Adaptive Security has raised $146.5M in under two years to fight AI-powered cyber threats, with backers that include OpenAI, NVIDIA, and Andreessen Horowitz. That kind of cap table is a signal, not a coincidence. It says AI-driven fraud is no longer a niche security concern — it’s a board-level risk that’s moving as fast as the underlying models.
If you’re building anything that touches payments, customer data, or internal workflows, this isn’t an abstract “future of AI” story. It’s a story about how quickly the threat surface is changing, and how fast a new category — human-layer security — is being built to cover the gaps.
From Marketing Tech to AI-Driven Security: The Founders’ Bold Shift
Brian Long and Andrew Jones are not security lifers. They’re the team behind Attentive, the SMS marketing platform that scaled into a unicorn by helping brands turn text messages into a high-conversion growth channel.
They know how to ship product fast, sell into the enterprise, and build high-volume communication systems. That background matters, because the new wave of AI-powered cybersecurity problems is happening in the same channels they already understand: email, SMS, voice, and chat.
After Attentive, they could have done another marketing or sales tool and had a pretty easy time raising money. Instead, they pivoted hard into cybersecurity, focusing specifically on AI threats that target people rather than infrastructure. According to their own account of the shift from marketing to cybersecurity, the catalyst was watching AI make social engineering dramatically cheaper, faster, and more convincing.
Adaptive Security was founded in 2023 and publicly launched in 2025. In that short window, they went from idea to a platform used by hundreds of enterprises facing real AI fraud attempts in production.
That speed isn’t just about founder pedigree. It reflects how quickly AI attacks have gone from “interesting demos” to “we just wired money to a fake CFO” incidents that CISOs have to explain to their boards.
Why Deepfakes and Voice Impersonations Are Breaking Traditional Security
The attack pattern is shifting from “click this bad link” to “talk to this fake human.” Deepfake phishing, voice impersonation of executives, and multi-channel social engineering are now cheap to run and hard to spot.
Think about a coordinated attack where a target gets: a spoofed email from the CEO, a Slack DM from a “colleague,” and a live phone call with a voice that sounds 95% like the real CFO. That’s no longer sci-fi — it’s what large enterprises are actually reporting.
Legacy security controls were never built for this. Email gateways can’t evaluate the emotional tone of a conversation thread. MFA doesn’t help when the attacker convinces a real employee to approve a real transaction. Training people to “hover over links” doesn’t work when the main payload is a convincing voice on a call.
Brian Long has been blunt about this shift, warning that AI impersonations are becoming an everyday occurrence, not an edge case. When anyone can spin up a cloned voice or a realistic video with a few minutes of source material, the old mental models for “this seems off” start to fail.
This is where the idea of human-layer security comes in. Instead of just scanning packets or URLs, tools like Adaptive are trying to understand the patterns of human interaction — who usually talks to whom, from where, about what, and with what timing — and flag when something looks off, even if the content looks polished and “professional.”
Investor Confidence Reveals a New Security Frontier for Founders
Follow the money and you see how seriously the market is taking this. Adaptive Security has raised a total of $146.5M, including an $81M Series B to “stop AI-powered cyber threats,” as detailed in their funding announcement and coverage from fintech and security analysts.
Investors like OpenAI and NVIDIA don’t need more generic cybersecurity exposure. Their bet here is specific: the human side of security is now the weak link, and AI is making that weakness exponentially worse. If AI can convincingly mimic your leadership team, your customers, or your vendors, then “who is this person really?” becomes a core security question.
Adaptive reports more than 500 enterprise customers already on the platform, with strong retention and high NPS. You don’t get that kind of traction in security unless customers are feeling real pain and seeing real value quickly.
For founders, this is the wake-up call. AI fraud risks are not just a CISO problem or a “once we’re bigger” problem. If your product moves money, controls access, or mediates communication between humans, you’re now in the blast radius of these attacks.
Boards are starting to ask: how are we protecting our people from AI-enabled social engineering? Not just our servers, not just our S3 buckets — our finance team, our sales reps, our support agents. That’s the new frontier Adaptive is selling into, and the market response suggests it’s going to be a durable category, not a fad.
What Adaptive Security’s Rise Means for Startup Leaders Facing AI Risks
The story here isn’t “wow, big funding round.” It’s that in roughly 24 months, AI took social engineering from a manual, one-off craft to something scalable and automated — and a new company built an entire AI-powered cybersecurity stack around that shift.
As a founder, you don’t need to become a deepfake expert, but you do need to assume that AI impersonation and multi-channel fraud will touch your company sooner than you think. That means treating human-layer security as part of your core risk model, not as optional awareness training once a year.
Adaptive’s rise is a proof point that the threat is real and the market is moving. A team that previously built marketing tech saw the opportunity, pivoted, and built a fast-growing security company around a problem that barely existed five years ago.
The practical takeaway: when you think about AI in your business, don’t just think about features and productivity. Think about how AI will be used against your team, your customers, and your workflows — and what it would look like to adapt as quickly as the attackers are.
