When Guy Arazi and Yoni Ramon began pitching their cybersecurity startup last year, they had a provocative opening line: What if your security team could remember every vulnerability it had ever seen—and prevent the same mistakes before a single line of bad code shipped?
That premise, part sales pitch and part technical moonshot, appears to have resonated. On June 10, 2026, the Israeli-American startup Pi announced it had raised $35 million in total funding, emerging from stealth with a product it describes as an "agentic product security platform." The announcement bundled together a $25 million Series A led by Third Point Ventures and a $10 million seed round from Brightmind Partners. Forbes reported the company's valuation at around $100 million, based on sources familiar with the financing.
The round attracted a notable roster of security-industry names. CrowdStrike CEO George Kurtz participated, as did Yevgeny Dibrov and Nadir Izrael, the co-founders of Armis. Recursive Ventures and QP Ventures also joined.
The Pedigree
Arazi and Ramon—both veterans of large-scale security operations—founded Pi in 2025. Arazi spent several years as a security researcher at Microsoft, where he focused on cloud vulnerabilities and contributed to both Defender and Azure's security architecture. Before that, he helped establish XCloud at Palo Alto Networks, a detail the company now highlights in investor materials.
Ramon's résumé reads like a tour of Elon Musk's corporate ecosystem. He led offensive security at Tesla, where his team probed weaknesses in both vehicle systems and the company's robotics infrastructure. Prior roles included security work at SpaceX and Twitter. Forbes, in its June 10 coverage, dubbed him "Elon Musk's go-to hacker"—a characterization the startup has not discouraged. The publication also reported, based on unnamed sources, that xAI is among Pi's early customers, though the company has not confirmed that publicly.
Pi's website currently lists customer logos for Lemonade, Teramind, and Navan. Detailed case studies, however, remain unpublished.
The Product Pitch
At its core, Pi's platform ingests vast amounts of development data: source code, design documents, cloud infrastructure configurations, even internal Slack and Teams conversations. The system then constructs what the company calls "institutional product security memory"—essentially, a knowledge graph of an organization's security posture over time.
The idea is to distinguish signal from noise. Security teams are drowning in alerts, many of them false positives. Pi claims its platform can filter out the irrelevant, surface genuine threats, and—this is the ambitious part—generate contextual fixes automatically. The system is also designed to enforce guardrails, theoretically preventing the same vulnerability from cropping up again six months later.
The platform includes an AI assistant named Sloane (the naming convention feels very 2025-era tech branding) that learns a company's codebase and security history. Pi has shared some eye-catching metrics: 70% reduction in manual triage, 85% fewer repeat bug bounty payouts, 95% of vulnerabilities blocked before production. Those figures come from the company itself and have not been independently verified—standard caveats in early-stage enterprise software.
A Crowded Moment
Pi's emergence coincides with what has become a remarkably crowded field. The first half of this year alone has seen a flurry of similar raises in the "agentic security" category. Above Security announced a $50 million round in March for insider-risk detection powered by autonomous agents. Capsule Security launched in April with $7 million for runtime controls over AI agents. And just days before Pi's announcement, Opal Security disclosed a $23 million raise for identity governance that spans both human users and AI systems.
The underlying bet is the same across all these companies: as AI agents proliferate inside enterprises, the old playbook for securing code and infrastructure won't scale. Whether the market can support this many players remains an open question.
"Pi are fixing the part of the problem the industry has ignored," Stephen Ward of Brightmind Partners said in the company's press release. "They close the loop between finding and fixing." Sapir Harosh of Third Point Ventures added that Pi is "building an agentic system that understands the full development environment and ships fixes autonomously."
Third Point Ventures, which maintains a Tel Aviv office, previously led the Series B for Grip Security in 2023—a signal of sustained interest in Israeli cybersecurity startups, a sector that has long punched above its weight in venture funding.
What Comes Next
As of the June 2026 announcement, Pi employed roughly 23 people split between San Francisco and Tel Aviv. The company has not disclosed specific plans for the capital. In a market this competitive, speed matters—perhaps more than the founders expected when they first started building.
For now, Pi is betting that enterprises are ready to trust an AI system not just to flag vulnerabilities, but to fix them. That's a significant leap of faith, both technically and culturally. Security teams tend to be conservative, for good reason. But if the product delivers on even a fraction of its promises, the payoff could be substantial.
Whether Sloane becomes a household name in security circles—or just another chatbot lost in the noise—will depend on execution in the months ahead.
