Original Research · India · Currently fielding
The 2026 India Family Office Technology Benchmark
India went from roughly 45 family offices in 2018 to more than 300 today, at around $30 billion. Nobody has measured what they actually run on. This is that survey.
Status
This survey is fielding. No findings are published yet. Every figure below is a marked placeholder, not a result. They are shown so the method and the reporting rules are public before the data exists, which is the only order that makes a benchmark trustworthy. Nothing here should be cited as a finding.
Technology spend
All-in annual spend covering software, implementation and dedicated staff, expressed both absolutely and as a share of AUM. Nobody publishes this figure for the Indian market, which means every office currently benchmarks its budget against nothing.
Stack composition
What is actually used for consolidated reporting, whether an alternatives-data tool sits in front of it, and how custodian and PMS positions reach the system: automated feed, manual upload, or re-keying. The last of those is more common than the category admits.
Reporting cycle time
Days from period close to a consolidated view the family trusts, how many people it takes to produce, and whether it survives the most knowledgeable person being unavailable for a month.
Decision ownership
Who actually signs off technology purchases, whether anyone's job includes technology at all, and whether decisions have been deferred because nobody felt qualified to make them.
AI adoption
What is in production rather than in a pilot, what it is used for, and whether a written policy exists on what family data may be entered into these tools.
Data governance
Whether access is defined by role or by whoever has the shared drive link, and whether the office could produce a record of who saw which entity's data last quarter.
—
Median technology spend as a share of AUM
Q8, banded by AUM · publishable at n ≥ 40
—
Share with no consolidated reporting system
Q10 · publishable at n ≥ 30
—
Median reporting cycle time from period close
Q15 · publishable at n ≥ 30
—
Share where the cycle depends on one person
Q17 · publishable at n ≥ 30
—
Share with nobody owning technology, even partly
Q19 · publishable at n ≥ 30
—
Share with a written policy on AI and family data
Q23 · publishable at n ≥ 30
—
Share that cannot produce a data-access record
Q24 · publishable at n ≥ 40
—
Platform implementation completion rate
Q14, among those who attempted · publishable at n ≥ 40
- →Medians with an interquartile range, never means. One $2B office would drag every average.
- →Any cut with fewer than 8 respondents is described in words, never given as a percentage.
- →n is stated on every figure, on this page and in the report.
- →The instrument is published alongside the findings. A benchmark whose questions are not visible is a marketing asset, not research.
Sample target is 40 or more offices, drawn across Chennai advisory relationships, the Tykhe LP network, and IVCA and family office forums, so no single network dominates. Responses are confidential and only aggregates are published. No participating family is named, ever.
Run a family office in India? Take the survey.
Eight minutes. Responses are confidential, only aggregates are published, and every respondent gets the findings before they are public.
Background on why this gap exists: India went from 45 family offices to 300+.