Insight Partners, one of the largest growth-equity firms in the world, is taking a notably different posture in the artificial intelligence investment race, according to a profile and interview published by TechCrunch. The firm's managing director Deven Parekh spoke candidly about the strategy: while much of the venture and private-equity world scrambles to concentrate capital into OpenAI and Anthropic, Insight is deliberately spreading its bets across the AI landscape rather than doubling down on any single foundation-model player.
To understand why this matters, it helps to appreciate the gravitational pull those two names currently exert on institutional capital. OpenAI's recent funding rounds have drawn in sovereign wealth funds, traditional asset managers, and technology conglomerates at valuations that would have been unthinkable only a few years ago. Anthropic has attracted similar enthusiasm, particularly from strategic investors with cloud-infrastructure interests. The implicit thesis behind those concentrated bets is straightforward: foundation models are the oil wells of the AI era, and whoever controls the most capable general-purpose models will extract the most value from everything built on top of them. That logic has proven persuasive enough to reshape the portfolios of firms that built their reputations on disciplined diversification.
Parekh's counterargument, as TechCrunch reports it, is essentially that nobody actually knows how the stack will settle. The history of platform transitions offers some support for this skepticism. During the early internet era, investors who bet everything on browser makers or portal companies often fared worse than those who spread exposure across infrastructure, applications, and enabling tools. The firms that commanded the most attention and capital at any given moment were rarely the ones that generated the most durable returns. A similar dynamic played out in cloud computing, where the early hype around pure-play infrastructure providers eventually gave way to a recognition that the value was migrating to the application and integration layers.
Insight's position is also shaped by its size. Managing a fund of roughly ninety billion dollars, as TechCrunch notes, creates different constraints than managing a two-billion-dollar venture fund. At that scale, taking a meaningful ownership position in a company already valued in the hundreds of billions is mathematically difficult to make accretive to the portfolio. Concentration at the top of the market, in other words, is partly a structural luxury that smaller, more agile funds can afford in ways that Insight cannot. Parekh's stated comfort with holding stakes across competing AI labs, including ones that might be positioned as rivals, reflects a portfolio-construction logic more than a philosophical neutrality about which technology will win.
The mention of Legora, the legal-tech AI startup that TechCrunch reports went to General Catalyst rather than Insight, is worth pausing on. Losing a competitive deal to a firm as well-resourced as General Catalyst is hardly unusual in today's market, but the fact that Parekh discussed it openly suggests a degree of comfort with the firm's own lane. General Catalyst has made no secret of its ambitions to be a more operationally involved, higher-conviction investor in the AI wave. The contrast between the two firms' approaches is a useful lens: one concentrating bets on the belief that picking winners early is paramount, the other treating the current moment as sufficiently uncertain to warrant breadth.
The likely consequences of Insight's approach will take years to fully evaluate, but the near-term implications are already visible in how the firm is perceived by founders. A diversified investor is, by definition, not going to offer the same signaling value to a startup that a concentrated true-believer can. Founders raising from firms that have made OpenAI or Anthropic central to their identity are essentially borrowing some of that halo. Insight's pitch to founders is more transactional and arguably more honest: the firm believes the AI transition will create value broadly, it wants exposure to that broad value creation, and it is not making theological commitments about which layer of the stack will prove most durable.
For the wider industry, the real question Insight's posture raises is whether the concentrated bets being made elsewhere are the product of genuine conviction or of competitive pressure among funds that fear being left out of the defining technology cycle of the decade. The two motivations can produce identical behavior in the short run and very different outcomes over a longer horizon.
What to watch for next: whether Insight's diversification thesis gets stress-tested by a scenario in which one or two foundation-model players genuinely do consolidate the market in the way the most bullish investors expect. If OpenAI or Anthropic establishes something approaching a durable platform monopoly, the opportunity cost of Parekh's caution will be measurable and visible. Conversely, if the application and tooling layers generate returns that rival or exceed those of the model providers themselves, Insight's restraint will look prescient. The answer is probably still several years away, which is precisely what makes the current moment so interesting to watch.
