Devin Parekh: AI Investment Stays Diversified
Insight Partners co-leader Devin Parekh admitted in an interview with TechCrunch that the company recently lost a funding bid for Swedish AI legal tech firm Legora to General Catalyst. He elaborated on why, despite peers going all in on OpenAI and Anthropic, Insight Partners, which manages $90 billion, insists on maintaining a diversified approach in the AI sector.
Regarding the loss of the Legora deal, Parekh revealed that company partner Jeff Horing personally flew to Stockholm to advocate for the investment but ultimately failed. He attributed the failure to "the other party selling their value proposition better than we did," rather than a loss on price or valuation.
Meanwhile, Insight Partners' AI strategy involves holding equity in both OpenAI and Anthropic, direct competitors. Parekh does not shy away from this and clearly states he is comfortable betting on both "competing labs," contrasting with many peers who choose to "pick sides."
In terms of fund data, Insight Partners currently manages $90 billion and is operating its 13th fund, having recently returned over $20 billion to limited partners (LPs) through strategic sales and IPOs. Notable investment cases mentioned by Parekh include: data platform Databricks (multiple rounds of leading or co-leading), security company Wiz (seed round and follow-on investments), enterprise AI agent platform Wonderful (valued at $5 billion), security company Armis (later sold to ServiceNow for $7 billion), and fintech company Ramp.
Parekh summarizes his investment philosophy in one sentence: "Let the portfolio speak for itself... our performance should prove itself." He also expressed cautious observation regarding the current valuation pace in the primary market, pointing out that the speed of valuation increases has reached 2021 levels, with the time interval between multiple follow-on investments being "so short that there is almost no incremental data to judge."
Market mechanisms reflect the current funding game in the AI primary market—top AI labs (OpenAI, Anthropic) have become scarce resources that multiple top institutions are competing for. Institutions like General Catalyst are securing leading targets in niche sectors like Legora through more aggressive on-site public relations and transaction execution speed. Insight Partners' response strategy is not to heavily bet on a single star target but to hold equity in multiple competitors through large-scale funds like Fund 13, relying on real return data from exited projects to maintain LP confidence. Essentially, this is a hedge against the systemic risk of "valuation growth comparable to the 2021 bubble period" through "diversification + realized returns."
Source: Public information
ABAB AI Insight
Devin Parekh has been with Insight Partners for 26 years, and the company's historical strategy has been "growth investments in the software industry"—from early bets on SaaS to the current focus on AI, Insight has maintained a style of "casting a wide net and making multiple bets." Admitting the loss of Legora continues its consistent communication style of "openly acknowledging failures and letting overall portfolio performance speak," rather than selectively promoting only success stories.
In terms of funding pathways, Insight Partners is simultaneously raising new funds through Fund 13 to invest in top AI labs like OpenAI and Anthropic while returning over $20 billion to LPs through exits from invested projects (Armis sold for $7 billion to ServiceNow, multiple IPOs), creating a cycle of "new money in, old money out." The motivation behind this approach is to maintain the fund's long-term fundraising capability—if all funds are concentrated on a few unexited AI star targets, LPs will not see real returns in the short term, and realized exit cases are the most convincing currency for persuading LPs to continue subscribing to new funds.
Holding equity in both OpenAI and Anthropic is akin to the traditional private equity strategy of "hedged positioning" among multiple competitors within the same vertical, essentially transferring the risk of "which AI lab will ultimately succeed" to the market itself rather than relying on the institution's predictive capabilities. The loss in bidding for niche sectors like Legora indicates that AI startup financing is evolving into a direct confrontation between institutions like General Catalyst and traditional growth funds. In industry positioning, the AI primary market is shifting from "top labs monopolizing attention" to a stage where "vertical niche fields are flourishing."
This essentially reflects a risk hedging strategy adjustment in the context of capital concentration—when the industry generally believes that the AI sector will ultimately form a "winner-takes-all" pattern, the instinctive reaction of institutions is to go all in on top players. However, Parekh represents an alternative judgment: at a stage where the landscape has not truly settled and valuation growth has reached levels comparable to the 2021 bubble, the risk of overly concentrating bets on a single winner may outweigh the risks of a diversified approach relying on realized exits to maintain cash flow. Mechanically, this is a direct game between the expectations of "winner-takes-all" and "bubble valuation risks" in the same capital market, and Insight chooses to stand on the side of "preserving portfolio flexibility and not betting on a single outcome."
ABAB News · Cognitive Law
- Losing a star deal is not scary; what’s scary is not having realized returns to back it up.
- Betting on both competitors is not about who wins; it’s about not having to guess.
- When valuations rise like in 2021, only cautious money dares to slow down.