Insight Partners Discovers Diversification While Industry Sleepwalks Into AI

Deven Parekh, the head of a $90 billion investment firm, has made a shocking discovery: diversification works. In remarks this week, the Insight Partners executive explained that while the entire venture capital industry has developed what can only be described as acute OpenAI-and-Anthropic tunnel vision, his firm is sensibly holding positions across multiple AI laboratories. One might expect this announcement to be met with universal agreement, perhaps a gentle pat on the back for remembering Portfolio Theory 101. Instead, it reads like a press release from the only sober person at an all-night casino. The tragedy here is not that Parekh is wrong—it's that stating the obvious has become brave.

The context makes this more amusing: Insight Partners recently lost Legora to General Catalyst, a deal that presumably stung enough to warrant public commentary about strategic positioning. Rather than frame this as a loss, Parekh has pivoted to celebrating the firm's deliberate choice to avoid overconcentration in the mega-cap AI narrative. This is the investment equivalent of saying, "We're actually fine with that house we didn't buy, because we own three others." The underlying message is clear: while other firms have bet their entire institutional credibility on the premise that OpenAI and Anthropic will single-handedly reshape civilization, Insight is hedging. Which raises an uncomfortable question for everyone else: why wasn't this obvious?

The genius of this positioning is that it requires no actual contrarian thesis—just basic risk management dressed up as prescience. Parekh isn't claiming that OpenAI or Anthropic will fail, or that some dark-horse AI lab will crush them. He's simply stating that a $90 billion firm can afford to own stakes in rival laboratories without declaring exclusive fealty to the chosen two. This is not a hot take. This is what fiduciaries are supposed to do. Yet in an industry currently exhibiting the collective investment discipline of a sports bar during March Madness, the ability to articulate basic diversification sounds like cutting-edge strategy.

The irony is suffocating. Venture capital has spent the last two years executing the financial equivalent of an all-in poker bet on a single hand—or rather, two hands. The narrative has been relentless: OpenAI and Anthropic are the only games that matter, everyone else is a rounding error, and the smart money goes where the incumbents have already won. Parekh's comments suggest that perhaps, just perhaps, this wisdom might be incomplete. That holding optionality across multiple AI systems—rather than genuflecting at the altar of two—might be a prudent institutional strategy. The audacity of suggesting that concentration risk exists in a $100+ billion market is somehow rendered controversial by sheer repetition of the opposite view.

What could go wrong with this strategy depends entirely on whether the AI revolution follows the playbook written by its current cheerleaders. If OpenAI and Anthropic genuinely monopolize the frontier of capability for the next decade, then Parekh's diversification looks like expensive insurance against a disaster that never materializes. If, however, the AI landscape fragments—if regulatory pressure, capability parity, or simple market dynamics spread value across a wider ecosystem—then the firms that bet everything on two horses will have made an irreversible error. History suggests that concentration bets in emerging technology often end badly, but history also suggests that the current cycle will be different, because it always is.

The broader industry commentary is bleak: we have reached a point where advocating for portfolio construction principles requires explanation and defense. The fact that a $90 billion firm must publish a press release explaining why it isn't putting all its capital into two companies suggests that the industry has collectively abandoned prudence for FOMO dressed in intellectual clothing. Parekh isn't being brave; he's being normal. The rest of the market is being extremely stupid. The tragedy is that this will only become obvious once the losses are locked in.

So here's the thing: being right about diversification doesn't make you a genius. It just makes you someone who didn't forget how to manage risk while everyone else was refreshing Twitter for OpenAI press releases.