic way to control them. That's the gap Oasis Security is closing. They've built the control plane for agentic AI: infrastructure to manage non-human identities (agents, bots, service accounts, automated tools) across the enterprise, with real-time visibility into what those identities are, what they can access, and what they're doing. Static snapshots of an environment aren't enough when the attack surface is dynamic and adversaries are automated. Danny Brickman and Amit Zimerman saw this shift early, and it's showing in their customer list. Organizations with some of the most demanding security standards in the world, such as LifeLabs, Blue Cross Blue Shield, and Citizens, are all standardizing on Oasis. We believe every enterprise running AI agents will need Oasis, and we're excited to partner with Danny, Amit, and the team on the journey. Link in comments to learn more.

Oasis Security raised $120M Series B to secure the way enterprises deploy and manage AI agents, and we're proud to be part of it alongside Craft Ventures. Agents are already inside your enterprise, and most security teams aren't set up to govern them. Since OpenClaw and Opus 4.6 dropped, agentic AI stopped being a concept and moved into production. They operate autonomously, are always-on, and move at machine speed. The challenge is that the security infrastructure most enterprises run was built for a completely different era. It was built to manage human access, defined by known identities, predictable behavior, and relatively static permissions. AI agents don't fit that model. Every one of those access events is a potential risk, and most organizations have no systematic way to control them. That's the gap Oasis Security is closing. They've built the control plane for agentic AI: infrastructure to manage non-human identities (agents, bots, service accounts, automated tools) across the enterprise, with real-time visibility into what those identities are, what they can access, and what they're doing. Static snapshots of an environment aren't enough when the attack surface is dynamic and adversaries are automated. Danny Brickman and Amit Zimerman saw this shift early, and it's showing in their customer list. Organizations with some of the most demanding security standards in the world, such as LifeLabs, Blue Cross Blue Shield, and Citizens, are all standardizing on Oasis. We believe every enterprise running AI agents will need Oasis, and we're excited to partner with Danny, Amit, and the team on the journey. Link in comments to learn more.

Important work from the Upwind Security here. This is what modern attacks look like: targeted, adaptive, and designed to evade traditional detection. And this is why runtime security matters. Static scans won't catch what happens in real time. Upwind identified this early and is already supporting customers as it happens, exactly the kind of response you want when things move this quickly. Proud to be on this journey with the team.

We're proud to participate in Oasis Security's $120M Series B and to back a team building at the center of one of the most important shifts in security. At its core, every system relies on identity: knowing who or what is accessing a system, and what actions they're allowed to take. But that model was built for humans. We're now seeing the limits of that model as AI agents become active participants in how work gets done. They make decisions, take actions, and interact with systems autonomously. And every one of those actions requires access. The result is an explosion of non-human identities (NHIs) operating across environments that were never designed to manage them, creating a rapidly expanding and largely ungoverned attack surface. Oasis Security was built for this new reality, providing the infrastructure to secure agentic AI. It's impressive how early the team saw this shift and how quickly that vision is translating into real traction with forward-thinking enterprises such as LifeLabs, Blue Cross Blue Shield, and Citizens. This latest round accelerates that momentum, with focus on deepening the platform and expanding go-to-market. Congrats to Danny Brickman, Amit Zimerman, and the entire Oasis team on this incredible milestone. We're excited to support you as you shape the future of identity security in an AI-native world.

🚨 Big milestone for Upwind Security 🚨 Upwind is now live inside the AWS Security Hub Console as one of a select group of CNAPP extension partners. For customers, that means direct access to Upwind inside the AWS console, with unified billing and zero deployment friction. Milestones like this are the result of strong execution, customer trust, and a product that meets the bar of the world's largest cloud provider. It's also meaningful validation of the runtime-first approach we believed in from the start. Proud to support Amiram Shachar and the team as they continue to raise the bar in cloud security.

We’re proud to continue partnering with Matia as the company announces its $21M Series A. From the start, we believed the modern data stack had become too fragmented. Data teams managed too many tools across the full data stack. And that was creating complexity, cost, and a lack of trust in the data itself. Matia was built around a clear bet on consolidation, combining ETL, observability, reverse ETL, and catalog into a single, tightly integrated platform. That approach closely aligned with our thesis, which has proven true far faster than we could have imagined. Matia grew 10x over the past year driven by the most innovative and demanding customers in the world, like Ramp, Drata, Lemonade, and Rogo. And as AI systems now move into production, having reliable data has become more important than ever. Building on the foundation they've established, the team is developing a natural-language platform designed to support what they call the "AI data engineer," enabling teams to easily build data pipelines and validate data. Think 𝑪𝒖𝒓𝒔𝒐𝒓 𝒇𝒐𝒓 𝒅𝒂𝒕𝒂 𝒆𝒏𝒈𝒊𝒏𝒆𝒆𝒓𝒔. Congrats to the entire Matia team on this milestone. It’s been a privilege to work alongside Benjamin Segal and Geva Segal. We're excited to continue the journey and see what you build next. More on the announcement (links in comments): 👉 Gideon Hayden reflects on why we backed Matia at Seed and why we’re leaning in at Series A 👉 Ben Segal shares Matia’s journey and what’s coming next

🚨 Two years after coming out of stealth, Upwind Security has become the unicorn that solves cloud security better than anyone else. 🚨 A snapshot of the response to Upwind's Series B, including thoughtful coverage on the team and growing validation of a runtime-first approach to cloud security. 🎉

We’re proud to participate in Upwind’s $250M Series B and to continue supporting a team that’s turning vision into real momentum. From the start, we believed cloud security would move toward runtime. Static, point-in-time scans create noise and fall short in dynamic cloud and AI-driven environments that demand real-time context. Upwind Security was built for this shift, focusing on what’s actually running, reachable, and relevant in production. Leading enterprises agree, with customers like Siemens, Nubank, Peloton Interactive, and Roku adopting the platform. We partnered with Upwind at the Seed stage, and it has been a privilege to work alongside Amiram Shachar and the team as the company scaled with precision and discipline to achieve meaningful results: 🎉 900% year-over-year revenue growth 🎉 200% logo growth 🎉 4,000% overall growth in just two years This new round of funding fuels the next phase, accelerating product development, expanding go-to-market efforts, and continuing to invest deeply in customer support. Congrats to the entire Upwind team on this incredible milestone. We're excited to continue the journey as you shape the future of cloud security. More on the announcement (links in comments): 👉 Gideon Hayden reflects on working with Amiram over the years and why Leaders Fund continues to double down on Upwind. 👉 Amiram talks us through Upwind's journey and what's next.

Congratulations to Thiago Da Costa, Dov Amihod, and the entire Datagrid team on your acquisition by Procore Technologies. In a few short years since we led the Series A, you've been able to build an incredible platform that rapidly achieved something all startup entrepreneurs are after: market pull! As a result of your maniacal focus on bringing the promise of AI to heavy industries and enabling your customers to see transformative results, you delivered 500% ARR growth in 2025, and became a must-have acquisition. Now, as part of Procore, we wish you continued success as you bring more innovation and impact to the industry. Thank you for delivering impressive returns for our investors :) We look for exceptional founders building must-have AI solutions and support them in building enduring companies. If this is you, we'd love to chat.

🚨 Major launch announcement 🚨 Despite all the AI tools still available, many teams still spend a surprising amount of time on manual, repetitive work. With Vellum, teams can easily build an agent (in plain language) and offload that work. In other words, Vellum gives teams time back to focus on work that actually moves the business forward. Proud to support the Vellum team as they bring this to market.