← shawndimantha.com
AI Rollups · March 2026 · PEF

Investing in
AI Rollups: Lessons & Opportunities

Shawn Dimantha
Shawn Dimantha
Founder · Investor · Product Exec
Alphabet/Google — AI products (YouTube · Verily · X)
Former exited health × AI founder · Hydra Ventures
2014
Pharma holdco models: Roivant, BridgeBio and Andrew Lo
2016
Peterson Center on Healthcare — Hired by Blackstone co-founder Pete Peterson for tech-driven primary care transformation. Secured Gates Ventures investment to scale.
2022+
Hydra Ventures — AI & crypto investment vehicle. First checks in early-stage AI startups. Evolved toward AI rollups for barbell risk-adjusted exposure.
Now
Active AI rollup portfolio — 5+ investments across the spectrum: AI-maximalist holdcos, PE-style service rollups, and pharma.

Four Bets Across the Spectrum

AI Maximalist → Traditional PE
AI Maximalist

Rocketable ↗

YC W25 · $6.5M Seed

Acquire profitable SaaS, replace the human org chart with AI agents. Run software companies with <5 people. Ex-Cruise, ex-Uber ATG founder proved model with first acquisition.

Zero-employee software co
PE / Growth Equity

Legion Holdings ↗

PE & Growth Equity Veterans

Vertical service rollups via SPVs. Ops & eng talent from Invisible Technologies, Bridgewater. First deal in healthcare staffing. Acquire + bolt-on, layer agent automations.

Deal-by-deal SPV structure
Pharma / Biotech

Babylon Bio ↗

Neurodegenerative Focus

Acquire neurodegenerative clinical assets, spin into separate entities. Co-developed SOTA clinical trial prediction model with OpenAI. The Roivant model for the AI era.

SOTA model with OpenAI
Emerging Models

+ More TBA

New Structures · 🎙 PE Funcast

AI transformation cos taking outcome-based risk on what were previously consulting engagements. Skin in the game on business outcomes, not hourly billing.

Outcome-based risk sharing
← Replace humans entirely Augment + transform operations →
What Hasn't Worked

Patterns That Struggle

✕
PE operators without agentic AI domain depth. PE associates and search fund pivots raising to "apply AI" without hands-on experience building or deploying agents. Investors want proven AI-native operators, not just financial modelers with a thesis.
✕
Tech-first founders with limited M&A experience. International and technical founders targeting US acquisitions often struggle to raise at reasonable dilution levels without in-market deal networks and relationships.
✕
Team retention risk. Sequence Holdings lost co-founder Neal Wu (to Thinking Machines) before their $75M raise. The AI talent market is brutal.
✕
Dilution from mega-funds. GC-style deals taking ~50% upfront. Ithaca ($1B+), Thrive ($1B+), Thoma Bravo, Vista all active in the space.
Validation & Challenges
⚡
OpenAI & Anthropic PE JVs — Foundation model cos partnering directly with PE firms. Validates the thesis, but may disintermediate smaller AI rollup teams — or become a tailwind as infrastructure access improves.
Key formula for AI rollups success: founding team / market fit + reasonable entry multiples + domain depth + room to grow in white space.
shawndimantha.com/investing-ai-rollups.html
White Space & Opportunities

Where to Look Next

→
Underserved service verticals — Insurance brokerage, healthcare, wealth mgmt, tax, marketing agencies. Strong PE attention but less AI rollup competition.
→
Neofirms — Build AI-native service cos from scratch vs. acquiring legacy. Crosby (Sequoia-backed AI law firm) as template; Eudia for AI-native to rollup evolution.
→
Public company plays — Acquire meaningful stake + board seat, drive AI transformation from that position. Ithaca Holdings signaling this approach.
Let's Talk
Looking to deepen my footprint in the space by leaning into more actively building AI rollups and holdcos. If that resonates, reach out.
✉️ Email me LinkedIn 🗺️ My AI Rollups Market Map