In the last post we argued the data layer is the moat. This post is about why nobody has built that data layer end-to-end yet — even though the puzzle pieces have existed for years.
The short version: the publics-aggregation world and the alts-aggregation world grew up as two different industries, with two different funding stories, two different customer bases, and two different sets of technical problems. They have not converged. The gap between them is where most of HNW reality actually lives.
Let us explicitly tell you a wild story. A CEO of one of these players said to us, word for word: “You need to be super disciplined about doing only one little thing, and API everything else.” This really is the market view.
Bucket one — the public aggregators
These are the firms that read your brokerage and bank accounts. Mature category, mostly stable APIs, screen-scraping wars with the big custodians mostly settled. By their own counts, the leading players connect to ten-thousand-plus institutions and power a meaningful share of U.S. consumer fintech.
These aggregators are very, very good at some things: pulling positions, balances, and transactions from custodial accounts. Public equities. Public bonds. Mutual funds. ETFs. 401(k)s. 529s. Crypto exchanges. Very clean stuff.
They are not built for the messy stuff. Ask one of these for the unrealized capital account in a Delaware LP holding a stake in a sponsor-led continuation vehicle, and it will, politely, ignore you.
Bucket two — the alts aggregators
A growing list of AI-document-extraction shops, most of them founded post-2018, most of them riding the wave of family-office allocation into alternatives. The category processes documents at industrial scale — public disclosures cite more than 150,000 documents handled for just one platform partner’s clients. The category exists because, as Family Wealth Report has noted, family offices were literally hiring half-time and full-time employees just to download statements from GP portals before these tools existed.
These firms are very, very good at some things: extracting structured data from unstructured alts documents — capital calls, distribution notices, quarterly statements, K-1s — and pushing it into accounting systems.
They are not built to read your Schwab account. Ask one of these to pull your daily public-equity positions, and it will, politely, ignore you.
The gap in the middle
Here is the thing nobody on either side likes to say out loud.
A UHNW family’s reality is not “publics” or “alts.” It is publics held at three brokerages, alts held across forty private vehicles, two operating businesses that throw K-1s, four real-estate LLCs with mortgages, a 412(e)(3) plan, several intercompany loans, art, two yachts, a plane held in a partnership for tax reasons, and a charitable remainder trust that owns part of the public-equity portfolio.
Roughly 90% of that family’s actual balance sheet does not fit cleanly into just one product. Mortgages — handle them as liabilities but don’t tie them to the underlying real estate entity. Real-estate LLCs with GP economics — not built for them. Operating-company K-1s — neither. Carry interests — neither. Oh, I had a great idea with staking stablecoin — neither. Direct private investments not in a fund wrapper — neither. Crypto held in self-custody alongside crypto on Coinbase — partial coverage, often wrong. Vehicles registered to LLCs for liability reasons — neither.
This is what we mean when we say the gap is where reality lives.
Not to mention there are all sorts of bizarre assets out there that neither the public guys nor the private guys really focus on. You end up needing your analyst to work on your weird, hyper-juiced waterfall you got on that SPV you invested in last year because they couldn’t raise it and you got in last-money so you could dictate some wild terms. Good luck automating that with an API.
And to be fair — none of these vendors claim to. They aren’t doing that. It isn’t their business. They are so good at their business — why do that business? That’s hard work, messy work. Let’s not do that work, obviously, they say.
Why the two worlds haven’t merged
Three reasons.
One — different technical problems. Pulling a clean position file from a custodian is an API problem. Extracting a capital balance from an unstandardized PDF is an ML-on-documents problem. Different stacks, different talent, different ROI curves.
Two — different customers. The publics aggregators sell into B2C fintech and mass-affluent advisory. Their unit economics work at $5–$20 per account per year. The alts aggregators sell into family offices and institutional allocators. Their unit economics work at six-figure contracts per relationship. You don’t naturally merge those motions.
Three — different exit logic. The publics aggregators got bought into payment and reporting platforms. The alts aggregators are being absorbed into family-office service providers and large platforms. Different acquirers, different roadmaps.
What the convergence looks like
We think the convergence is now inevitable, and we think it happens at the family-office and UHNW end of the market first, for two reasons.
First, alternatives allocation among HNW and UHNW households keeps climbing. Bain projects private-market AUM at $60–$65T globally by 2032. Capgemini reports that 93% of young HNW investors plan to increase their alts exposure further. The mix is shifting faster than the data infrastructure is.
Second, AI changes the unit economics of handling the weird stuff. A document-extraction problem that used to require a full-time analyst at $90K can now be handled by an AI agent supervised by a human at a fraction of the cost — but only if you’ve built the supervision model thoughtfully, which we’ll get into in post six.
What buyers should ask
If you’re evaluating data infrastructure for a family office, RIA, or private-bank group in 2026, the test we’d run is:
(1) Bring your weirdest five assets. The carry stub, the K-1 from the operating co, the mortgage on the LLC-held real estate, the direct private investment, the foreign-domiciled trust position.
(2) Ask the vendor to ingest them and show you a clean, structured representation tied back to the right entity and ownership path.
(3) Then ask them to do the same for your top three public-brokerage accounts.
The vendor that can do both, end-to-end, with a clean ownership graph and reconcilable balances, is the data layer. The ones that can only do one side are tools you’ll use to build that layer. Both are useful. Only one is the moat.
The reframe
The market has built two halves of the structured database — one for publics, one for privates. Neither half is enough. The 90% of HNW reality that lives in the gap is where the durable infrastructure gets built. That gap is closing. It is not yet closed.