The pattern is the same. Different technology. Different cycle.
In 2020, everyone was building in Web3. The technology was genuinely new. The potential was real. And the marketing ecosystem that grew up around it was, by and large, a disaster.
Projects with legitimate utility drowned in a sea of identical messaging. Every protocol was "revolutionizing finance." Every NFT project was "building community." Every Layer 2 was "scaling the future." The market got overwhelmed, then skeptical, then hostile.
The products that survived weren't always the best ones. They were the ones that found a way to be understood.
Now look at AI in 2025.
Genuine technology. Real potential. And a marketing ecosystem that's making almost identical mistakes.
Every AI startup is "transforming workflows." Every foundation model is "changing the way we work." Every vertical application is "the future of" whatever industry it's targeting.
The terminology is different. The pattern is identical.
What happened in Web3 should be instructive here.
The market didn't turn against the technology. It turned against the communication style. Against the vagueness. Against the hype that couldn't be substantiated. Against the demos that were impressive but didn't translate into clear value.
The companies that survived the Web3 winter — and there are good ones that did — shared a few things in common. They knew exactly who they were for. They could explain what they did without a whitepaper. They had built enough trust that when things got hard, people stuck around.
AI companies have a shorter window than they think.
The tolerance for "transformative potential" is thinning faster than most founders realize. Enterprises are starting to ask harder questions. The press is starting to push back. Audiences who got burned in the last cycle are applying that skepticism here.
The market is moving from curiosity to evaluation. And most AI marketing hasn't caught up.
The companies that figure out how to close the gap between what they build and what the market understands — with precision, not with hype — are the ones that will matter in three years.
The ones that don't will become case studies in a familiar pattern.