I recently came across a particularly biting viewpoint from a veteran in the investment world—Naval. On a podcast, he said a line that feels like it’s effectively passed a death sentence on the entire software industry: pure software isn’t worth investing in.

At first it sounds alarmist, but on second thought, he’s right. Naval isn’t the type to talk off the cuff. As the founder of AngelList, he has invested in companies like Twitter, Uber, and Notion over the past twenty years, and his instincts about where capital flows have long been proven. So when he says this, it’s not mere commentary—it’s a formal verdict.

Where does the problem really lie? Look at Apple. The company’s $3 trillion valuation, bluntly speaking, is built on one thing: supporting the profits of premium hardware through a finely crafted software experience. But now that logic is being completely destroyed by AI. Imagine that in the future, when you open your phone, it’s no longer tapping app icons—you just start talking directly with an AI agent, which generates the interface you need in real time. If interfaces can be generated instantly by AI running on any phone, what’s the point of Apple’s carefully engineered app store, design standards, ecosystem lock-in, and all of that?

This reminds me of the Microsoft lesson Naval mentioned. Back then, Microsoft ruled the PC era, but in the mobile era it completely lost—because they refused to build a touch system from scratch. They assumed the old paradigm would continue. In the end, Apple won the next decade. Now Apple is repeating that same mistake: they’ve outsourced the AI experience to Google and handed their most core elements to their competitors. It’s like giving away your lifeline.

For entrepreneurs, the situation is even more severe. Why does your SaaS company exist? Because building the product used to be hard. Why could you raise funding? Because you needed a technically capable team to execute. But that difficulty is collapsing now. With two people using Claude, you can replicate 80% of most B2B SaaS products within three months—not as toy versions, but as genuinely usable products. The remaining 20% may look crucial, but that’s mostly just friction—friction that gets compressed every quarter by the next generation of AI agents.

Look at what’s already happening. Adobe spent $20 billion to buy Figma, claiming it was hard to replicate. Now independent developers can build design tools with 70% of Figma’s functionality in just a few months. Salesforce, the most valuable SaaS company in history, is now being eaten away by AI-native CRMs. Giants like Workday, ServiceNow, and Atlassian are now all potential replacements.

So what can survive? Naval’s answer is: not the software itself, but the things AI can’t replicate. Distribution channels, network effects, data flywheels, hardware integration, branding, communities, regulatory barriers—these are the real moats in the new world.

If your honest answer to “What’s our moat?” is “Our product is better,” then you only have 18 months to find a real moat. Otherwise, at your next round of fundraising, your valuation will evaporate by 70% to 90%. This isn’t alarmism. It’s Naval’s judgment based on twenty years of investing experience.

But there’s another side to this story as well. It’s precisely this massive collapse that has opened a new era for individual entrepreneurs. Software isn’t dead—it’s been democratized. Historically, there have been examples like Notch developing Minecraft entirely on his own, Markus Frind building Plenty of Fish into a business with annual profit in the tens of millions while working solo, and Instagram being acquired by Facebook for $1 billion with only 13 people. These were outliers.

What’s changing now is the ceiling. In the past, solo founders could make interesting things, but when scaling up they hit a hard wall—requiring teams, coordination, compromises, and diluting the vision. Naval’s vision is a one-person company operating at the pace of a 50-person team. Users report bugs, AI agents review them, write fix plans, and the founder approves and releases. Customer support is handled by AI, feature requests are voted on by users, and AI is responsible for building. No internal collaboration friction, no office politics—your ideas go straight from your head to launch, without distortion.

This isn’t theory. Pieter Levels, as an independent operator, has built multiple businesses with seven-figure revenue. More and more independent developers are running businesses that, three years ago, would have required Series A funding to get off the ground. The next billion-dollar company might have only one employee. The next hundred-billion-dollar company might have no more than ten people.

That’s where things stand now: either it’s the worst era ever for building general-purpose software, or it’s the best era ever for building differentiated products—and both are true at the same time.

You have three paths in front of you. First: treat all of this as hype, convince yourself that Apple can’t fail, your company is different, and AI is being exaggerated. Most entrepreneurs will choose this, and most will fail because of it. Second: panic—slash your cash runway fast and pivot in a scramble. That’s what happens when realization comes too late. Third: take the 18-month window seriously. Audit your moat, start building distribution channels, find angles AI can’t copy, and position yourself for the world that’s coming.

Naval isn’t playing it by ear. When he says “pure software isn’t worth investing in,” that’s a judgment, not just an opinion. It’s a conclusion spoken by someone who spent twenty years deciding what’s worth investing in—and is now judging that most of what’s being funded today is no longer worth investing in.

Apple is dead. Most SaaS entrepreneurs are the next wave. And those who survive will be the ones who heard this warning and acted before everyone else caught on. The window is open right now, but it won’t stay open forever. In the next 18 months, are you building a moat that can survive—or are you watching your existing moat erode in real time? Most won’t make it. A few will. The difference comes down to what you do this quarter.