AI's musical chairs | PitchBook News Summary

1. Key Themes

AI Capital Is Concentrating Into Fewer, Larger Bets

The gap between AI's share of VC dollars vs. deal count is widening dramatically, signaling that capital is flowing to a handful of winners rather than spreading broadly across the ecosystem.

"By 2026, AI accounts for 76.7% of all VC dollars but just 35.9% of deals as capital concentrates."

And the velocity is striking: Q1 2026 alone has already surpassed all of 2025's AI funding.

"Q1 funding for the vertical has already eclipsed 2025's full-year total, reaching $255.5 billion. Driven by historic fundraising by companies like OpenAI and Anthropic, only three deals accounted for 67% of the vertical's total funding."

AI Infrastructure (Physical Layer) Is the New Frontier for Expansion-Stage Capital

Capital is no longer just chasing software — it's moving into the physical systems that make AI run: chips, energy, transportation, and commercial products.

"Expansion-stage deal value across semiconductors, energy, commercial products, and commercial transportation reached record highs, reflecting rising investor focus on the capacity and infrastructure required to sustain digital transformation. Collectively, expansion-stage dealmaking across those four industries exceeded $44 billion in 2025, a 69 percent increase from 2024."

Wealth Managers Are Disintermediating Alt Asset Managers

A structural shift is underway in private markets distribution: sophisticated RIAs are no longer just passing client capital to third-party fund managers — they're standing up their own GP vehicles.

"An ascendant group of wealth managers is cutting out the middleman, standing up their own institutional-grade drawdown fund vehicles and seeding them with capital from ultra-high-net-worth investors."

This is being enabled by scale: firms that previously lacked the AUM to hire dedicated private markets talent can now do so.

"These are firms that have the economics where they can actually hire dedicated, experienced, professionally trained private markets experts when it comes to sourcing and underwriting private investment vehicles. You didn't see that previously because these firms were all sub-scale." — Don Calcagni, CIO, Mercer Advisors

Female-Founded Startups Are Having a Breakout Fundraising Moment

The Q2 2026 numbers represent a historically rare milestone for female-led ventures.

"Startups with at least one female founder had a buzzworthy Q2, raising $73.2 billion across 622 deals" — described as "their second-best quarter on record."

Big Tech IP Wars Are Coming for AI

Apple's lawsuit against OpenAI signals that hardware trade secret litigation will become a recurring feature of the AI talent wars — mirroring the autonomous vehicle IP battles of the prior decade.

"Apple sued OpenAI, alleging it stole hardware trade secrets. The legal battle, involving former Apple employees who have joined OpenAI, harkens back to Google's similar lawsuit against Uber in 2017 over autonomous driving tech."

2. Contrarian Perspectives

AI deal count is actually shrinking as a share of VC, even as dollar dominance grows. The conventional framing is that "AI is eating venture." But the data shows it's eating dollars, not necessarily deals. AI now commands 76.7% of VC dollars but only 35.9% of deals — meaning the vast majority of venture transactions are happening outside AI, while a small number of mega-rounds distort the dollar figures. Founders and investors outside AI may face less competition for deals than headlines suggest.

"AI and ML's deal value and deal count shares of global VC activity tracked closely through 2022. Since 2023, they've diverged sharply; by 2026, AI accounts for 76.7% of all VC dollars but just 35.9% of deals as capital concentrates."

The "wealth channel opportunity" for alt managers may already be closing — because wealth managers are becoming GPs themselves. The consensus view is that alt managers (Blackstone, Apollo, etc.) will capture trillions from retail wealth via intermediary RIAs. But the emerging reality is that large RIAs are bypassing those managers entirely.

"The evolution of wealth advisers into a new breed of GP highlights the effects of two closely watched trends: the private market's growing dependence on the wealth channel, and wealth advisers' newfound capability to run capital-allocation strategies on a par with professional institutional investors."

The irony: the wealth channel that alt managers spent years building pipelines to access is now competing with them directly.

"Alternative asset managers have spent years building pipelines to tap into the adviser-led wealth channel... But now a quiet reversal is underway."

Three deals drove two-thirds of all AI VC funding in Q1 2026 — raising questions about whether "AI VC" is a category or just a handful of bets. The breadth implied by "$255.5 billion in AI funding" is misleading given that just three transactions accounted for 67% of it. The AI VC market is, in practical terms, dominated by a tiny number of foundation model bets — not a broad ecosystem of investable companies.

"Only three deals accounted for 67% of the vertical's total funding."

3. Companies Identified

OpenAI Description: Leading generative AI company Why mentioned: Named as a historic fundraising driver in Q1 2026, and as defendant in Apple's hardware trade secret lawsuit

"Driven by historic fundraising by companies like OpenAI and Anthropic, only three deals accounted for 67% of the vertical's total funding." "Apple sued OpenAI, alleging it stole hardware trade secrets."

Anthropic Description: AI safety-focused large language model company Why mentioned: Cited as one of the historic fundraisers driving AI VC concentration in Q1 2026

"Driven by historic fundraising by companies like OpenAI and Anthropic, only three deals accounted for 67% of the vertical's total funding."

Mercor Description: San Francisco-based startup providing specialized data to train AI models Why mentioned: Reportedly in discussions for a new funding round at a $20 billion valuation; also separately mentioned for its novel data-generation approach (contractors role-playing office workplaces for AI training) and for acquiring Deeptune

"San Francisco-based startup Mercor, which provides specialized data to train AI models, is discussing a new funding round at a $20 billion valuation." "Mercor is getting contractors to role-play entire workplaces for AI to study. Employees act out whole office dynamics so models can finally answer: What does a banker actually do all day?"

Mercer Advisors Description: $115 billion RIA catering to wealthy clients Why mentioned: Case study for the trend of large wealth managers building in-house private markets capabilities and bypassing traditional alt managers

"These are firms that have the economics where they can actually hire dedicated, experienced, professionally trained private markets experts when it comes to sourcing and underwriting private investment vehicles."

Apollo Global Management Description: Major alternative asset manager Why mentioned: Submitted an unsolicited £5.7 billion bid for easyJet, trumping a rival offer; also acquired a minority stake in Bayer's contraceptives unit for €3 billion

"Apollo Global Management has gatecrashed the takeover of easyJet with an unsolicited £5.7 billion ($7.7 billion) bid."

EQT Description: Global private equity firm Why mentioned: Acquired Copia Power (energy campuses for AI infrastructure) from Carlyle for $2.6 billion — a notable AI infrastructure play

"EQT agreed to acquire Copia Power, a developer of energy campuses for high-voltage transmission and AI infrastructure, from The Carlyle Group in a $2.6 billion deal."

Copia Power Description: Developer of energy campuses for high-voltage transmission and AI infrastructure Why mentioned: Subject of a $2.6B acquisition by EQT from Carlyle, illustrating PE appetite for AI physical infrastructure

"EQT agreed to acquire Copia Power, a developer of energy campuses for high-voltage transmission and AI infrastructure."

Databento Description: Market data platform for trading and financial firms Why mentioned: Raised a $97 million Series B led by NEA

"Databento, which provides a market data platform for trading and financial firms, raised a $97 million Series B led by NEA."

Warburg Pincus / PantherRx Rare Description: Specialty pharmacy focused on rare diseases; PE target Why mentioned: Warburg Pincus reportedly in talks to acquire PantherRx from General Atlantic for over $7 billion

"Warburg Pincus is in talks to acquire specialty pharmacy company Pantherx Rare from investors including General Atlantic for over $7 billion."

XGS Energy Description: Geothermal energy startup Why mentioned: In talks to pursue an IPO — a notable exit signal for the energy-for-AI infrastructure theme

"XGS Energy, a geothermal energy startup, is in talks to pursue an IPO."

Fuchs & Eule Description: Berlin-based AI-powered retrofit platform for energy-efficient buildings Why mentioned: Raised €10M, illustrating European VC activity at the intersection of AI and climate tech

"Fuchs & Eule, a Berlin-based provider of an AI-powered retrofit platform for energy-efficient buildings, raised €10 million in a round led by GET Fund."

HarbourVest Partners Description: Global private markets investment firm Why mentioned: Closed Co-Investment Fund VII on $4.75 billion — a significant fundraising data point

"HarbourVest Partners closed its Co-Investment Fund VII Program on $4.75 billion."

Sequoia Capital Description: Leading venture capital firm Why mentioned: Retains the #1 spot as the most active AI investor for the fourth consecutive year

"For the fourth year in a row, Sequoia Capital retains the top spot as the most active investor in AI."

4. People Identified

Don Calcagni Description: Chief Investment Officer, Mercer Advisors ($115B RIA) Why mentioned: Quoted as a primary voice explaining why scaled RIAs are now building in-house private markets teams and bypassing traditional alt managers

"These are firms that have the economics where they can actually hire dedicated, experienced, professionally trained private markets experts when it comes to sourcing and underwriting private investment vehicles. You didn't see that previously because these firms were all sub-scale." "The entire industry is focused on the wealth channel."

Jacob Robbins Description: Technology Reporter, PitchBook News Why mentioned: Author of the AI investor rankings piece (Byline attribution; no direct quote.)

Alec Davis Description: Head of Enterprise Reporting, PitchBook News Why mentioned: Author of the wealth managers/private markets piece (Byline attribution; no direct quote.)

5. Operating Insights

1. Scale your RIA to unlock private markets GP economics — the sub-scale era is over. The article makes clear that the new competitive moat for wealth managers is size: firms that have crossed the threshold where they can afford dedicated private markets professionals are now running their own drawdown vehicles, cutting out the fund-of-fund layer entirely. For RIA operators, the strategic implication is consolidation or organic growth to reach the AUM level where institutional-grade hiring becomes viable.

"These are firms that have the economics where they can actually hire dedicated, experienced, professionally trained private markets experts... You didn't see that previously because these firms were all sub-scale."

2. For AI startups, the data layer is where the next $20B valuations are being built. Mercor's trajectory — a company that generates synthetic training data by having contractors role-play workplace scenarios — illustrates that the bottleneck in AI development has shifted from models to the quality and specificity of training data. Founders and operators building in data generation, curation, and domain-specific simulation are operating in a high-value niche.

"Mercor is getting contractors to role-play entire workplaces for AI to study. Employees act out whole office dynamics so models can finally answer: What does a banker actually do all day?" "San Francisco-based startup Mercor... is discussing a new funding round at a $20 billion valuation."

3. Physical infrastructure is the next large VC/PE allocation — don't wait for software multiples to compress. With expansion-stage deal value across semiconductors, energy, transportation, and commercial products surging 69% YoY to $44B+, operators and investors in these sectors should expect intensifying competition for assets and talent. The window to enter at reasonable valuations may be narrowing.

"Collectively, expansion-stage dealmaking across those four industries exceeded $44 billion in 2025, a 69 percent increase from 2024."

6. Overlooked Insights

1. European VC is quietly strengthening — not just a US/AI story. Buried in the "Catch Up Quick" section, European VC dealmaking is described as strengthening in H1 2026, driven by AI adoption and larger round sizes. This is notable given the persistent narrative of European VC lagging the US. The deals section corroborates this with multiple European rounds (Fuchs & Eule, Bohr Energie, Marker in London, Andera Partners closing above its hard cap). For investors, Europe may represent a relative-value opportunity as AI tailwinds reach a less crowded market.

"Europe's VC dealmaking strengthened in the first half of the year, driven by continued AI adoption and larger funding rounds."

2. UHNWIs are being re-categorized as institutional-equivalent investors — with major implications for how private markets products are structured and marketed. The article notes a subtle but important reclassification: ultra-high-net-worth individuals are increasingly being treated as institutional clients by large wealth managers, not merely as wealthy retail. This has downstream implications for product design (drawdown structures vs. evergreen), fee models, and regulatory positioning.

"As their wealth has grown, UHNWIs are increasingly attracting the attention of the biggest wealth managers due to their similarities to institutional clients in terms of risk appetite, market savvy, investable assets and investment time horizons."