Insight Partners’ Parekh Defends Diversification Strategy Amid AI Fund Concentration Trend

Devin Parekh, who has co-led Insight Partners for 26 years, recently emphasized his firm’s commitment to portfolio diversification during an interview at TechCrunch’s StrictlyVC event in New York. While many venture capital firms are allocating massive portions of their funds to frontier AI labs like OpenAI and Anthropic, Parekh argued that the long-term data supports spreading risk across a broader range of investments.

Parekh addressed concerns regarding artificial intelligence safety, noting that while risks such as access to open-source models by non-state actors exist, the potential benefits are substantial. Citing his role on the board of NYU Langone, he highlighted AI’s ability to analyze millions of patient records to predict health events, describing the net impact as highly positive. He compared current AI anxieties to historical fears surrounding new technologies like drone warfare, asserting that society has consistently raised living standards despite emerging risks.

The conversation shifted to Insight Partners’ investment methodology, which Parekh described as intentionally low-profile compared to peers who frequently engage in public discourse. He stated that the firm prefers to let its performance speak for itself rather than seeking media attention. Currently managing $90 billion in assets, Insight operates across early-stage, growth, and buyout sectors globally, with no fixed geographic or strategic allocations.

Parekh noted that buyout activity has slowed due to high interest rates and unfavorable debt markets, with the firm not completing a major buyout since 2024. Conversely, he observed that venture valuations are rising at a pace reminiscent of 2021. He pointed out that recent funding rounds often lack incremental data, meaning investors are paying higher prices without reducing risk. Consequently, Insight has shifted toward earlier-stage investments, utilizing smaller checks in its scale fund to maximize returns on winners like Wiz while limiting downside exposure.

On the topic of geographic concentration, Parekh acknowledged that while AI infrastructure talent remains dense in San Francisco, other verticals like financial services tech see talent clustered in cities like New York. He admitted that Insight lost the deal for AI legal-tech company Legora to General Catalyst, attributing the loss to the competitor’s superior pitch rather than any strategic failing.

The firm holds stakes in rival AI companies OpenAI and Anthropic, a strategy Parekh defended as a matter of stage appropriateness. He explained that early-stage exclusivity agreements prevented simultaneous investment in direct competitors, but later-stage investments resemble purchasing public equity. He viewed OpenAI as the dominant consumer play and Anthropic as having a stronger enterprise focus, though he noted these positions are evolving.

Parekh also touched on physical AI, describing robotics as currently being more science project than viable business. He humorously noted the generational divide on the topic with his son, a fellow venture capitalist who believes physical AI is the most exciting sector available. Despite this, Insight remains watchful but not yet committed.

Addressing the concentration of capital in OpenAI and Anthropic, which absorbed roughly half of all venture dollars in the first half of the year, Parekh expressed concern for funds pitching 35–40% allocation to a single company. While he acknowledged that concentrated strategies by firms like Founders Fund have succeeded, he stressed that Insight, now on its 13th fund, prioritizes long-term diversification over short-term performance boosts.

The discussion also covered secondaries and liquidity. Parekh criticized funds that have failed to return capital to limited partners, arguing that LPs require liquidity even if a position continues to grow. He revealed that Insight has returned over $20 billion to LPs in the last two years through strategic sales and IPOs. He referenced Elad Gill’s advice that founders should consider de-risking during frothy valuation periods, though he admitted founders often disregard such counsel.

Looking ahead to the anticipated IPOs of Anthropic and OpenAI, Parekh suggested that while these trillion-dollar entities would be absorbed by the market similar to SpaceX, the more critical question is what bar the next tier of companies will set. He predicted a wave of AI-related IPOs over the next 18 months as these companies mature into normal-growth enterprises.

Finally, Parekh shared examples of the firm’s dynamic approach to portfolio management, citing Armis as a key success story. After missing an initial deal with Sequoia, Insight made a small $5 million investment, later bought out the entire cap table, and sold the security company to ServiceNow for $7 billion. The firm recently conducted a review of 300 portfolio companies to identify inflection points for additional investment or exit, underscoring its active management style.