The best unicorns aren't the… | PitchBook News Summary

1. Key Themes

Theme 1: AI Dominates Unicorn Valuations, But Not Business Quality

AI has captured an extraordinary share of unicorn market value, yet that dominance does not translate into superior business fundamentals.

"AI now represents nearly 62% of all unicorn value, even though AI companies make up over 40% of active unicorns. Anthropic and OpenAI alone account for $1.8 trillion of that, or 22% of the $8.2 trillion universe."

"That dominance doesn't mean AI companies always make the best businesses... Anthropic and OpenAI rank 11th and 15th out of 18 major unicorns on PitchBook's Business Quality framework."

The structural weakness: model builders depend on rented compute, which compresses margins and limits operational control.

"The reason has a lot to do with the fact that model builders must rely on rented computing power, which leaves them with thinner margins and less control over their operations."

Theme 2: Climate Tech & Robotics Are the Real Capital Efficiency Leaders

Despite AI's valuation dominance, cleantech and robotics are generating more value per dollar invested — a signal for where smart capital may be underpenetrated.

"Cleantech leads every sector in the unicorn market, with a median relative velocity of value creation of nearly 330% in the first half of 2026, followed by robotics and drones at 202.5%. AI posted an RVVC of 183.7%."

"Billion-dollar data center deals are climate tech's new calling. Investments in nuclear fission, geothermal, energy storage and efficient cooling technologies all illustrate the climate tech sector's shift toward infrastructure."

Theme 3: AI Is Reshaping M&A Strategy and Creating New Deal Categories

Investment bank CEOs are flagging that AI is fundamentally altering the M&A landscape — both by prompting strategic buyers to consolidate and by creating distress among software companies whose valuations have eroded.

"The rapid spread of AI technology is making strategic buyers question whether they need to make 'stronger and bigger' acquisitions to compete with their peers." — John Weinberg, Evercore

"AI means more take-privates of 'healthy companies that can't seem to regain their prior valuations' and more companies feeling they need to be part of a bigger entity to protect their business models." — Paul Taubman, PJT Partners

Theme 4: Private Credit Recovery Is Under Pressure, Especially in Software

Private credit is entering a more challenging phase, with recovery expectations — particularly for software — likely to disappoint.

"As the private credit market enters a challenging new phase marked by worsening loan performance, recoveries are likely to underperform expectations, especially for the software sector." — Michael Gross, SLR Capital Partners

Theme 5: AI Is Compressing Pricing in Financial Services While Expanding TAM

A nuanced dynamic is emerging in financial services: AI is simultaneously deflating pricing in established business lines while opening new market opportunities.

"AI is already driving down pricing in some of its businesses, including in its valuations business, one of the largest in the sponsor-backed market. At the same time, AI has led to a massive expansion of the total addressable market, driving strong growth even with weaker pricing power." — Scott Adelson, Houlihan Lokey CEO

2. Contrarian Perspectives

Perspective 1: The Best Unicorns Aren't the Biggest-Name AI Companies

The consensus assumption is that investing in the largest, most talked-about AI unicorns (OpenAI, Anthropic) is the surest path to returns. PitchBook's data directly challenges this.

"Anthropic and OpenAI rank 11th and 15th out of 18 major unicorns on PitchBook's Business Quality framework. (Databricks and Stripe topped our scoring framework.)"

The implication: infrastructure-layer and application-layer companies with owned compute or distribution moats score materially better on business quality than the headline model builders.

Perspective 2: Cleantech — Not AI — Is the Superior Capital Efficiency Story Right Now

While investor attention and media coverage overwhelmingly favor AI, the data shows cleantech is dramatically outperforming on value creation per dollar invested.

"Cleantech leads every sector in the unicorn market, with a median relative velocity of value creation of nearly 330% in the first half of 2026... AI posted an RVVC of 183.7%."

Cleantech RVVC is nearly 80% higher than AI's — a gap that suggests significant underpricing of climate tech opportunities relative to the capital flowing into AI.

Perspective 3: AI Will Worsen — Not Solve — PE's Distribution Problem

Conventional wisdom holds that improving AI-driven productivity and deal flow will help PE firms generate liquidity. PJT's Taubman argues the opposite.

"AI will also exacerbate a problem that private equity firms thought was fading: the challenge of generating distributions for investors... This could drive the use of continuation funds, which allow PE firms to extend their hold on businesses while providing liquidity to investors."

AI-induced valuation dislocation means more portfolio companies will struggle to achieve exit-worthy valuations, prolonging holding periods and accelerating the adoption of continuation fund structures as a workaround.

3. Companies Identified

Anthropic Description: Leading AI model developer Why mentioned: Despite a massive valuation ($1.8T combined with OpenAI), ranks 11th out of 18 on PitchBook's Business Quality framework — underperforming on fundamental business metrics Quote: "Anthropic and OpenAI rank 11th and 15th out of 18 major unicorns on PitchBook's Business Quality framework."

OpenAI Description: Leading AI model developer Why mentioned: Same context as Anthropic — enormous valuation share but relatively weak business quality score due to compute dependency Quote: "Anthropic and OpenAI alone account for $1.8 trillion of that, or 22% of the $8.2 trillion universe."

Databricks Description: Data and AI platform company Why mentioned: Cited as a top scorer on PitchBook's Business Quality framework for unicorns, contrasting favorably with model builders Quote: "Databricks and Stripe topped our scoring framework."

Stripe Description: Payments infrastructure company Why mentioned: Also topped PitchBook's Business Quality framework, suggesting infrastructure/fintech unicorns offer stronger fundamental business profiles than model builders Quote: "Databricks and Stripe topped our scoring framework."

River AI Description: Platform helping developers and enterprises train and own custom AI models Why mentioned: Raised $1.1B led by General Catalyst and AMP PBC; notable for its founder pedigree and positioning as a compute-ownership alternative to renting Quote: "River AI, which helps developers and enterprises train and own custom AI models, raised $1.1 billion led by General Catalyst and AMP PBC."

Applied Compute Description: Startup developing customized AI models for businesses Why mentioned: In talks to raise at a $3B valuation; signals strong investor interest in enterprise-customized AI versus general-purpose models Quote: "Applied Compute...is in talks to raise a new round led by Elad Gil at a $3 billion valuation."

Neros Description: California-based defense tech startup Why mentioned: Secured $250M at a $2.5B valuation, led by Sequoia — signals continued strong VC interest in defense tech Quote: "California-based defense tech startup Neros secured a $250 million round led by Sequoia and American Strategic Technology Fund at a $2.5 billion valuation."

DayOne Data Centers Description: Singapore-based data center operator backed by Coatue Why mentioned: Confidentially filed for a US IPO — a potential bellwether for data center infrastructure exits Quote: "Singapore-based data center operator DayOne Data Centers, which is backed by Coatue, confidentially filed for a US IPO."

Mountaintop Beverage / Monogram Capital Partners Description: Beverage co-packer; PE firm closed a $300M continuation vehicle Why mentioned: Illustrates how continuation vehicles are being used in non-tech sectors seen as tricky to enter; Apollo S3 participated as lead investor Quote: "The PE firm closed a $300 million CV for Mountaintop Beverage, luring Apollo S3 as lead investor in a category that's tricky to enter."

Houlihan Lokey Description: Global investment bank Why mentioned: CEO described how AI is simultaneously compressing pricing in their valuations business while expanding total addressable market Quote: "AI is already driving down pricing in some of its businesses, including in its valuations business, one of the largest in the sponsor-backed market."

Evercore Description: Independent investment bank Why mentioned: Chairman/CEO commentary on how AI is prompting strategic buyers toward larger acquisitions, while also generating restructuring opportunities in software Quote: "Evercore's restructuring business is holding more conversations with software companies whose valuations have taken a hit and are considering restructuring their debt."

PJT Partners Description: Independent investment bank Why mentioned: CEO provided the most pointed macro warning: AI will worsen PE's distribution problem and accelerate continuation fund usage Quote: "AI will also exacerbate a problem that private equity firms thought was fading: the challenge of generating distributions for investors."

Layer Global Description: New investment firm launched by former General Atlantic co-president Anton Levy Why mentioned: Raised over $1.1B for its first fund — a notable fundraise for a debut vehicle Quote: "Layer Global, a firm launched by former General Atlantic co-president Anton Levy, raised over $1.1 billion so far for its first fund."

Craft Ventures Description: VC firm founded by David Sacks Why mentioned: Targeting a $1B new fund raise Quote: "David Sacks' Craft Ventures is raising a new fund with a $1 billion target."

4. People Identified

John Weinberg Role: Chairman and CEO, Evercore Why mentioned: Provided insight on how AI is driving two simultaneous M&A dynamics — defensive consolidation among strategic buyers and debt restructuring among software companies Quote: "The rapid spread of AI technology is making strategic buyers question whether they need to make 'stronger and bigger' acquisitions to compete with their peers."

Paul Taubman Role: Chairman and CEO, PJT Partners Why mentioned: Offered the most bearish and contrarian institutional view on AI's impact on private equity exits and distributions Quote: "AI will also exacerbate a problem that private equity firms thought was fading: the challenge of generating distributions for investors."

Scott Adelson Role: CEO, Houlihan Lokey Why mentioned: Described a nuanced dual effect of AI — pricing compression in existing businesses alongside TAM expansion Quote: "AI is already driving down pricing in some of its businesses... At the same time, AI has led to a massive expansion of the total addressable market."

Igor Babuschkin Role: Co-founder, River AI; former co-founder of xAI Why mentioned: Founding pedigree (xAI/Elon Musk's AI company) lends credibility to River AI's $1.1B raise Quote: "The company was founded by Igor Babuschkin, a co-founder of xAI."

Michael Gross Role: Co-founder, SLR Capital Partners Why mentioned: Issued a warning on private credit recovery expectations, particularly for software Quote: "Recoveries are likely to underperform expectations, especially for the software sector."

Anton Levy Role: Founder, Layer Global; former co-president, General Atlantic Why mentioned: Successfully raised $1.1B+ for his debut fund, a significant endorsement from LPs for a first-time manager Quote: "Layer Global, a firm launched by former General Atlantic co-president Anton Levy, raised over $1.1 billion so far for its first fund."

Elad Gil Role: Investor leading Applied Compute's new round Why mentioned: His backing at a $3B valuation signals conviction in the enterprise custom AI model space Quote: "Applied Compute...is in talks to raise a new round led by Elad Gil at a $3 billion valuation."

Franco Granda & Harrison Rolfes Role: Authors, PitchBook Unicorn Tracker report Why mentioned: Produced the analysis underpinning the core thesis of the issue — that unicorn size and business quality are decoupled Quote: Byline attribution on the unicorn tracker analysis

5. Operating Insights

Insight 1: Owning Your Compute Stack Is a Structural Advantage The article makes clear that the core weakness of AI model builders like OpenAI and Anthropic is dependence on rented compute. Companies like River AI are explicitly positioning against this by helping customers "train and own" models. For operators building AI-native businesses, securing owned or dedicated compute infrastructure — rather than remaining wholly dependent on cloud providers — materially improves margin structure and business quality scores.

"The reason has a lot to do with the fact that model builders must rely on rented computing power, which leaves them with thinner margins and less control over their operations."

Insight 2: AI-Driven Pricing Compression Is Coming for Professional Services — Plan for It Houlihan Lokey is already experiencing AI-induced pricing pressure in its valuations business. Operators running advisory, consulting, or any knowledge-work business should anticipate margin compression in commoditizable service lines, while simultaneously identifying which parts of their TAM AI will expand.

"AI is already driving down pricing in some of its businesses, including in its valuations business... At the same time, AI has led to a massive expansion of the total addressable market, driving strong growth even with weaker pricing power."

Insight 3: Continuation Vehicles Are Becoming a Mainstream Liquidity Tool PE firms and their investors should treat continuation funds not as a niche workaround but as a structural feature of the current exit environment, especially as AI disrupts software valuations.

"This could drive the use of continuation funds, which allow PE firms to extend their hold on businesses while providing liquidity to investors."

6. Overlooked Insights

Insight 1: Defense Tech Is Quietly Attracting Top-Tier VC at Scale While AI and climate tech dominate the headlines, two defense tech deals were quietly noted — Neros raising $250M at a $2.5B valuation led by Sequoia, and London-based Pyra looking to raise ~$200M. The involvement of Sequoia (typically consumer/enterprise-focused) in defense signals a broadening of institutional VC appetite for the sector.

"California-based defense tech startup Neros secured a $250 million round led by Sequoia and American Strategic Technology Fund at a $2.5 billion valuation. London-based defense-tech startup Pyra is looking to raise around $200 million."

Insight 2: Smaller LPs Now Have Pathways Into Co-Investment Briefly flagged in the newsletter but potentially significant for the LP community: there are emerging structures allowing smaller LPs to access co-investment (no-fee deals) — historically reserved for large institutional investors.

"You don't need to be a big fish to co-invest. There are ways for smaller LPs to get in on the no-fee investment action."