AI IPOs are off to the races | PitchBook News Summary

1. Key Themes

🏁 The AI IPO Race Is Accelerating — and Valuation Quality Varies

OpenAI, SpaceX, and Anthropic are all moving toward public markets simultaneously, creating a defining moment for the AI investment landscape.

"OpenAI confidentially filed its S-1 yesterday, closely following SpaceX and Anthropic. Of the three, OpenAI has the highest price per unit of quality, according to our recent analyst note."

⚖️ Legal AI Startups Face an Existential Threat from Big Law Going In-House

Despite raising billions in VC, Harvey and Legora are being squeezed from two sides: law firms building proprietary AI and foundation models commoditizing their core offering.

"Harvey and Legora's issue is that there's starting to be increasingly less differentiation between what they offer and what Claude offers directly." — Dan Block, Director, Sterne Kessler

"Investors are increasingly questioning both the lofty valuations of these companies and their ability to capture a large enough share of the market."

💧 The PE Liquidity Crisis Is Cascading — Even Secondaries Need an Exit

The distribution drought has become so severe that secondary funds — which exist specifically to provide liquidity — are now themselves illiquid, forcing novel structured solutions.

"In 2025, distributions as a proportion of PE net asset value dropped to 9%, the lowest level since 2010, according to Goldman Sachs Asset Management."

"This suggests that secondary funds, which exist to provide liquidity to PE fund managers and LPs, are increasingly in need of liquidity themselves."

🦄 Unicorn Creation Remains Robust in 2026

Despite macro headwinds, startup formation at scale continues at a healthy pace, suggesting the private market innovation engine has not stalled.

"2026 continues to see solid unicorn creation, as 50 startups have crossed the billion-dollar valuation threshold so far in Q2 2026, bringing the total up to 139 for the year."

🏥 Healthcare Displaces Software as the Top Sector in the Leveraged Loan Market

A notable sector rotation is underway in private credit markets, with potential implications for deal sourcing and credit risk underwriting.

"Software has lost its throne in the leveraged loan market, with healthcare claiming the top spot for the first time since 2015."

2. Contrarian Perspectives

Large Law Firms May Be Better Positioned to Win the Legal AI Race Than VC-Backed Startups

The consensus view has been that well-funded legal AI startups like Harvey would disrupt law firms. The contrarian read: the largest firms have both the resources and the proprietary data to beat them at their own game — and the incentive to keep that data in-house.

"Claude Code and new AI infra solutions make it easier than ever to completely rebuild Harvey or Legora in short order, while giving [firms] full control over the user experience, functionality, workflows and, most importantly, their data." — Ethan Batraski, GP, Venrock

Supporting evidence: Kirkland & Ellis has committed $500 million to internal AI development, and 62 of the top 100 US law firms posted gross revenue over $1 billion in 2025 — meaning the capex for in-house AI is achievable for a meaningful portion of the addressable market.

Tertiary Secondaries ("Secondaries of Secondaries") Could Become a Legitimate $20B+ Asset Class

It sounds like financial engineering run amok — but the structural illiquidity of the secondary market is creating durable demand for a new layer of liquidity provision.

"Netley Capital...said it has invested more than $825 million in cashing out investors in secondary funds, and expects the total market for tertiary stakes to one day exceed $20 billion."

The counterintuitive implication: the very funds designed to solve illiquidity have become illiquid themselves, suggesting the problem is structural, not cyclical — and therefore a persistent opportunity.

Wealth Management Faces a Generational Cliff, Not Just an AI Disruption Threat

The risk to traditional wealth advisors is often framed as an AI story. The deeper risk is demographic: heirs simply don't want their parents' advisors.

"The vast majority of high-net-worth heirs say they won't keep their parents' financial adviser. With 80% of Gen Z already using AI for personal finances, the $500,000-a-year wealth management job is at the mercy of a ticking generational clock."

3. Companies Identified

Harvey

Description: Legal AI startup
Why mentioned: Raised over $1B+ in VC; now under pressure from Big Law building in-house AI and from LLMs commoditizing its core product
Quote: "Harvey and Legora's issue is that there's starting to be increasingly less differentiation between what they offer and what Claude offers directly."

Legora

Description: Legal AI startup
Why mentioned: Co-mentioned with Harvey as a target of investor skepticism; collectively the two have raised over $2B; declined to comment on fears of becoming a competitor to their own clients
Quote: "Legal AI stars Harvey and Legora have raised over $2 billion in VC funding collectively."

Kirkland & Ellis

Description: One of the world's largest and highest-grossing law firms
Why mentioned: Cited as the bellwether for Big Law building proprietary AI, with a $500M internal commitment and plans to train a model on its attorneys' collective intelligence
Quote: "Kirkland & Ellis last month said it will spend $500 million on internal AI projects."

Freshfields

Description: Global Magic Circle law firm
Why mentioned: Partnered directly with Anthropic to build customized legal AI, receiving early model access in exchange — a model of direct LLM-to-enterprise partnership that bypasses legal AI vendors
Quote: "Freshfields recently inked a deal with Anthropic to train a legal AI model."

PhysicsX

Description: European industrial AI company
Why mentioned: Raised a $300M Series C at a $2.4B valuation led by Temasek — a signal of sovereign wealth backing the next wave of industrial AI
Quote: "PhysicsX just raised a $300 million Series C at a $2.4 billion valuation, led by Singapore's state investor Temasek, signaling that sovereign wealth and top-tier VC are lining up behind the next wave of European industrial AI."

Blackstone (Strategic Partners)

Description: Global alternative asset manager; Strategic Partners is its secondaries vehicle
Why mentioned: Seeking to sell $2B+ in secondaries fund positions via a collateralized fund obligation — a landmark illustration of the liquidity crisis cascading into the secondaries layer
Quote: "Blackstone is looking to sell more than $2 billion in fund positions held by its Strategic Partners secondaries vehicles."

Netley Capital

Description: Specialist firm acquiring stakes in secondary funds ("tertiary" market)
Why mentioned: Has invested $825M+ in cashing out secondary fund investors; projecting a $20B+ tertiary market
Quote: "Netley Capital...said it has invested more than $825 million in cashing out investors in secondary funds, and expects the total market for tertiary stakes to one day exceed $20 billion."

Bending Spoons

Description: Italian tech conglomerate known for acquiring distressed or undervalued consumer apps (e.g., AOL, Eventbrite)
Why mentioned: Filed for a US IPO, representing a notable non-AI tech listing story
Quote: "Italian tech conglomerate Bending Spoons, which has built its reputation by acquiring an eclectic mix of companies such as AOL and Eventbrite, has filed to go public."

Revolut

Description: London-based fintech / neobank
Why mentioned: Preparing a $750M secondary share sale at an implied ~$109B valuation — one of the largest private fintech valuations globally
Quote: "Revolut is preparing a $750 million secondary share sale, which could value the London-based fintech specialist at around $109 billion."

Moonshot AI

Description: China-based AI chatbot developer
Why mentioned: Seeking up to $2B in a new round at a $30B valuation — signals continued mega-round activity in Chinese AI
Quote: "Moonshot AI...is looking to raise up to $2 billion in a round that would give the startup a $30 billion valuation."

OpenAI

Description: Leading US AI company, creator of ChatGPT and GPT model series
Why mentioned: Confidentially filed its S-1, entering the IPO race alongside SpaceX and Anthropic; flagged as the most expensive on a quality-adjusted basis
Quote: "OpenAI has the highest price per unit of quality, according to our recent analyst note."

Zepto

Description: India-based instant delivery platform
Why mentioned: Planning an India IPO targeting $836M raise, backed by General Catalyst — a signal of emerging market liquidity events
Quote: "India-based instant delivery platform provider Zepto...is planning to raise 80 billion rupees ($836 million) in its India IPO."

4. People Identified

Ethan Batraski

Title: General Partner, Venrock
Why mentioned: Offered the sharpest critique of legal AI vendor moats, arguing that LLM commoditization makes Harvey/Legora easily replicable
Quote: "Claude Code and new AI infra solutions make it easier than ever to completely rebuild Harvey or Legora in short order, while giving [firms] full control over the user experience, functionality, workflows and, most importantly, their data."

Dan Block

Title: Director, Electronics Practice Group, Sterne Kessler
Why mentioned: Provided a practitioner's view on the narrowing differentiation between legal AI vendors and direct LLM capabilities
Quote: "Harvey and Legora's issue is that there's starting to be increasingly less differentiation between what they offer and what Claude offers directly."

Sri Pangulur

Title: Partner, Mayfield
Why mentioned: Articulated the threshold at which in-house legal AI investment becomes rational — and the strategic ambition behind it for elite firms
Quote: "For those types of firms…it's about redesigning their entire service delivery model."

Caspar Berendsen

Title: Founder, Netley Capital; former Cinven buyout executive
Why mentioned: Pioneer of the tertiary secondaries market; framing it as market maturation rather than distressed opportunism
Quote: "Whether these developments are a sign of market maturation, as Berendsen suggests, or an opportunistic response to a liquidity shortage will be an ongoing debate."

5. Operating Insights

For Legal AI Founders: Data Moats Must Be Proprietary — or They Are No Moat At All

The article makes clear that the fatal flaw of legal AI vendors is their dependency on client data they don't own. The strategic warning for any vertical AI startup: if your competitive advantage rests on the client's data processed through a third-party model, you are one enterprise build-vs-buy decision away from irrelevance.

"Using legal AI vendors also requires firms to entrust a valuable asset: their collective knowledge. Many industry insiders foresee a future in which those vendors evolve into competitors."

Implication: Vertical AI startups must either (a) own proprietary training data that clients cannot replicate, (b) create deep workflow integrations that raise switching costs, or (c) accept that foundation model providers will commoditize them.

For PE/Fund Managers: Proactively Engineer Liquidity — Don't Wait for It

The article reveals that even sophisticated secondary fund managers are caught in the distribution drought. Managers who wait passively for exits are ceding control. The market is rewarding those who structure liquidity — via CFOs, continuation vehicles, tertiary sales, or LP-friendly secondaries — rather than waiting for traditional M&A or IPO windows.

"In 2025, distributions as a proportion of PE net asset value dropped to 9%, the lowest level since 2010."

"The distribution drought represents a generational buying opportunity for secondary funds."

6. Overlooked Insights

Healthcare's Structural Ascent in Private Credit Deserves Attention Beyond the Headline

The displacement of software as the top leveraged loan sector by healthcare — for the first time since 2015 — is noted briefly but has significant implications for credit underwriting, deal flow, and sector specialization in private credit. Healthcare is more capital-intensive, more recession-resistant, and more regulation-sensitive than software, meaning this shift could alter the risk/return profile of the broader leveraged loan market.

"Software has lost its throne in the leveraged loan market, with healthcare claiming the top spot for the first time since 2015."

The Freshfields-Anthropic "Early Access for Data" Model May Become a Template for LLM Commercialization

Buried in the legal AI story is a structurally important deal mechanic: Freshfields gets early model access; Anthropic gets a rich legal training corpus. This barter-for-data model — where enterprises contribute domain-specific knowledge in exchange for preferential LLM access — could become a blueprint for how foundation model providers penetrate high-value verticals without going through intermediary application vendors.

"Freshfields partnered directly with Anthropic to build customized tools integrated into the firm's systems. In exchange, Freshfields will get early access to Anthropic's latest models."