Insight Partners is not trying to win the AI boom by picking one mega-lab and hoping the rest of the market bends around it. That is the message from Deven Parekh, the longtime Insight Partners investor whose firm manages roughly $90 billion and has spent years backing software companies before they became obvious.
While much of Silicon Valley is obsessing over enormous checks into OpenAI and Anthropic, Parekh is taking a broader view. His argument is simple: artificial intelligence is not one company, one model, or one infrastructure cycle. It is a rewiring of enterprise software, legal work, cybersecurity, data platforms, developer tools, customer support, and nearly every workflow that businesses already pay for.
Insight Partners’ AI investing strategy is built around diversification
Parekh appears comfortable avoiding the herd mentality that currently defines a large part of venture capital AI investing. The hottest names, especially OpenAI and Anthropic, are attracting enormous amounts of capital at valuations that leave little room for error. Insight is not ignoring foundation models, but it is also not betting the farm on a single AI lab.
That matters because the AI market is still unstable. Model costs are changing. Open-source systems keep improving. Enterprise buyers are experimenting but still cautious. Regulation is evolving. In that kind of environment, owning a wide set of AI-adjacent companies can be safer than trying to predict the one winner at the top of the stack.
For Insight Partners, the opportunity is less about chasing headlines and more about finding where AI becomes daily software. That includes startups that use AI to automate niche but expensive business tasks, as well as established portfolio companies adding AI features to products customers already trust.
Why OpenAI and Anthropic are not the only AI venture capital story
OpenAI and Anthropic dominate the conversation because they sit at the center of the generative AI platform race. They have the talent, the brand power, and the cloud partnerships to shape the market. But their capital needs are also massive, and their valuations reflect expectations that are almost impossibly high.
Parekh’s more diversified approach suggests a belief that the biggest venture returns may not always come from the most visible companies. Plenty of value could be created in the application layer, where AI tools solve specific problems for lawyers, accountants, doctors, engineers, sales teams, and back-office departments.
That is also where Insight Partners has deep experience. The firm has long specialized in growth-stage software businesses, often backing companies once they have product-market fit and recurring revenue. In AI, that playbook can still work, even if the pace is faster and the competitive field is noisier.
The Legora loss to General Catalyst shows how competitive AI deals have become
One notable example is Legora, the AI legaltech startup that Insight wanted but ultimately lost to General Catalyst. The miss is revealing. Premium AI startups now attract fierce competition from top-tier venture firms, sovereign funds, corporate investors, and crossover funds all fighting for access.
For Parekh, losing a deal like Legora does not necessarily invalidate the strategy. If anything, it highlights why diversification matters. In a market where the best AI startups can become expensive almost overnight, investors need discipline. Chasing every hot round can quickly turn a smart thesis into valuation risk.
Legal AI is a particularly attractive category because law firms and corporate legal departments handle huge volumes of repetitive, document-heavy work. If a startup can reduce research time, contract review, or drafting costs without sacrificing accuracy, customers have a clear reason to pay. That explains why companies like Legora are drawing intense investor interest.
Holding stakes in rival AI labs is no longer unusual
Parekh also seems untroubled by the idea of holding positions in companies that may compete with one another. In traditional software investing, that could raise eyebrows. In AI, the lines are blurrier. Startups may compete in one area while partnering in another, and the market is expanding quickly enough that multiple strong players can grow at the same time.
This is especially true for AI infrastructure and model ecosystems. An investor might back companies using different large language models, selling to different industries, or building tools that work across several AI platforms. The goal is not ideological loyalty to one lab. The goal is exposure to the parts of the AI economy that become durable businesses.
What Deven Parekh’s comments reveal about the future of AI startups
The broader takeaway is that Insight Partners is treating AI as a long software cycle, not a short speculative bubble. That does not mean every AI startup will survive. Many will be copied, crushed by incumbents, or exposed as features rather than companies. But businesses with real distribution, defensible data, strong workflows, and measurable customer savings still have room to win.
Parekh’s stance is a useful counterweight to the current AI frenzy. The market may remember the giant OpenAI and Anthropic rounds first, but the next decade of AI investing will likely be shaped by hundreds of quieter companies turning artificial intelligence into practical business software.
For Insight Partners, that is the bet: do not choose one throne before the kingdom is built.
Tags: #InsightPartners #DevenParekh #AIInvesting #OpenAI #Anthropic
About the Author
