Insight Partners’ Devin Parekh on why the firm is diversifying while everyone else bets the farm on OpenAI and Anthropic

Devin Parekh, who has co-led Insight Partners for 26 years, said the firm is intentionally keeping a broad, stage-agnostic investment profile rather than concentrating capital in a small set of frontier AI labs. Parekh argued that performance should be driven by the portfolio rather than publicity, noting Insight’s quieter public profile despite managing $90 billion in assets. He framed the firm’s allocation as temporal and adaptive: percentages across early-stage, growth and buyouts change fund-to-fund and by market conditions. Parekh criticized the current pace of late-stage valuations, likening it to 2021 when quick follow-on rounds meant paying more without gaining much new information. That dynamic has pushed Insight to move earlier in many cases, making smaller initial commitments — he cited typical early checks of $20–25 million rather than single large bets of $500 million — and then backing winners with follow-on funding. He used Wiz as an example where writing multiple checks after a Series A produced outsized returns for the firm. On competition and conflicts, Parekh acknowledged internal debate about holding positions in rival AI labs but said the question was often stage-dependent: early exclusivity can prevent backing multiple competitors, while later-stage investments are more like buying a public stock. Insight enforces information-sharing restrictions at Series A/B and avoids investing in companies that directly compete; he also noted some founders object to even small overlaps in revenue. Parekh described losing the deal for Legora to General Catalyst as down to the competitor selling itself better on that occasion. Regarding technology focus, Parekh said Insight is cautious about physical-AI and robotics plays, viewing many such companies as still essentially science projects where timing and adoption are uncertain. He contrasted that with areas like AI applied to healthcare, where he serves on the board of NYU Langone and sees concrete near-term benefits from using large patient datasets. Talent density, he added, remains concentrated for AI infrastructure in San Francisco while vertical-specific talent can be geographically dispersed — for example, financial-services-focused companies cluster in New York.