TOMER WEINGARTEN

SentinelOne's early architecture reflected Weingarten's engineering instincts in a way that matters: the endpoint protection was designed from day one to work even when the cloud was unavailable. Most SaaS security companies assumed uptime. Weingarten designed for failure. "You have to make sure that you're prepared to fail," he said, "and that when you fail, you fail gracefully." It reads like a management aphorism. It was an architectural constraint that saved customer networks when competitors' cloud services went dark.

The early proof came in an unexpected way. During a beta proof-of-concept with a major streaming entertainment company, an executive wandered into a Forbes interview and described the legacy antivirus category in terms that editors cleaned up for publication. His actual message: "these things are going away." He pointed to SentinelOne as the replacement. Weingarten hadn't planned the press hit. The inbound demand that followed was something he had to build fast enough to keep up with.

The IPO in June 2021 was a moment of unusual specificity. SentinelOne raised $1.2 billion. The implied valuation was $8.9 billion - surpassing CrowdStrike's own record-setting debut from 2019 to become the highest-valued cybersecurity IPO in history. The stock rose 21 percent on the first day of trading on the NYSE under the ticker "S." That single letter - confident, almost terse - fits. On the morning SentinelOne went public, Weingarten had already been building the next phase for two years.

The next phase is AI. Not the marketing department's version - bolt-on chatbots and press releases about generative AI. Weingarten's view on the noise is characteristically uncharitable: "There is not a single LLM in the world today that is secure by any degree." He believes the real moat in AI-powered security is foundational models trained on real, curated telemetry data. Not ChatGPT integrations. SentinelOne's Purple AI, launched in 2023, is the first commercial product in that thesis - an agentic security analyst that can investigate threats, correlate across the Singularity data lake, and respond autonomously.

At OneCon 2025, Weingarten stood on stage and told the industry its entire premise was broken: cybersecurity today is reactive, and reactive is unacceptable. The solution he outlined was one where AI and human analysts operate together in a proactive and predictive loop - not waiting for breaches, but anticipating them. He then went to RSA Conference 2026 and made the claim bigger, arguing that the future of cybersecurity had become inseparable from the future of civilization itself.

The personal cost of building a company of that magnitude is rarely polished. In 2021, a legal settlement required Weingarten to transfer approximately $124 million in SentinelOne shares to his ex-wife - one of the largest executive divorce settlements in Israeli tech history. He showed up to work. The observation isn't remarkable because of the amount; it's remarkable because it was the same year the company went public, and the company got better.

His philosophy of leadership - built on listening before proposing, on ego-free culture, on treating analyst skepticism as fuel - runs counter to the style that gets CEOs profiled in soft-focus magazine features. Weingarten's interest is in the problem. The competition is a useful irritant. When asked about Microsoft's growing dominance in enterprise security, his answer was pointed: "The power that Microsoft has is very corruptive." He didn't mean Microsoft is corrupt. He meant concentrated power without accountability corrupts an entire market's incentives. It's a distinction that requires actually understanding the market, not just selling into it.

SentinelOne now serves more than 11,000 organizations globally. It has led the MITRE ATT&CK evaluations five consecutive times. It crossed $1 billion in ARR in 2024 - a milestone Weingarten announced with four words on LinkedIn: "A historic moment." The understatement, deliberate or not, is consistent with how he has always talked about the company. Not the biggest. Not the loudest. Just the one that works.

There's a version of this story where Tomer Weingarten is the guy from Petah Tikva who got lucky when antivirus became a punchline. That version is wrong. The version that holds is the one where an engineer who learned to listen, built for failure, and refused to confuse activity with progress turned a second-grade friendship into the category-defining platform for AI-native enterprise security. The computer he took apart in Petah Tikva never got put back together. Neither did the industry that thought it had already solved the problem.