FROM THE 2026 ANNUAL REPORT · SECTOR SUMMARY
What the Asset Owners sector shows
Among the principal stewards of capital, disclosed control sits at the model layer and remains, at the typical level, partial. The median posture across the sector reaches an evolving stage only in how models are operated, with the surrounding cells largely undisclosed. The range within the sector is wide: a small number of stewards disclose named internal control mechanisms — staged review, confidence thresholds, mandatory human checks — that reach an evidenced-control standard, while others disclose little about their own use of AI even as they engage actively with the technology as an investment theme or as a matter of oversight of the companies they hold. That distinction matters for this report: stewardship of AI in portfolio companies, however sophisticated, is not the same as control of the steward's own AI, and only the latter is read here.
Heat-map classifications reflect publicly available information reviewed under the methodology described in this report. They are not assessments or certifications of any institution’s actual internal AI capabilities or controls. Grey (Not Disclosed) indicates the absence of public disclosure, not the absence of control.
The report does not evaluate, rate, certify, or benchmark any individual institution; the tiers reflect the completeness of public disclosure as our review found it, not an assessment of any institution's actual controls.
STEWARDSHIP IS NOT A PASSIVE ACTIVITY
STEWARDSHIP IS NEVER PASSIVE. NEITHER IS AI CONTROL.
Asset owners govern USD 119 trillion in capital across the global financial system. The fiduciary obligations attached to that capital — to beneficiaries, sovereign citizens, policyholders, families, and institutional missions — predate every commercial contract with every AI provider. The institutions that govern their AI with the same precision they govern capital will lead. The ones that do not will operate at the permission of those who do.
.png/:/rs=w:1240,cg:true,m)
The asset owner's AI mandate rests on three obligations no other institutional category faces simultaneously.
Fiduciary accountability to beneficiaries who cannot protect themselves. Regulatory obligations that demand explainability and auditability under the strictest interpretive standards. A governance cascade that turns the institution's AI posture into the standard every manager, advisor, and counterparty must meet.
Each asset owner category — pension funds, sovereign wealth funds, insurance companies, family offices, endowments and foundations — faces a distinct version of that mandate.
Pension Funds
USD 68.3T globally. The Total Portfolio Approach is reshaping governance as portfolio complexity intensifies.
Sovereign Wealth Funds
USD 27T. The active management revival is reasserting strategic intent across geopolitical volatility.
Insurance Companies
USD 23T. 87% restructuring operating models as private credit and infrastructure debt redefine the asset mix.
Endowments & Foundations
USD 944B across 657 institutions. Mission-driven spending pressures intersect with private market complexity.
Family Offices
USD 651B covered. The largest wealth transfer in history is accelerating governance professionalization.
THE AI CONTROL GAp
THE AI CONTROL GAP
43% of large asset owners now identify AI as the single most influential macro factor over the next five to ten years. 69% have neither implemented nor begun developing an AI policy.
AI is contributing to investment decisions, risk assessments, manager selection, compliance determinations, beneficiary communications, and operational execution. For most asset owners, the systems making those contributions are running on provider infrastructure under standard commercial terms — not under fiduciary, regulatory, or institutional control.
The infrastructure is being built faster than the governance.
— Mercer 2025 Large Asset Owner Barometer (74 institutions, USD 2T+ AUM)
WHAT WE DO FOR ASSET OWNERS
AI Control Assessment
A structured diagnostic built on the proprietary 5×5 Control Matrix that scores the institution's current control posture across five AI ecosystems (Power, Compute, Data Centers, Models, Agents) and five pillars of control (Jurisdictional, Logical, Technical, Operational, Contractual). The output is a sector-benchmarked control profile that identifies exactly which intersections are exposed and where investment will have the greatest impact.
Strategic SCENARIO PLANNING Engagements
Translation of diagnostic findings into operating model design, governance frameworks, and board-ready strategy. Built on the Oxford Scenario Planning Approach to stress-test the institution's AI direction against multiple plausible futures.
Architecture Retainers
Ongoing technical control as institutions scale their AI deployments. The Institutional AI Stack™ defines what to build; OLTAIX™ enforces it as the control tower across the institution's AI infrastructure.
REQUEST The State of AI Control in Institutional Finance
The State of AI Control in Institutional Finance is available on a relationship basis to qualifying
All discussions covered under NDA. Tiers reflect public-disclosure completeness, not assessments of any institution's actual controls.
AI is a given. Control is not.™
