Insight Partners Deven Parekh Explains Diversified AI Bet Beyond OpenAI and Anthropic
TL;DR
- Insight Partners managing director Deven Parekh says losing the Legora deal to General Catalyst was about price discipline, not conviction, calling the AI legal tech star a great company at a valuation Insight wouldn't chase.
- The $90 billion firm is deliberately spreading its AI bets across rival foundation model labs and application-layer winners instead of concentrating only on OpenAI and Anthropic.
- Parekh argues the real enterprise value in AI will accrue to vertical workflow disruptors and diversified infrastructure picks, a strategy Insight is backing with fresh capital despite a slower exit market.
Losing Legora to General Catalyst
Deven Parekh isn't pretending Insight Partners wins every competitive AI deal. In a new interview, the longtime managing director addressed one of the firm's most talked-about recent misses: Legora, the fast-growing AI platform for lawyers.
Legora, formerly known as Leya and based in Sweden, has become one of Europe's breakout legal AI stories, riding surging demand from top law firms for contract review, research automation, and agentic drafting workflows. When the company raised its latest large round led by General Catalyst, Insight Partners was in the mix and ultimately walked away.
Parekh described Legora as an excellent business with strong product-market fit, but said Insight held a firm line on entry price. General Catalyst was willing to pay up to win, and Insight let it go. For Parekh, that discipline is core to how Insight has operated through multiple hype cycles, from SaaS to mobile to cloud.
He framed the loss as painful but consistent: Insight would rather miss an outlier than overpay across a dozen similar growth-stage AI applications and break its return model.
A $90 Billion Firm Playing a Different AI Game
The Legora story matters because it opens a window into Insight Partners' much larger AI playbook. With around $90 billion in assets under management, Insight is one of the largest software-focused investors in the world, and its choices move markets.
Unlike some crossover and venture giants that have concentrated massive capital in OpenAI and Anthropic, Insight has taken a more diversified approach. Parekh said the firm intentionally avoided putting all of its foundation-model exposure into the two most highly valued U.S. labs.
Instead, Insight has been backing a basket of rival model builders and infrastructure players, including bets in Europe and open-weights ecosystems, alongside a wide portfolio of vertical AI applications in legal, healthcare, cybersecurity, sales, and software development.
The logic, Parekh explained, is simple: no one knows which lab will win on capabilities, cost, distribution, or regulation, and concentration at $60 billion-plus valuations leaves little room for venture-scale upside.
Why Not Just OpenAI and Anthropic
Parekh was direct about the concentration risk he sees in AI investing right now. OpenAI and Anthropic dominate headlines, talent, and enterprise pilots, but their rounds are enormous, highly competitive, and priced for perfection.
For a growth investor like Insight, writing a $100 million-plus check at peak valuation into a single lab creates binary risk. If model commoditization accelerates, inference prices collapse, or open-source alternatives close the gap, returns compress quickly.
By backing challengers and alternative architectures, Insight is betting that innovation will remain multipolar. European sovereign AI initiatives, enterprise demand for custom and private models, and developer preference for open and cheaper models all support a world with multiple winners, not a duopoly.
Parekh also pointed to leverage: smaller checks into emerging labs can buy meaningful ownership and board influence, whereas a late-stage investment in an incumbent buys access but little control.
Where Insight Sees Real AI Returns
If foundation models are the expensive lottery tickets, Parekh believes the more reliable returns will come from the application and workflow layer.
Insight is doubling down on AI-native vertical SaaS companies that own customer workflows end-to-end — tools that don't just add a chatbot but replace hours of billable work, clinical documentation, security triage, or sales prospecting. Legora itself fits that thesis perfectly, which is why losing it stung.
He also highlighted developer tools, data infrastructure, evaluation and observability startups, and AI security firms as critical picks-and-shovels bets that benefit regardless of which model wins.
According to Parekh, enterprises are moving from experimentation to procurement, and that shift favors vendors with deep integrations, compliance readiness, and measurable ROI — areas where Insight's decades of scaling B2B software companies gives it an edge in company building.
What Comes Next for Insight in AI
Despite a sluggish IPO market and longer hold periods, Parekh said Insight is still deploying aggressively into AI and has dry powder to lead both early growth and late-stage rounds.
The firm is telling founders it can be patient on exits while providing go-to-market playbooks, talent networks, and M&A support. In return, it is demanding clearer paths to gross-margin expansion and defensibility beyond a wrapper around a third-party API.
Losing Legora, in that context, isn't a retreat — it's a signal. Insight wants to be the diversified AI house: exposed to multiple labs, overweight in vertical disruptors, and unwilling to chase any single deal, even a hot one, at any price.
Get All The Latest Updates Delivered Straight To Your Inbox For Free!