While the venture capital world appears locked in a feeding frenzy over foundation models and large language models, one of Silicon Valley’s largest investment firms is deliberately pumping the brakes. With $90 billion in assets under management, Insight Partners has positioned itself as a contrarian player in the AI investment landscape—and that deliberate diversification might prove more prescient than the current consensus suggests.
The technology news cycle has become almost monotonously focused on OpenAI and Anthropic. Every funding round, every partnership announcement, every emerging capability gets filtered through the lens of which closed ecosystem will dominate enterprise AI. But Devin Parekh and his team at Insight Partners are asking a fundamentally different question: what if betting the entire portfolio on two companies is exactly backwards?
The Case Against Monoculture Investing
Concentration risk exists in venture capital just as it does in any portfolio strategy. When the majority of institutional capital flows toward the same handful of companies, several structural problems emerge. First, valuation inflation becomes inevitable—if everyone is bidding on the same assets, prices disconnect from underlying fundamentals. Second, the broader ecosystem suffers from capital starvation. Third, and perhaps most importantly, the industry misses the heterodox approaches that often generate outsized returns.
Insight Partners’ approach reflects mature capital management principles borrowed from traditional investing. Diversification wasn’t invented in tech; it’s been a cornerstone of sound money management since modern portfolio theory emerged decades ago. Yet the venture world, particularly around transformative technology categories, often abandons diversification in favor of herd behavior. Parekh’s willingness to hold stakes across multiple AI labs—even those competing directly with portfolio companies—suggests confidence in a landscape where multiple players capture value rather than a winner-take-most dynamic.
The Legora Lesson and Strategic Recalibration
Losing the Legora investment to General Catalyst, while potentially stinging in the moment, provides useful real-world data about market dynamics. Rather than viewing it as a failure to recognize AI excellence, Insight Partners appears to have processed it as evidence that maintaining optionality matters more than chasing every hot deal. In industry trends we see this play out repeatedly: the venture firms that remain most profitable over multi-decade horizons are those that resist fomo-driven decision making.
The product launch velocity across AI startups has accelerated dramatically, and it’s easy to fall into the trap of assuming that superior product automatically generates market dominance and investor returns. History suggests otherwise. Some of the best-funded, most technically impressive products have failed to achieve commercial success at the scale their supporters anticipated. Conversely, seemingly ordinary products backed by strong distribution or business models have generated exceptional returns.
Reading the Macro Signal
A $90 billion asset manager making strategic decisions reflects signal about their own thesis regarding AI’s evolution. If Insight Partners leadership genuinely believed that OpenAI and Anthropic represent a binary choice—where investing heavily in these two firms at current valuations and supporting infrastructure would generate superior risk-adjusted returns—we would expect to see capital concentrated there. Instead, the firm is explicitly choosing to remain diversified.
This suggests conviction that: foundation models will eventually become commoditized, multiple architectures and approaches will prove viable long-term, and value capture will flow to companies solving specific enterprise problems rather than those controlling the base layer. That’s a much different bet than the predominant narrative in technology news currently suggests. Whether distributed, fine-tuned models prove more economically sustainable than centralized APIs remains genuinely unsettled. Edge AI, specialized model architectures, and domain-specific solutions are all viable value vectors that don’t require betting everything on OpenAI’s compute advantages or Anthropic’s safety research.
Key takeaway: Insight Partners’ refusal to follow the herd on AI concentration reveals a sophisticated approach to portfolio construction in emerging technology categories. Rather than assuming the obvious frontrunners will capture all value, maintaining diversified exposure hedges against the very real possibility that AI’s commercial evolution follows a different path than current consensus assumes.
The venture capital industry often talks about backing founders with contrarian views, yet somehow investors themselves default to consensus thinking when allocating capital. Does Insight Partners’ diversification strategy suggest the smartest money in the room is skeptical of the current AI narrative—or are they simply being prudent in an area where fundamental uncertainty remains genuine?
Get Tech Savvy Digest in your inbox
IT news, cybersecurity, and crypto — the signal, not the noise. No spam, unsubscribe anytime.