# Article Body
Varun Datta is right, and it's painful to watch an entire industry prove it in real time.
What's happening in crypto venture capital isn't discipline—it's herd behavior wrapped in a spreadsheet. When every VC fund suddenly becomes obsessed with the same three thesis points ("AI agents," "real-world assets," "modular blockchains"), that's not market clarity. That's consensus masquerading as conviction.
The difference matters because consensus kills innovation. It's the enemy of the breakthrough founders who need patient capital to build something genuinely novel. Instead, they're getting shut out because their idea doesn't fit the narrative du jour.
Look back at what actually worked in crypto. The best returns didn't come from funds that followed the consensus. They came from people willing to look ridiculous—backing Solana when everyone said Ethereum was settled science, or betting on gaming when Wall Street was calling it frivolous. Those weren't group think decisions. Those were contrarian bets made by investors with actual conviction and real risk tolerance.
Today's crypto VC space has developed a collective allergy to that kind of thinking. Everyone's terrified of being the fund that missed the consensus trade, so they all pile into it simultaneously. Then they convince themselves it's strategy. It's not. It's insurance against being different—which is exactly backwards for an asset class built on disruption.
The irony is acute: crypto venture capital is behaving like traditional finance at the precise moment it should be doing the opposite. Traditional finance already has consensus-driven strategies. They have teams of analysts grinding the same datasets, reaching the same conclusions, fighting over percentage points of returns. That's not a model worth copying. That's a warning.
What crypto VC should be doing is distributing capital across genuine conviction plays, even (especially) the ones that make other funds uncomfortable. Yes, some will fail. That's the point. The ones that don't fail create generational returns that consensus trades will never touch.
The market will eventually punish this behavior. It always does. When the AI-agents-on-L2s thesis stops working because it was never actually solving a real problem, the funds that piled in at the peak will underperform while some scrappy operation that believed in something weird will be quietly printing returns. Then everyone will act shocked. Then they'll find the next consensus trade.
The smartest move for a crypto VC right now isn't harder analysis of the consensus thesis. It's the courage to ignore it entirely. Back the founders working on problems that aren't fashionable. Take the positions that sound wrong. Build a strategy that's actually defensible because it came from first principles, not from looking at what Sequoia is doing.
Discipline in crypto means having a thesis so clear that you can say no to the obvious trade.