Bold Security Named to 2027 Fortune Cyber 60 in First Year Out of Stealth

Bold Security, a startup building real-time data loss prevention (DLP) for the endpoint, has been named to the 2027 Fortune Cyber 60 list of the most important venture-backed cybersecurity startups. According to the company, it made the list in its first year out of stealth, in the early stage category.
- Bold Security was named to the 2027 Fortune Cyber 60 in the early stage category, in its first year out of stealth.
- The company runs small language models on each device to classify data by meaning and block or coach users in the moment.
- Bold was founded in 2024 by the team behind Polyrize and has raised $40 million from Bessemer Venture Partners, Picture Capital and Red Dot Capital Partners.
- The Fortune Cyber 60 is developed with Lightspeed Venture Partners and AWS, and groups 60 startups into three stages of 20.
What the Fortune Cyber 60 recognizes
The Fortune Cyber 60 is an annual list of venture-backed cybersecurity startups that have not had an IPO, acquisition or other significant exit event. It was developed in partnership with Lightspeed Venture Partners and AWS, and it names 60 companies.
According to the announcement, Lightspeed evaluates hundreds of venture-backed cybersecurity companies on annual recurring revenue, year-over-year growth and validated operating data. It then groups the list into three stages of 20. Bold Security appears in the early stage category.
Rethinking DLP around on-device AI
Bold argues that data loss prevention has long been plagued by noise and false confidence. In the company's description, legacy DLP matches patterns, so it catches PII and PCI data but misses source code, financial models and board documents. The patterns it does match also fire on legitimate work, which floods teams with alerts and leads them to turn blocking off. In that case, the prevention tool becomes a detection tool.
Bold says AI is making the problem harder, because work is moving back onto the device through desktop assistants and AI agents that run locally, outside the paths legacy DLP was built to watch.
The company's answer is to run AI on the device itself. According to Bold, small language models on every device classify data by meaning rather than pattern, read the context around each action, and block or coach the user at the moment it happens. The company says this means no cloud round trip, no regex rules to write and no coverage gap when a device goes offline.
"People and AI agents move sensitive data through the device faster than a security team can follow, and DLP both overburdens and underdelivers," said Nati Hazut, Co-Founder and CEO of Bold Security.
Founders, funding and customers
Bold was founded in 2024 by the team behind Polyrize, the cloud data security company acquired by Varonis. It has raised $40 million from Bessemer Venture Partners, Picture Capital and Red Dot Capital Partners.
The company names Tekion, HiBob, Cresta and Shutterfly, among others, as customers. According to Bold, they use the product to prevent data loss, identify risky insiders and secure AI usage on the endpoint.
Amit Karp, Partner at Bessemer Venture Partners, said: "Every security team knows DLP is a problem, but very few companies have been willing to rethink how it works."
What it means for security teams
The listing is a third-party signal about the company's traction, based on the revenue, growth and operating data criteria described above. For security teams, the more practical point is the approach Bold represents: moving classification and enforcement onto the endpoint as employees and AI agents handle more sensitive data locally.
Bold says its product works from day one with zero configuration and no tuning, and covers any data, not only the patterns regex can catch. Those are the company's own claims, and the announcement does not include independent performance figures.


