Analysis
At TechCrunch Disrupt this week, Benchmark partners Eric Vishria, Chetan Puttagunta and Peter Fenton acknowledged on stage that the AI funding boom is straining one of venture capital's oldest informal rules: don't back two startups chasing the same customer. None of them declared the norm dead, but none defended it as fully intact either, according to TechCrunch.
The clearest evidence: OpenAI and Anthropic, the two biggest frontier AI labs, now share dozens of the same investors — a level of overlap that would have been unusual in earlier venture cycles, when firms more often picked a side in a category and stayed there. Benchmark itself has backed companies across competing corners of the AI stack, and Vishria argued that's simply the current dynamic: "you are going to get conflicts." He said the firm stays committed to each founder, keeps information confidential, and isn't acting as "product or technology experts" picking winners.
“Degnan said he had already resigned and denied sharing confidential information; he later joined Cognition as chief revenue officer.”
The conflict isn't theoretical. AI coding-agent startup Factory fired board observer Chris Degnan in late September, alleging "unethical conduct involving Cognition," a rival maker of the Devin coding agent. Degnan said he had already resigned and denied sharing confidential information; he later joined Cognition as chief revenue officer. Khosla Ventures, which backs both Factory and Cognition, saw founder Vinod Khosla weigh in publicly on X, calling Factory "a struggling second tier competitor" — the kind of public commentary firms used to keep private.
Puttagunta's framing is the more useful diligence lens: AI is accelerating the collision because founders now see a larger addressable market and are more likely to expand into territory a portfolio company already claimed, rather than starting as head-to-head rivals on day one. Fenton called the whole issue the firm's toughest problem, describing conflicts as "kryptonite" rather than a solved process question.
The counterweight is obvious and worth saying plainly: firms have every incentive to describe conflicts as inevitable rather than admit they chase every hot category regardless of existing portfolio commitments. Founders raising in AI right now should treat "we handle conflicts case by case" as a starting point for questions, not a reassurance — ask what specifically changes once your lead investor backs a direct competitor, and get it in writing rather than relying on partner-level discretion.