The Abu Dhabi Investment Authority does not disclose its size, its holdings, its actual weights, a benchmark or a target return. What it publishes is a set of allocation ranges that deliberately do not add to 100%, and two performance numbers: 6.6% a year over 20 years and 7.2% over 30, both to 31 December 2025. In 2025 it widened private equity to 15% to 20% and cut real estate to 2% to 7%.
Key numbers, and what they mean
6.6%
ADIA's 20-year annualised return to 31 December 2025, with 7.2% over 30 years. Both are realised results, reported against no benchmark, and neither is a forecast of anything.
15% to 20%
Its new private equity range, widened in 2025 from 12% to 17%. Changing the band is the decision it discloses. It does not say where inside the band the portfolio actually sits.
2% to 7%
Its new real estate range, cut from 5% to 10%. ADIA says the cut reflects other asset classes growing rather than any selling, and that its absolute exposure to real estate held steady.
57%
The share it manages internally, up from 55% in 2022, with 63% of the portfolio actively managed. Even its most specific disclosures describe process rather than positions.
Allocation ranges, 2025
Developed equities32% to 42%
Private equity15% to 20%
Government bonds7% to 15%
Emerging market equities7% to 15%
Financial alternatives7% to 12%
Real estate2% to 7%
ADIA also publishes ranges for small-cap equities, credit, infrastructure and cash. Its own footnote: these are ranges within which allocations can fluctuate, and so they do not total 100%.
The read-through
From the report
Takeaway: how to apply it to a portfolio
ADIA's footnote reads that these are ranges within which allocations can fluctuate, and hence do not total 100%. The minimums sum to about 75% and the maximums to about 135%.
A policy that sums to exactly 100% forces a trade every time anything moves, because every overweight is somebody else's underweight. Writing the policy as bands turns a constant rebalancing argument into an occasional governance question.
ADIA changed exactly three ranges in 2025, and said the real estate cut reflected the relative growth of other asset classes rather than a reduction in the amount invested.
Separate moving within a band from moving the band itself, and give the two different approvers. Note too the discipline of saying when a weight fell because the denominator grew. Most reports let the reader assume something was sold.
ADIA publishes no size, no actual weights, no benchmark and no target, and reports only a 20-year and a 30-year return.
Bands buy management a great deal of freedom, and they cost the reader the ability to judge it. If you adopt the structure, supply internally the accountability ADIA declines to supply externally, or being inside the range becomes the answer to every question.
The Argus lensADIA widened its private equity band and is moving into private credit as banks pull back. Widening a band is the easy part. Knowing whether the specific fund in front of you earns its fee against a cheap public alternative is the hard part, and that is the part we do.
Benchmark your deals in the Portal →
About the source
ADIA's 2025 Review, published September 2026, is its only substantive public document. It provides the allocation ranges, the 20-year and 30-year returns to 31 December 2025, the internal and active management shares, and the commentary on private credit and systematic strategies. ADIA has never published its assets under management; outside estimates cluster around $1 trillion and disagree with each other by roughly a quarter.