Steve Blank has never been one to sugarcoat things. The man who helped invent the Lean Startup methodology, who built and sold eight companies, who has taught entrepreneurship at Stanford, Berkeley, and Columbia for two decades — he’s now telling founders something they don’t want to hear: most of them are building companies that were doomed before they wrote their first line of code.
In a new essay published on his blog, Blank lays out a blunt, data-informed argument that the vast majority of startups fail not because of bad execution, underfunding, or tough competition, but because founders skip the most fundamental step in company building. They never seriously validate whether anyone actually wants what they’re making.
That sounds obvious. It isn’t.
Blank’s argument is sharper than the familiar “talk to your customers” advice that has become startup gospel. He contends that an entire generation of founders has absorbed the language of customer discovery without practicing the discipline. They conduct a handful of interviews, hear what they want to hear, build a minimum viable product, launch it into the void, and then wonder why nobody shows up. The process has become performative. A checkbox exercise designed to satisfy investors rather than to genuinely test whether a real market exists for a real product solving a real problem.
“The number one cause of startup failure isn’t running out of money,” Blank writes. “It’s building something nobody wants. And the reason founders build something nobody wants is that they never did the hard work of finding out what people actually need.”
He’s not wrong on the numbers. CB Insights’ widely cited post-mortem analysis of startup failures has consistently ranked “no market need” as the leading cause of death, accounting for roughly 35% of all failures. That figure has barely budged in years. Despite all the accelerators, all the pitch competitions, all the lean canvases taped to coworking space walls — founders keep building products for markets that don’t exist.
Blank traces the problem to what he calls the “founder delusion cycle.” A technically gifted entrepreneur has an idea. The idea feels right because it solves a problem the founder personally experiences. The founder talks to friends, colleagues, maybe a few people on LinkedIn. Everyone says it sounds cool. The founder interprets polite enthusiasm as market validation. Money gets raised. Code gets written. A product gets shipped. And then — silence.
The silence is devastating. Not just financially, but psychologically. Blank has seen it hundreds of times across his career as an investor, advisor, and educator. He describes founders who spent two years and burned through a million dollars building something that a few weeks of rigorous customer interviews would have revealed was fundamentally flawed.
So what does rigorous customer discovery actually look like? Blank is specific. It means talking to at least 100 potential customers before writing a single line of code. Not surveys. Not focus groups. One-on-one conversations where the founder asks open-ended questions about the customer’s problems, workflows, and existing solutions — without ever pitching the product idea. The goal is to understand the customer’s world so deeply that the right product becomes self-evident.
That takes time. Weeks. Sometimes months. And it requires a kind of intellectual humility that doesn’t come naturally to most entrepreneurs. Founders are, by temperament, builders. They want to make things. Sitting in coffee shops asking strangers about their pain points feels like stalling. It feels unproductive. But Blank argues it’s the most productive thing a founder can do, because it’s the only activity that reduces the existential risk of building something nobody wants.
This message is landing at a particularly fraught moment for the startup world. Venture capital funding, while recovering somewhat from the trough of 2023, remains well below the peaks of 2021. Investors are pickier. Runways are shorter. The margin for error has compressed dramatically. A startup that burns six months building the wrong product may not get a second chance.
And the rise of AI has made the problem both better and worse. On one hand, tools like GPT-4, Claude, and open-source models have radically reduced the cost and time required to build software prototypes. A solo founder can now build in a weekend what once took a team of engineers months to produce. That’s genuinely powerful. But it also means the temptation to skip customer discovery has never been greater. Why spend weeks talking to people when you can just build the thing and see what happens?
Blank addresses this directly. “AI makes it faster to build the wrong thing,” he writes. “Speed without direction is just a more efficient way to fail.”
That line has been circulating widely on X since the essay’s publication, resonating with founders and investors who’ve watched the AI gold rush produce a flood of products searching for problems. The pattern is familiar to anyone who lived through the mobile app boom of 2010-2013 or the blockchain frenzy of 2017-2018. New technology creates new capabilities. Founders build around the technology rather than around customer needs. Most of what gets built dies.
The essay also takes aim at the venture capital industry itself. Blank argues that VCs have become complicit in the founder delusion cycle by funding teams based on pedigree, pitch decks, and TAM calculations rather than evidence of genuine customer demand. He points to the common practice of raising a seed round on nothing more than a slide deck and a demo, with no proof that paying customers exist or that the problem being solved is painful enough to change behavior.
“Investors love to say they back founders, not ideas,” Blank writes. “But backing a founder who hasn’t validated their market is just gambling with extra steps.”
This critique won’t be popular on Sand Hill Road, but it’s hard to argue with given the data. The Kauffman Foundation has published research showing that the majority of VC-backed startups fail to return their invested capital. The industry’s returns are driven by a tiny number of massive winners — the Googles, the Ubers, the Airbnbs — while the vast middle of the portfolio quietly goes to zero. Blank’s argument is that better pre-investment customer discovery could meaningfully shift those odds, not by eliminating failure but by eliminating the most preventable kind of failure.
He’s also pushing back against a cultural tendency in Silicon Valley that valorizes speed above all else. Move fast and break things. Ship it. Iterate. These mantras have produced extraordinary companies, but they’ve also produced an enormous amount of waste — wasted capital, wasted talent, wasted time. Blank isn’t arguing against speed. He’s arguing against speed without foundation. Build fast, yes. But build the right thing.
There’s a tension here that Blank acknowledges. Customer discovery is inherently slow. Markets are messy. Customers often can’t articulate what they want. And sometimes — sometimes — a visionary founder really does see something the market can’t yet imagine. Steve Jobs didn’t do customer interviews before building the iPhone. But Blank’s counterargument is that for every Steve Jobs, there are ten thousand founders who thought they were Steve Jobs and weren’t. The base rate matters. And the base rate says: validate first.
The essay has drawn sharp reactions from both sides. Some founders on X have praised it as a necessary corrective, sharing their own stories of products that failed because they skipped the validation step. Others have pushed back, arguing that Blank’s framework is too conservative for markets that move at AI speed, where being first matters more than being right. One prominent angel investor posted: “If you spend three months doing customer discovery in AI right now, you’ll come back to find ten competitors already in market.”
That’s a fair point. But Blank would likely respond that being in market with a product nobody wants isn’t a competitive advantage. It’s just burning cash publicly instead of privately.
The broader context for this debate is a startup market that’s undergoing significant structural change. The cost of building software has collapsed. AI coding assistants like GitHub Copilot, Cursor, and Replit’s AI tools have made it possible for non-technical founders to build functional prototypes. The barrier to entry has never been lower. But the barrier to building something people will pay for hasn’t changed at all. If anything, it’s gotten higher, because the flood of new products has made customers more skeptical, more overwhelmed, and harder to reach.
Blank’s prescription isn’t complicated. Talk to people. Listen. Don’t pitch. Ask about their problems, not your solution. Do this a hundred times. Look for patterns. Then — and only then — start building. It’s the same advice he’s been giving for twenty years. But the fact that he feels compelled to say it again, more forcefully than ever, suggests that the message still hasn’t sunk in.
Maybe it never will. Entrepreneurship attracts optimists, and optimists are constitutionally inclined to believe their idea is the exception. That’s what makes them founders. It’s also what kills their companies.
The essay closes with a challenge to the startup community that’s worth quoting in full: “If you can’t find 10 people who will pay you before you build your product, you don’t have a business. You have a hobby. And hobbies are fine — but don’t raise venture capital for your hobby.”
Harsh. But given that roughly 90% of startups still fail, and that the single biggest reason remains building something nobody wants, maybe harsh is exactly what the moment requires. Blank isn’t trying to discourage entrepreneurship. He’s trying to make it less wasteful. Less heartbreaking. Less dumb.
Whether founders will listen is another question entirely. History suggests most won’t. They’ll read the essay, nod along, and then open their laptops and start coding. But for the few who take it seriously — who put down the keyboard and pick up the phone — the odds shift dramatically. Not to certainty. Never to certainty. But away from the kind of preventable, predictable, dead-on-arrival failure that Blank has spent his career trying to eliminate.
That might be the most important thing a founder can do in 2026. Not build faster. Not raise more. Just ask the question that most are too afraid or too impatient to ask: Does anyone actually want this?
WebProNews is an iEntry Publication