Skip to content
    CrossWorkRequest access
    Illustration: a bright semiconductor fab corridor

    Portfolio

    Late-stage companies that went on to define their categories.

    Early positions in companies that went on to define their categories. Shown as text, for identification only.

    Portfolio · Cerebras

    Wafer-scale compute. Cerebras builds the processor the size of the wafer.

    Most AI chips are cut from a wafer; Cerebras keeps the whole wafer as one processor.

    Cerebras: one wafer, not one chip

    Interactive: switch the compared measure
    Compare
    WSE-3 Turbo · one wafer
    Cerebras WSE-3 Turbo
    46,225 mm²
    NVIDIA H100 die
    814 mm²
    Illustration: macro view of a silicon wafer surface

    The wafer is about 57 times the area of a leading GPU die, drawn here to true relative scale.

    Keeping a model's working data on one wafer avoids shuttling it between chips; Cerebras says its systems run inference up to 30 times faster than GPU systems. For buyers, that is cost and speed per token.

    Illustrative geometry, sourced figures.

    Sources
    1. 18. Cerebras: Wafer-Scale Engine 3 Turbo (WSE-3T) specifications, product page.
    2. 19. NVIDIA, 2022: NVIDIA H100 Tensor Core GPU Architecture whitepaper.

    Portfolio · Stripe

    Payments at national scale. Stripe moves more money than many economies.

    Stripe's 2025 total volume, ranked against national GDP on one shared scale.

    If Stripe's 2025 volume were an economy

    1. Stripe volume$1.9T
    2. Canada$2.3T
    3. Spain$1.9T
    4. Australia$1.8T
    5. Mexico$1.8T
    6. Indonesia$1.4T
    7. Netherlands$1.3T
    8. Saudi Arabia$1.3T
    Stripe volume, 2025
    1.9 trillion dollars of total volume in 2025
    Illustration: a customer paying by phone at a café counter

    Equivalent to roughly 1.6% of global GDP, per Stripe.

    Payments infrastructure compounds with the economy it serves, and agents will soon be paying for things too.

    Nominal GDP, IMF WEO April 2026; Stripe total payment volume 2025, per Stripe. Volume processed is not revenue.

    Illustrative geometry, sourced figures.

    Sources
    1. 21. Stripe, February 24, 2026: Stripe 2025 annual letter.
    2. 33. IMF, April 2026: World Economic Outlook database, GDP at current prices (NGDPD), US$ billions, 2025.

    Portfolio · Perplexity

    Search, rebuilt. Perplexity turns search into answers.

    The same question asked two ways, and who does the reading and assembling in each.

    Perplexity: an answer engine, not a list of links

    Illustration: a bright, modern office floor
    • Search today · you do the work: Your question; Ten links to open; You read and compare; You write the answer.
    • Perplexity · it does the work for you: Your question; It reads the pages; One cited answer; Pages it cites.

    The work of reading and assembling moves from the person to the product. That is a different business from selling clicks.

    Illustrative geometry, sourced figures.

    In brief

    What it does, who buys it, why it compounds

    Three more holdings, each described in three short lines.

    AuditBoard

    What it does
    A connected platform for audit, risk and compliance teams.
    Who buys it
    Internal audit, risk and compliance leaders.
    Why it compounds
    Once a company's controls live in one system of record, every new regulation adds work to that system.

    More than 2,000 enterprises, including nearly 50% of the Fortune 500, at the time of the Hg agreement.

    Salsify

    What it does
    Product experience management: the product content brands publish to retailers and marketplaces.
    Who buys it
    Consumer brands and manufacturers.
    Why it compounds
    Every new retailer and marketplace multiplies the content a brand must keep correct.

    More than 2,000 customers, including many leading consumer brands, per Cinven.

    Neo4j

    What it does
    The graph database: data stored as connected entities rather than tables.
    Who buys it
    Engineering and data teams in large enterprises.
    Why it compounds
    Agents and retrieval work better on connected data, which puts graphs under more AI applications.

    Used by 84% of Fortune 100 companies, per Neo4j.

    Sources
    1. 22. Business Wire, May 23, 2024: AuditBoard agrees to be acquired by Hg.
    2. 12. Cinven, July 22, 2026: Cinven to acquire Salsify.
    3. 23. Neo4j, November 19, 2024: Neo4j surpasses revenue milestone (press release).

    Holdings

    Held, realized and exited. Every holding, grouped by status.

    Held first, then partially realized, announced, exited or restructured, shown by name and area.

    • Neo4j

      Held

      The graph database.

    • Qumulo

      Held

      Hybrid cloud file data storage.

    • SambaNova

      Held

      Enterprise AI systems.

    • DataRobot

      Held

      Enterprise machine learning.

    • ThoughtSpot

      Held

      AI-powered analytics.

    • Course Hero

      Held

      Study resources for students.

    • Perplexity

      Held

      An answer engine for the open web.

      See the explainer ↓

    • Verbit.ai

      Held

      AI transcription and captioning.

    • Aviatrix

      Held

      Cloud network security.

    • BigID

      Held

      Data security, privacy and AI governance.

    • Tanium

      Held

      Autonomous endpoint management.

    • Attentive

      Held

      AI marketing across SMS and email.

    • Rappi

      Held

      On-demand delivery across Latin America.

    • Stripe

      Held

      Financial infrastructure for the internet.

      See the explainer ↓

    • Tealium

      Held

      Real-time customer data platform.

    • Flexport

      Held

      Technology for global supply chains.

    • Stord

      Held

      Commerce fulfillment and logistics.

    • Tradeshift

      Held

      E-invoicing and procurement.

    • Invoca

      Held

      Conversation intelligence for revenue teams.

    • Outreach

      Held

      AI sales execution platform.

    • Podium

      Held

      AI for local businesses.

    • AngelList

      Held

      Software for startups and venture funds.

    • CityBlock Health

      Held

      Value-based care for underserved communities.

    • Response Pharmaceuticals

      Held

      Weight management for high-need patients.

    • Cerebras

      Partially realized

      IPO, May 2026.

      See the explainer ↓

    • Scale AI

      Partially realized

      Meta's 2025 investment in Scale AI.

    • Salsify

      Announced

      Agreed to be acquired by Cinven, 2026 (pending).

    • DataStax

      Exited

      Acquired by IBM, 2025.

    • Reltio

      Exited

      Acquired by SAP, 2026.

    • SingleStore

      Exited

      Growth buyout led by Vector Capital, 2025.

    • AuditBoard

      Exited

      Acquired by Hg, 2024.

    • Thrasio

      Restructured

      Chapter 11 restructuring, 2024.

    Company names and logos are trademarks of their respective owners and are shown to identify holdings. No endorsement is implied.

    Sources
    1. 1. CNBC, May 13, 2026: Cerebras prices IPO above expected range.
    2. 29. Bloomberg, June 13, 2025: Meta finalizes $14.3 billion Scale investment, hires its CEO.
    3. 11. Salsify, July 22, 2026: Cinven to acquire Salsify, Inc. (press release).
    4. 12. Cinven, July 22, 2026: Cinven to acquire Salsify.
    5. 28. Database Trends and Applications, May 2025: IBM officially closes acquisition of DataStax.
    6. 31. SAP News Center, May 2026: SAP completes acquisition of Reltio.
    7. 32. SingleStore, October 16, 2025: SingleStore and Vector Capital close growth buyout to accelerate Enterprise AI Innovation.
    8. 22. Business Wire, May 23, 2024: AuditBoard agrees to be acquired by Hg.
    9. 36. Bloomberg Law, February 28, 2024: Thrasio files for Chapter 11 in bid to restructure its debt.

    Exposure

    Two ways to own the shift. AI-native companies, and the businesses AI improves.

    The fund also invests in companies that benefit from AI, not only AI-native ones, so the portfolio is not tied to the AI cycle alone.

    a.

    AI-native infrastructure and platforms

    The companies building compute, models, data and agent platforms.

    AI compute · AI models · AI data · data infrastructure

    Illustration: an electrical substation feeding a data center

    b.

    AI beneficiaries

    Businesses whose margins and growth improve because of AI: cost-savings and productivity beneficiaries, and AI-enabled incumbents.

    Commerce software · risk and compliance software · logistics · payments

    Illustration: a liquid-cooling manifold between server racks

    Categories

    Where we are looking now. Four categories, each with a liquidity path.

    What a leader in each category looks like, and how it could reach liquidity.

    Representative categories, not commitments

    AI model and platform

    IPO or tender path

    Thesis

    Frontier models are becoming an input to most knowledge work. A small set of labs and platforms sets the pace for everyone building on top of them, and spend is concentrating where model quality meets distribution. We think this layer captures value now because the leaders are turning model quality into enterprise relationships and daily usage that are hard to displace.

    What a leader looks like at this stage

    We look for companies valued at $1B+ with $100M+ in revenue and a visible path to liquidity: an IPO, a tender or a strategic sale.

    • Enterprise depth: large organizations running the platform in production across several functions, not in pilots.
    • Gross-margin shape: margins that hold or widen as inference costs fall, because pricing tracks value delivered rather than tokens served.
    • Distribution: a consumer or developer surface that feeds enterprise demand, and partnerships with the clouds and devices people already use.
    • Moat: proprietary usage signal, research talent and the capital to keep training at the frontier.

    Route to liquidity

    1. IPO
      • Cerebras · IPO, May 2026
    2. Tender
      • Company-run tenders for employees and existing holders
    3. Strategic sale
      • Meta · Scale AI, strategic investment, June 2025

    Reference companies

    Named for reference only; not holdings or recommendations, except where marked as a holding.

    Application leaders
    Perplexity (a holding)

    Market context

    • PwC estimates generative and agentic AI could add US$2.6 to 4.4 trillion a year to global GDP by 2030.
    • 40%Gartner expects 40% of enterprise applications to embed task-specific AI agents by the end of 2026, up from under 5% in 2025.

    What we watch

    • Model-cost deflation: price per token falling faster than usage grows.
    • Dependence on a few compute suppliers and cloud partners.
    • Regulatory and copyright pressure on training data and model release.
    Sources
    1. 1. CNBC, May 13, 2026: Cerebras prices IPO above expected range.
    2. 29. Bloomberg, June 13, 2025: Meta finalizes $14.3 billion Scale investment, hires its CEO.
    3. 5. PwC Middle East, 2024: Agentic AI, the new frontier in GenAI.
    4. 27. Gartner, August 26, 2025: Gartner predicts 40% of enterprise apps will feature task-specific AI agents by 2026, up from less than 5% in 2025.

    Portal

    The full list is in the portal. Holdings vary by vehicle.

    Inclusion here is not a recommendation, and past performance is not indicative of future results.