The company used an AI-native platform to help companies fight threats.
What depthfirst is building
depthfirst has raised a $40 million Series A to expand an AI-native security platform aimed at helping companies fight threats. The framing matters: "AI-native" implies the machine learning is part of the core detection and response loop rather than a feature added to an existing rules engine. That distinction shapes how the product ingests data, what it can flag, and how much manual tuning a security team has to do.
For a firm at the Series A stage, funding at this level usually buys three things: engineering headcount to harden the platform, a larger data and threat-research effort, and a go-to-market team to move from early design partners to paying customers. Each of those is a bet that the AI approach detects real attacks earlier than the tools companies already run.
Why "AI-native" changes the detection model
Traditional security tooling leans on signatures and hand-written rules: you describe a known bad pattern, and the system alerts when it sees a match. That works well for threats you have already catalogued and poorly for novel or slow-moving ones. A platform built around models instead learns what normal looks like across logs, network traffic, and user behavior, then surfaces deviations that no one wrote a rule for.
The tradeoff is that anomaly-driven systems can generate noise. The engineering challenge behind a product like depthfirst's is not just catching more, but ranking findings so analysts spend time on the alerts that matter. Buyers should press on how the platform scores confidence, how it explains why something was flagged, and how it avoids drowning a team in low-value alerts.
How to evaluate a platform like this
If you are considering an AI-native security vendor, treat the AI as a claim to be tested, not a guarantee. A short structured evaluation tells you more than a demo.
- Run it against your own telemetry, not a curated sample, and measure both what it catches and what it misses.
- Ask how the models are trained and updated, and whether your data stays isolated from other customers.
- Check what happens when the AI is wrong: how easy is it to suppress a false positive without blinding yourself to a real one.
- Confirm it integrates with the systems you already use for alerting, ticketing, and response.
- Look at the human workflow — the goal is to make analysts faster, not to replace judgment they still need to apply.
What the funding signals for buyers
A Series A means the company has enough traction to convince investors but is still early. For a prospective customer, that carries both upside and risk. You may get responsive support, direct access to the engineering team, and influence over the roadmap. You also take on the possibility that features are still maturing and that the product will change as the company grows into its funding.
The practical move is to engage as a design partner rather than a passive buyer: define the threats you most need covered, agree on how success will be measured, and revisit that measurement after a real evaluation period. An AI-native platform earns its place by reducing the time between an attack starting and your team knowing about it — and that is something you can test directly rather than take on faith.