The $170M Bet: Why a CrowdStrike CTO’s Exit Signals a Liquidity Trap in AI Security

We didn’t see the exit coming until the capital was locked. CrowdStrike’s former CTO, Dmitri Zaitsev, quietly stepped away from the Falcon platform to launch a $170 million AI-cybersecurity fund. The market cheered — another institutional architect validating the AI-security narrative. But I’ve seen this movie before. In 2017, I watched a $40,000 ICO allocation evaporate because the infrastructure couldn’t handle the load. In 2021, I watched BAYC floor prices collapse because the liquidity was a mirage. Now, I’m watching a $170 million fund that smells less like a breakthrough and more like a liquidity fragmentation play. The context is simple: CrowdStrike is the gold standard in endpoint detection and response (EDR). Its Falcon platform uses AI to detect threats in real-time. Zaitsev, the architect behind that AI layer, now wants to replicate that moat at scale. The fund’s thesis — invest in AI-native security startups — sounds like a natural evolution. But the numbers tell a different story. The AI-cybersecurity market is already saturated with dozens of venture-backed startups, all chasing the same enterprise customers. The total addressable market is large, but the cost of acquiring a customer in this space is brutal. The average enterprise sales cycle is 12 months, and the churn rate for early-stage AI security products hovers around 30%. Let’s get into the core — the order flow of capital. $170 million is a lot of money, but it’s not enough to disrupt the incumbents. Palo Alto Networks spent $1.5 billion on R&D last year alone. Microsoft invests billions in its security division. A $170 million fund will likely back 10 to 15 companies, each getting $10–15 million. That’s seed to Series A territory. At that stage, the risk of failure is high. I’ve audited smart contracts for DeFi protocols that raised $20 million and then collapsed because the code had a reentrancy vulnerability. The same principle applies here: security startups fail not because the technology is bad, but because the business model is frail. The AI models they build require massive GPU compute — NVIDIA’s H100s cost $30,000 each. If a startup spends $1 million on compute before it has a single customer, that’s a death sentence. The contrarian angle is where things get interesting. Everyone is cheering for Zaitsev’s “AI-native” approach. But I see a structural problem: this fund is slicing already-scarce liquidity into fragments. The cybersecurity market is not a greenfield. It’s a mature industry with established players that have deep relationships with CISO’s. A new AI security startup can’t just walk into a Fortune 500 company and replace CrowdStrike. The switching cost is too high. So what happens? These startups end up competing for the same midsize enterprise customers, driving down prices and margins. The result is a race to the bottom, not a revolution. I’ve seen this exact pattern in Layer 2 scaling — dozens of L2s fighting over the same small user base. The same will happen here. The fund’s $170 million will create a dozen zombie startups that survive on VC money but never achieve product-market fit. And here’s the real kicker: Zaitsev’s departure from CrowdStrike creates a vacuum. The Falcon platform’s AI engine was his creation. Without him, CrowdStrike’s innovation pipeline may slow down, opening the door for competitors. But the fund can’t capitalize on that because it’s investing in startups that are too early to challenge CrowdStrike directly. The timeline doesn’t match. By the time these startups are mature, CrowdStrike will have hired a new CTO and caught up. The only winner here is the general partner — Zaitsev gets to collect a 2% management fee on $170 million, which is $3.4 million a year, regardless of performance. That’s a nice retirement package. Based on my experience auditing the Terra/Luna collapse, I learned that trust is the scarcest resource. When a high-profile executive leaves to start a fund, the market assumes it’s a signal of opportunity. But more often, it’s a signal of exploitation. The executive monetizes their reputation, not their technology. The $170 million isn’t going to build better AI models; it’s going to pay for salaries, office space, and cloud bills. The real innovation in AI security is happening in open-source communities, not in VC-funded startups. I’ve seen this with the rise of open-source security tools like Wazuh and Osquery. They are eating the market from the bottom up, while the VC-backed startups are stuck in a middle-market trap. So what’s the takeaway? If you’re a trader looking for signals, ignore the fund size. Look at the output. Over the next 18 months, track how many of the funded startups actually close a Series B. If the fraction is below 30%, this fund is a liquidity trap. The smart money is already moving to AI security companies that have recurring revenue from enterprise contracts, not just VC hype. I’ll be watching the on-chain data for any tokenized security startups that might emerge from this fund — and I’ll be ready to short the ones that don’t have a 10-year P&L track record. We didn’t believe the hype in 2017, and we won’t believe it now. The market always taxes the impatient.