Insight Partners' Deven Parekh: Why $90B firm won't bet the farm on OpenAI

Insight Partners managing director Deven Parekh is making a contrarian bet: while much of venture capital piles into OpenAI and Anthropic, his $90 billion firm is deliberately staying diversified. In a candid interview with TechCrunch, Parekh opened up about losing portfolio company Legora to General Catalyst, why he is comfortable holding stakes in rival AI labs, and why he believes a broad-based approach beats going all-in on a handful of frontier model makers. The message lands at a moment when the AI funding race has become a winner-take-most narrative, with mega-rounds and strategic alliances reshaping the industry. Parekh's stance is a reminder that not every large investor is chasing the same concentration risk, and that diversification remains a viable - if less flashy - strategy for capturing AI upside. For founders and fund managers watching the AI wave, the question is whether Parekh's patience will look prescient or like a missed opportunity as the market consolidates around a few dominant players. The first thing to understand is that Parekh is not anti-AI. He is simply skeptical of the idea that the only way to win is to pour billions into the two most prominent frontier labs. Instead, Insight Partners is spreading its bets across a range of AI-enabled companies, infrastructure plays, and application layers, betting that value will accrue across the stack rather than concentrate in a single model provider. That approach stands in sharp contrast to the strategy of firms like General Catalyst, which recently poached Legora, a legal AI startup, from Insight's portfolio. Losing a high-profile company to a rival is never easy, but Parekh framed it as part of the natural churn of venture investing rather than a sign of weakness. He also acknowledged that holding stakes in competing AI labs is not a contradiction, but a hedge: if the market ultimately crowns one model king, Insight wants exposure to that outcome, and if it does not, the firm still benefits from a diversified portfolio. The broader context matters here. The AI industry is in the middle of a massive capital supercycle, with OpenAI and Anthropic raising billions at valuations that would have been unthinkable a few years ago. Microsoft, Amazon, and Google have all made strategic bets on these labs, effectively tying their cloud businesses to the success of frontier models. For many VCs, the fear of missing out on the next trillion-dollar platform shift has overwhelmed traditional risk management. Parekh's comments suggest a different calculus: instead of trying to pick the single winner in a highly uncertain race, Insight is building a portfolio that can thrive under multiple scenarios. That is a classic venture approach, but it feels almost radical in the current environment, where concentration has become the default for many growth-stage investors. There is also a practical dimension to Parekh's strategy. Frontier model development is extraordinarily capital-intensive, and the returns are far from guaranteed. Regulatory scrutiny is rising, with governments in the US, EU, and elsewhere examining AI safety, competition, and the concentration of power in a handful of companies. A diversified approach insulates Insight from some of that regulatory and technological risk. If a major lab stumbles, or if regulators force structural changes, a broad portfolio is more resilient than a concentrated bet. Parekh's willingness to hold stakes in rival labs also reflects a pragmatic view of the market: the AI ecosystem is still young, and the eventual winners may not be the companies that dominate the headlines today. For decision-makers, the implications are clear. The AI race is not just a technology story; it is a capital allocation story. Boards and CFOs should ask whether their own AI strategies are overly concentrated in a single vendor, model, or partnership. The same logic that applies to Insight Partners applies to enterprises: diversification is a hedge against uncertainty, and the cost of being wrong on a single bet can be catastrophic. Parekh's comments are a useful counterweight to the prevailing narrative that the only rational move is to back the biggest names in AI. He is not predicting that OpenAI or Anthropic will fail; he is simply saying that the future is too uncertain to bet the firm on any single outcome. That is a message worth hearing for anyone making large technology bets in 2025. The strategic stakes for peers are significant. If Insight's diversified approach pays off, it will validate a more measured path through the AI boom. If it underperforms, it will be another cautionary tale about the dangers of being too cautious in a transformative moment. Either way, Parekh has drawn a clear line in the sand: his firm will not be swayed by the herd. For founders, the takeaway is that not every investor is chasing the same AI moonshot, and that there is still room for a thesis built on breadth rather than concentration. For investors, the lesson is that the AI market is not a monolith, and that the best portfolio may be one that reflects the full range of possible futures. As the capital supercycle continues, Parekh's willingness to speak openly about his strategy is a rare moment of clarity in a market often driven by hype. The next few years will determine whether his patience is a strength or a missed opportunity, but for now, he is betting that the smartest position in the AI race is not to pick a single horse, but to own the whole track.