From the outside, Upwind Security seems to have had a smooth ride so far. Just four years old, the cloud security startup is now valued at $1.5 billion, and boasts the likes of Siemens, Peloton, Roku, Wix, Nextdoor and Nubank among its clients. But if you ask the company’s co-founder and CEO Amiram Shachar, the journey to get here was uncertain.

“Three years ago, we spend a lot of time asking ourselves if we’re going in the right direction, and 80% of the time, it feels like we’re not,” Shachar bluntly told TechCrunch in an interview after the startup’s recent $250 million Series B.

“In the beginning, we always asked if the market needed our solution, if it would be too difficult to integrate with larger systems, or if customers would adopt it,” he recalls. “It’s difficult to create a new method; people are used to installing some agents on machines, but they don’t like to do it.”

Upwind likes to call that approach “runtime” security: Prioritizing alerts and remediation efforts around threats and vulnerabilities in active services in real time. As Shachar says, it’s an “inside-out” take on cloud security, where internal signals like network requests and API traffic function as context to help security teams separate immediate risks from those that can wait.

Developing that approach wasn’t easy, however, because Shachar and his co-founders didn’t have a traditional security background: they first built and sold a cloud computing brokerage called Spot.io, to NetApp to nearly $450 million by 2020.

“After joining NetApp post the Spot acquisition, I experienced first hand how difficult cloud security can be,” said Shachar. “The security team will scan our environment and report issues, but they lack critical context. From a DevOps background, we (Shachar and his team) understand the infrastructure very well, while security teams often don’t know how APIs are exposed or which packages are running. As a result, they flag many issues that are not really dangerous.

But Shachar and his team feel they have a better understanding of cloud environments because they run them. “The prevailing approach is agentless, an ‘out-of-the-box’ model where you scan external environments,” he explains. “It’s easy to deploy, but it creates a lot of noise because you only see what’s visible on the outside.”

The team realized that the context provided by internal signals was more useful to security teams, because they could see what was happening on the network, in real time. But marketing their new take on cloud security has proven challenging, as security teams often lack the permission to deploy software internally and default to more traditional tools.

So Upwind sales take time. “It was not clear at first, and there was a lot of uncertainty; customers were hesitant,” said Shachar.

“But we saw something that no one else saw,” he explained. “Inside-out is not an advanced option; it’s the only way to solve the next generation of problems. With ephemeral infrastructure like containers, serverless workloads, AI agents talking to each other, and data moving constantly through APIs, you can’t map it from the outside. It needs to be inside.

However, the company has to contend with a crowded security market. Security teams are already overwhelmed by the number of tools, and customers don’t want multiple products just to manage cloud security. “From the beginning, it was clear that Upwind needed to build a broad, integrated platform,” Shachar said. “Otherwise, customers will not engage or allow us to deploy our technology.”

The company’s logic eventually speaks to its target customers: large, data-intensive organizations with large cloud footprints. Since then $100 million Series A in 2024Upwind grew rapidly, posting 900% year-over-year revenue growth and doubling its customer base. The company is also expanding from its core markets of the US, UK and Israel to emerging markets including Australia, India, Singapore, and Japan.

The $250 million Series B was led by Bessemer Venture Partners, with participation from Salesforce Ventures and Picture Capital. The fresh money will be used for product development and go-to-market activities, and the startup plans to invest in its AI security capabilities within its core cloud security platform and “expand its approach closer to developers to help prevent misconfigurations before they reach production.”