Liquidity engineering, mega-round concentration, and proprietary data moats reshape venture investing

Continuation Vehicles Turn Liquidity Into a Portfolio Management Skill

On Sept. 9, 2026, Peterson Partners launched a $510 million single-asset continuation vehicle for Kelso Industries, rolling its stake into a new fund with NorthSands Capital anchoring more than $450 million. The firm’s message was direct: Kelso still has room to grow through acquisitions, talent investment, capability building, and new-market expansion, so Peterson chose extension over a forced sale or IPO.

Partners Group made the same point at larger scale with an €800 million continuation vehicle that will include loans from multiple Private Markets Credit Strategies vintages and its Multi-Asset Credit strategy, while giving existing LPs the choice to roll or cash out through an early payout. Together, the deals show continuation vehicles moving from a niche PE tool to a mainstream liquidity mechanism in uneven exit markets.

For VC and growth investors, the practical implication is clear: liquidity is becoming an active structuring decision, not a binary exit event. The most valuable people on your team will be those who can judge when to extend hold periods, defend valuation and process in GP-led transactions, and manage LP elections and communications. Exit planning now requires secondary-structuring fluency, not just IPO or M&A readiness.

Late-Stage Mega-Rounds Tighten the Funnel Around a Few Giant Checks

US venture deal count fell 32% from July even as total dollars rose, because late-stage financings took 54% of capital across just 57 deals. The top 10 rounds captured $6.9 billion, or 38% of deployed capital, led by River AI’s $1.1 billion raise, Base Power’s $1.0 billion round, and Castelion’s $800 million financing, concentrating activity in AI and infrastructure. That extends the concentration story from exits and fund performance into the deal market itself: capital is increasingly flowing through a narrow set of oversized checks rather than a broad base of financings. For investors, this means broad coverage matters less than conviction-led execution on a handful of oversized opportunities, while the middle market demands tighter screening and lower-touch diligence. Career leverage now comes from domain depth, technical underwriting, and speed on breakout platforms, not from tracking more companies.

Proprietary Data Graphs Are Becoming the Real Moat

Veridion raised a $20 million Series A led by Hoxton Ventures, with Underline Ventures, OTB Ventures, Gapminder, Day One Capital, and LAUNCHub Ventures participating, to expand its AI business graph. The company is not pitching a broad vertical AI platform; it is building a live, API-driven data layer from billions of signals across more than 640 million businesses. That matters because the round is a bet on foundational company data infrastructure, not a single workflow, with use cases spanning risk, insurance, procurement, and market intelligence.

Hoxton’s framing of Veridion as reshaping how business intelligence is built underscores the investor logic: the value sits in the graph, entity resolution, and enrichment layer itself. For venture teams, this reinforces a sourcing pattern where the best opportunities are thesis-led infrastructure companies with proprietary, continuously refreshed datasets that can power multiple enterprise products and AI agents. The practical implication for analysts and associates is clear: diligence now has to go deeper than demos and TAM slides. The edge comes from spotting durable data moats early and pressure-testing freshness, coverage, integration depth, and whether one data asset can support several workflows.