As AI valuations surge, the firm is choosing a less fashionable path—backing more companies early and investing more heavily only after results emerge.
Devin Parekh has co-led investment firm Insight Partners for 26 years. The firm manages $90 billion in assets. Unlike many venture capital investors, it rarely draws attention with bold statements, instead focusing on the performance of companies in its portfolio.
As reported by Techcrunch
Insight Partners’ strategy differs from that of funds effectively making their biggest bets on OpenAI and Anthropic. The firm has invested in both labs and participated in numerous funding rounds for Databricks, but it has not abandoned broad diversification.
Over the long term, venture capital has always rewarded diversification. We are already on our 13th fund, so we have to think not in terms of one fund, but ten.
– Devin Parekh
Why Insight Partners Diversifies Its Investments
Capital allocation across early-stage startups, scaling companies, buyout deals, and secondary sales does not follow a fixed ratio. As Parekh explained, the portfolio’s structure changes along with market conditions.
Buyouts have become more difficult because of high interest rates, weak demand for debt financing for software companies, and falling valuations. Insight Partners has not completed any major deals of this kind since 2024.
At the same time, venture capital valuations are rising again at rates reminiscent of 2021. Parekh believes that rapid funding rounds often fail to provide investors with enough new information while forcing them to pay more. That is why the firm tries to invest in companies at earlier stages, initially committing smaller amounts and then increasing its investments in startups that deliver the strongest results.
This model was applied to Wiz, among others. Insight Partners invested in the company during its Series A round and continued to finance its growth thereafter. According to Parekh, steadily increasing stakes in the most successful companies has become one of the funds’ main sources of returns.
Deal Geography and Competition for Startups
The regional distribution of investments depends on the specific industry. For example, artificial intelligence infrastructure specialists are concentrated primarily in San Francisco, while fintech has a stronger talent base in New York.
Insight Partners also tried to invest in legal startup Legora but lost out to General Catalyst. Company partner Jeff Horing personally flew to Stockholm to present an offer to the founder. Parekh suggested that the competitor had done a better job of explaining its value at the right moment, but stressed that investors do not have to win every deal.
How the Firm Assesses OpenAI and Anthropic
Insight Partners’ simultaneous participation in funding OpenAI and Anthropic might once have been viewed as a conflict of interest. However, according to Parekh, the risks depend on the stage of the investment.
At early stages, the firm does not invest in direct competitors and follows restrictions on information sharing. At later stages, when an investor does not sit on the board and is not involved in management, such a deal is more like acquiring a stake in a promising company.
Initially, Insight Partners viewed OpenAI as a leader in the consumer segment and Anthropic as a company with a clearer focus on enterprise customers. However, the lines between these models are quickly blurring. In addition, the labs’ need for tens of billions of dollars reduces their ability to demand exclusivity from investors.
Parekh takes a more cautious view of physical AI and robotics. In his opinion, many such companies remain primarily research projects for now. Investors must simultaneously assess both the likelihood that the technology will be developed and the timeline for its mass adoption.
The Risks of Excessive Capital Concentration
According to Parekh, some funds plan to direct 35–40% of their capital into OpenAI or Anthropic. He does not deny these companies’ potential but considers such concentration risky. Even if a bet on one lab temporarily improves a fund’s performance, long-term results do not demonstrate the superiority of this model for most investors.
Secondary deals and returning capital to investors remain separate challenges. A large number of funds raised substantial sums in 2021–2023 but have yet to return them to their investors. Over the past two years, Insight Partners has returned more than $20 billion to investors through strategic sales and initial public offerings. The firm expects to receive several billion dollars more in the future.
Parekh also believes it makes sense for founders to take some profits off the table during periods of rapidly rising valuations. They do not have to sell their entire stake, but reducing their exposure by 10–20% may be justified, since valuations cannot rise indefinitely as a matter of mathematics.
Over the next 18 months, he believes more large technology companies could enter the public markets. IPOs by Anthropic, OpenAI, and SpaceX could become important events for the market. At the same time, the real test will be the next group of companies going public – those that grow more slowly and will have to prove their value to public-market investors.
Insight Partners’ approach is based on a broad portfolio, early-stage investments, and subsequently increasing stakes in the strongest companies. The story of Armis – which was initially acquired as a small stake, then bought out in full and sold to ServiceNow for $7 billion – illustrates the firm’s core principle: a profit can come from either a small or a large check if an investor recognizes a strong founder in a promising market at the right time.
