The 4,000 figure is McKinsey’s: the number of single family offices in Hong Kong and Singapore combined, which it says has quadrupled since 2020 to about 4,000. The official counts have since moved past it. A Deloitte study commissioned by InvestHK put Hong Kong at more than 3,380 SFOs at end-2025, according to the Hong Kong government; MAS says more than 2,000 Singapore SFOs were receiving tax incentives at end-December 2025. Add the two separate counts and the two-hub total is above 5,380 — before counting Tokyo, Mumbai, Dubai’s Asian-affiliated structures, or other regional centres. The two counts measure different things (a market study versus incentive awards), so the sum is an order-of-magnitude guide, not a census. Either way, the two-hub SFO count is well above its 2020 level, and the trajectory shows no sign of slowing. The rest of this note uses 4,000 as the milestone it marks; the arguments apply with more force to the 5,000-plus count.
Source: Asia Capital Notes illustrative estimate; neither InvestHK nor MAS publishes an AUM-tier breakdown
The headline framing is that Asia has reached a tipping point as the global gravity center for family office activity. There is something true about this framing. There is also a lot it obscures about what the 4,000 number actually represents, what kind of wealth sits behind it, and what practitioners should expect from the next phase of the regional growth.
This piece reads the tipping point through three observations about what 4,000 actually means, and what it does not.
Observation 1: 4,000 SFOs is not 4,000 wealth families
The most important read on the 4,000 number is that it is a count of structures, not a count of wealth-holding families. The two numbers diverge meaningfully in both directions.
Some wealth-holding families operate multiple SFO structures — typically because they have separate vehicles for different generations, separate vehicles for different asset classes, or separate vehicles in different jurisdictions for different tax-residency family members. A single ultra-high-net-worth family with global assets might be represented by 3-7 separate SFO structures across Hong Kong, Singapore, and other jurisdictions. The structural multiplication factor for the wealthier end of the spectrum is real and meaningful.
In the other direction, a substantial portion of the 4,000 structures are at relatively modest AUM levels — USD 30-100M range — and represent the family wealth of one or two families. These are the structures that the post-2023 substance updates in Singapore selected for — the mid-tier cohort I unpack in Singapore’s family office boom — and that Hong Kong’s 2023 FIHV regime has been attracting in significant volume. They count as one SFO each, but they each represent a single wealth-holding family without significant structural multiplication.
Translated, on my rough estimate: 4,000 SFOs probably represents somewhere between 2,500 and 3,200 distinct wealth-holding families. Put the other way, distinct families number roughly 20-37.5% fewer than the headline 4,000 once you account for structural multiplication, while the headline understates the count of structures held by each substantial family.
For LP and B2B service-provider sizing of the addressable market, this distinction matters substantially. The number of distinct family decision-units is the relevant metric for pipeline sizing; the number of structures is the relevant metric for fund admin and compliance volume. Conflating the two leads to systematically wrong market-sizing assumptions.
Observation 2: The wealth-mass behind 4,000 is concentrated, not distributed
The second important read is that the 4,000 SFO count does not correspond to a uniform wealth distribution. The underlying AUM is heavily concentrated in a small number of very large structures, with a long tail of mid-tier and smaller offices.
No official source publishes an AUM-tier breakdown, so the table below is an illustrative estimate rather than reported data. Read against McKinsey’s estimate that Asia-Pacific HNW and UHNW families will pass on USD 5.8 trillion between 2023 and 2030, the AUM distribution looks roughly like:
| AUM band | Approx count | Share of total AUM |
|---|---|---|
| USD 1B+ | ~150-250 | ~50-60% |
| USD 200M-1B | ~600-900 | ~25-30% |
| USD 50M-200M | ~1,500-2,200 | ~10-15% |
| USD 30M-50M | ~800-1,200 | ~3-5% |
| Below USD 30M | ~500-800 | ~1-2% |
Asia SFO AUM distribution, illustrative estimate (not reported data).
The shape is what you’d expect — heavy long tail at the top end, broad distribution in the middle, modest representation at the smallest end where the substance economics get tighter. But the takeaway is non-obvious: the 4,000 count is dominated, in headcount terms, by structures in the USD 30-200M AUM range, while the total AUM is dominated by the much smaller number of USD 1B+ structures.
This has implications for service-provider strategy. A B2B family office service provider whose business model is built on serving the bulk of the 4,000 by headcount is targeting the mid-tier and the long tail — structures whose individual fee budgets are modest but whose volume is large. A B2B service provider whose business model is built on AUM-share — sophisticated multi-family services, complex structural advisory, alternative investment access — is targeting a much smaller pool of 200-400 structures whose individual fee budgets are large but whose volume is much smaller.
These are essentially different businesses, and the right strategy for each is materially different.
A service model sized to the whole 4,000-structure market fits none of it well. The few hundred largest structures need high-touch service, the mid-tier needs a scaled model, and one model priced for both serves neither.
Observation 3: The growth is selective, not broadening
The third important read is what the trajectory looks like going forward. The growth since 2020 — roughly four-fold on McKinsey’s own count, with each hub’s official count running higher since — is not broadening growth. It is concentrated growth, and the underlying drivers point toward continued concentration.
Three structural drivers of where the next wave of SFO formation is happening:
Driver 1: Asian-origin generational wealth transfer. McKinsey estimates USD 5.8 trillion of intergenerational wealth transfer among Asia-Pacific HNW and UHNW families between 2023 and 2030, with UHNW families accounting for about 60% of it. The arithmetic of this transfer means that families currently operating without formal SFO structures will increasingly create structures to manage the next-generation transition — preserving family governance, providing fiscal optimization, enabling cross-border family member coordination. This is the largest single source of new SFO formation through the back half of this decade. It is also concentrated: it primarily affects families in the USD 100M+ band where the structural overhead is justified.
Driver 2: External wealth migration to Asia. Continued migration of wealth from China, India, and increasingly from broader emerging markets into Asian wealth hubs (Singapore, Hong Kong, Dubai, increasingly Japan) drives a steady second source of new SFO formation. This source is harder to size precisely; my rough estimate, not a published figure, is 200-400 new SFOs per year across the region. It is concentrated in mid-tier wealth (USD 30-200M) where the migration economics work but the structural overhead remains tractable.
Driver 3: Existing-wealth restructuring. A meaningful but smaller source of headline SFO count growth is existing wealth that was previously held through less formal structures (private investment companies, family-owned holding entities, offshore vehicles) being restructured into formal SFO frameworks to access tax incentives, improve governance, or modernize operational infrastructure. This is largely a substitution effect — the same wealth being recategorized into the SFO count — and it overstates the underlying new-wealth growth.
Pull these three drivers together and the forward trajectory looks like: the two-hub count, already past 5,380, keeps rising through 2030 — but more slowly in distinct families than in structures. The AUM mass behind these structures grows somewhat faster, driven primarily by the USD 5.8 trillion generational transfer. The structures that grow fastest in headcount are at the mid-tier USD 30-200M band; the structures that grow fastest in AUM are at the USD 1B+ band.
What this means for practitioners
Three practitioner implications from the 4,000 milestone, ordered by their immediate operational relevance.
Implication 1: Service segmentation is no longer optional. The 4,000-and-growing population is too heterogeneous for any single service model to address economically. Practitioners — fund administrators, family office consultancies, multi-family office services, legal counsel — increasingly need to segment their target market explicitly: high-touch for the top tier, tech-enabled for the mid-tier, and either a structured value-tier offering or explicit non-coverage for the smaller end. The all-segments-one-service approach that worked at the 2020 scale does not scale to today’s 5,000-plus.
Implication 2: Talent supply is the binding constraint, not capital. The wealth that wants Asian SFO structuring continues to grow. The supply of senior practitioners — Investment Professionals, family office CIOs, fund admin partners with family office specialization — is growing materially slower. This gap shows up in compensation pressure (already visible across all major hubs), in service quality variance (more structures competing for finite practitioner attention), and in the structural friction described in the family office hiring squeeze.
Implication 3: The wealth-hub competition is no longer Singapore-vs-Hong-Kong. Up through 2023, industry coverage framed Asian SFO competition as a binary choice between Singapore and Hong Kong. By 2026, the competitive landscape has broadened — Dubai, Tokyo (for niche profiles), India’s GIFT City, and informal but growing Mumbai HNW activity are all credible alternatives for specific profiles. The 4,000 milestone itself is a two-hub count, and Singapore and Hong Kong remain the dominant destinations; the point is that the next leg of regional growth is no longer confined to them.
The summary I’d offer for the milestone: 4,000 SFOs across Hong Kong and Singapore is real, and the official counts are already past it. It represents roughly 20-37.5% fewer distinct families than the headline implies. It is dominated, by AUM, by a much smaller pool of very large structures. Its growth from here is going to be selective — driven by generational transfer at the top end and external migration at the mid-tier — rather than broadly distributed. Practitioners and service providers building strategy off the headline number without disaggregating these layers will end up with strategies that aim at the wrong segment of the population.