Upwind reels in $300M for its automated cloud security platform
SiliconANGLE Maria Deutscher
Upwind just raised $300 million more for cloud security. Its pitch: keep up with fast-changing systems by mapping them every 30 seconds.
Based on reporting by SiliconANGLE, Maria Deutscher — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Upwind Security has pulled in another $300 million, just eight months after its last nine-figure round. Bessemer and TCV led the Series C, with Salesforce Ventures, Greylock, Craft Ventures and others joining in. CTech reported the deal on Wednesday. The new money puts the company at a $3.8 billion valuation, which is $2 billion higher than at the start of the year.
The bet here is simple: cloud environments move constantly, and security teams are always behind. Developers add and remove workloads. Instances change. User accounts change. Policies that looked fine a minute ago can already be stale. Upwind’s platform is built to keep that picture fresh, automatically mapping cloud assets and updating the view every 30 seconds.
To do that, Upwind leans on Linux and eBPF, a feature that lets observability code run in an isolated sandbox instead of forcing risky kernel changes. The company says its system collects data on instances, encryption keys, configuration scripts and other cloud assets, then shows how those pieces connect. One example: administrators can see which documents a data visualization app can reach.
The company is also pushing hard into AI security. Its platform can generate AI-BOMs, or bills of materials for AI applications, so teams can spot risky software more easily. It scans cloud assets for weak points, checks configurations against security best practices and regulatory requirements, and runs simulated attacks against important targets, including a possible connection attempt to a cloud instance holding a SQL database.
Upwind says it uses an AI agent called Red to filter out low-risk flaws and another called Blue to spot hacking attempts. It’s not saying how it will spend the new cash, but the recent rollout of about a half-dozen third-party cloud integrations points to the obvious direction: more coverage, more systems, more surface area to police. That’s the whole business now — not just finding problems, but staying close enough to the churn to notice them first.
My take — AI-written commentary, not fact-checked reporting
Cloud security vendors keep selling the same promise in different costumes: we’ll see everything, all the time. Upwind’s version is more credible than most because it’s built around change, not static snapshots, and because 30-second refreshes sound like a response to reality rather than a marketing slogan. The hard part isn’t finding more dashboards; it’s admitting the cloud never sits still long enough to be managed by hand.
Read more about this at: SiliconANGLE
