Source status: Public methodology pages and public source descriptions were prioritized. Commercial databases, paywalled tables, and blocked pages are identified as verification targets rather than treated as fully observed data.
Evidence Boundaries: A league table or index can describe reported performance, assets, lifetime net gains, strategy cohorts, awards, or legal identity. It does not establish AI depth, GenAI deployment, model ownership, investment quality, or performance causality.
There is no single public, transparent league table that simultaneously covers quantitative/systematic hedge funds, strategy focus, emerging managers, and AI/ML depth. The usable approach is a linked evidence model:
- use HFR, Aurum, and With Intelligence for strategy taxonomies and benchmark cohorts;
- use Preqin, HFR, and regulatory registers for manager and fund identity;
- use dated Preqin, With Intelligence, BarclayHedge, or other performance sources for separately labeled return observations;
- use Pensions & Investments, Institutional Investor, and LCH only for their specific AUM, earnings, or lifetime-net-gain measures;
- attach AI evidence from firm pages, technical artifacts, hiring, papers, conference recordings, or vendor/customer disclosures as a separate field.
Publicly usable source families
| Source family | What it measures | Main caveat |
|---|---|---|
| HFR | Fund-level performance, assets, liquidity, descriptions, and strategy classifications | Commercial, reporting-dependent, and not an AI measure |
| Preqin | Dated net-return league tables with explicit eligibility rules; manager and fund profiles | Full-track-record requirements exclude many emerging managers; reported data is not independently verified in the public methodology notice |
| With Intelligence / Eurekahedge | Strategy and regional indices with an explicit AUM screen, equal weighting, caps, and survivorship adjustment | Selected index constituents; legacy Eurekahedge indices are transitioning/sunsetting |
| Aurum | Manager/fund database with AUM, P&L, flows, qualitative data, and nine master strategies including Quant | Proprietary and not audited; Aurum may reclassify or suppress data |
| BarclayHedge | Hedge-fund and managed-futures index series | Index construction and reporting requirements create selection and investability effects |
| HFM / With Intelligence awards | Award categories using absolute performance, volatility, and benchmark-relative results | Entrant, nomination, and database-participation bias; not a full universe |
| Institutional Investor | Highest-earning managers and selected large-manager/AUM lists | Editorial, paywall, estimates, and large-manager bias; not a quant taxonomy |
| Pensions & Investments | Large-manager AUM ranking candidate | The primary page was blocked during retrieval; definitions and survey coverage remain to be verified |
| LCH | Estimated net gains since inception after fees and annual gains for selected managers | Scale, age, survival, and absolute-dollar bias; not a complete manager universe or strategy ranking |
| SEC IAPD, FCA register, AFM registers | Legal identity, registration, permissions, and filings | Regulatory status is not proof of active fund, strategy, AI, or performance |
Recommended data model
Keep these as separate linked records rather than a composite score:
manager -> legal entity -> fund/vehicle -> normalized strategy -> performance observation
-> AUM observation -> league/index membership -> AI evidence -> personnel evidence
Minimum fields are defined in the accompanying raw source ledger at sources/06-industry-verticals/hedge-fund-league-tables-manager-directories-2026-raw.md, including source date, retrieval date, metric, period, eligibility rule, taxonomy, access status, selection bias, AI evidence tier, and disqualification reason.
For AI evidence, use discrete labels—none located, public positioning, hiring intent, named workflow, firm-owned artifact, production claim, and independently evaluated. A league-table position must never upgrade an AI evidence label.
Candidate universe from the TC43 pass
The linked TC43 emerging-manager intake records public leads including Machina Capital, Seldon Capital, Castle Ridge Asset Management, Machine Capital, TradeWell Capital, Bayswater, Sparkline, AXQ, CapitalsAI, and other smaller or newer managers. The intake keeps name collisions, pre-launch structures, platform/adviser businesses, and thinly documented firms explicitly labeled rather than silently merging or excluding them. Each candidate still requires a legal-entity, regulatory, active-vehicle, strategy, and current-status check.
What not to publish
- Do not publish a “best AI hedge fund” table assembled from marketing labels.
- Do not equate systematic, quantitative, or automated with GenAI.
- Do not combine AUM, lifetime net gains, one-year returns, awards, and AI evidence into one score.
- Do not treat a database presence as proof of current employment, active fund status, or production AI.
- Do not treat the absence of a manager from a public table as evidence that the manager is absent from the market.
Source-ledger artifact
The full row-level ledger, source URLs, methodology notes, dates, access limitations, and candidate intake records are maintained in the repository source artifact sources/06-industry-verticals/hedge-fund-league-tables-manager-directories-2026-raw.md.