Skip to main content

The Robo-Advisors That Don't Call Themselves Robo-Advisors

The SEC has a checkbox for internet advisers. Acorns, Stash, Betterment, Wealthfront, SoFi and Robinhood all leave it blank. Reading Form ADV structurally instead of taking the declaration at face value finds 40 firms managing money for 13.2 million Americans — with 454 people doing the advising.

Fidelon Research · Note 04September 3, 2026Revision 1.1

Basis: every figure on this page reflects filings the SEC had on record as of May 12, 2025, and has not been restated since publication. Form ADV client and asset counts are annual and firms revise them freely, so current filings will differ. A later review of filings dated September 7, 2026 reproduces this same group of firms — 40 of them — and the headline finding at 87.4%.

210SEC-registered firms check the internet-adviser box0.96% of all filers
2.0MClients served by all 210 of them, combinedItem 5.F(2)(f)
4.8MClients at Acorns alone, which checks "No"One undeclared firm
86%Of app-managed clients sit at firms the flag misses11.3M of 13.2M

The problem

A self-declaration nobody declares

A rule lets an adviser register with the SEC on the grounds that it gives advice through an interactive website. Form ADV Item 2.A(11) is where a firm says so. It is the regulator’s own robo-adviser flag, and it is free for anyone to check.

It is also nearly empty. Of 21,893 SEC-registered investment advisers on file, 210 check it — just under 1%. The median declaring firm reports 55 clients. Ninety-two of them report fewer than a thousand.

The reason is mundane: the internet-adviser box is a registration path, not a business description. A firm with $100 million in assets or a broker-dealer affiliate already qualifies for SEC registration on other grounds, so it never checks that box — no matter how thoroughly its product is an app. The checkbox tracks how a firm got in the door, not how it delivers advice.

Every firm that declares itself an internet adviser serves 2.0 million clients between them. Acorns, which does not declare, serves 4.8 million by itself.

Any product that classifies advisors by trusting that field inherits the gap. Consumers searching the names they actually know — Betterment, Wealthfront, Stash — land on firms the label has quietly filed as ordinary human advisory practices.

Method

Three structural tests

Rather than ask a firm what it is, ask what its own filing implies it must be. Three measures, all already in what firms report to the SEC, all arithmetic — no model, no guessing, no starting list of names.

  1. 01
    clients ≥ 10,000

    A firm with a retail app has retail scale. This is the floor for the question being asked; below it, the ratios get noisy.

  2. 02
    clients per advisory employee ≥ 1,000

    Item 5.B(1) counts only staff performing advisory functions — not engineers, marketers or support. This is the load-bearing test: no human practice reaches 1,000 clients per advising employee. The industry median is 57.

  3. 03
    AUM per client < $100,000

    Separates consumer apps from institutional sub-advisors and TAMPs, which also run enormous client counts on tiny advisory headcounts — but whose accounts are large. Without this test, Envestnet PMC and Schwab Investment Management land in the consumer bucket.

Why staff headcount, not total headcount

Total employee count and advisory staff count look interchangeable, and are not. Wealthfront Advisers reports far more staff overall than the small team actually doing the advising. Using total headcount, the signal is diluted by everyone who builds the app; using advisory staff only, it resolves.

Stash reports 124 employees in total. It reports 5 people doing the actual advising.

Evidence

The line runs through a quiet region

Plotted together, the three groups occupy separate regions. On this filing, the highest assets-per-client among the 40 app-native firms is $97,205 and the lowest among the excluded institutional managers is $122,351: no firm falls between, a gap of about 26%.

That gap is real but narrow, and it is a property of this one filing date rather than a fixed law. On filings dated September 7, 2026, the same boundary runs from $94,764 to $100,888 — still empty, but only 6.5% wide, with the nearest excluded firm clearing the line by $888. The $100,000 line is best read as a chosen cut that happens to land in a thin part of the distribution, not as a natural line between two kinds of firm.

Figure 1 — Advisory leverage against account size

Both axes are logarithmic. Each mark is one SEC-registered firm reporting client count, advisory headcount and regulatory assets. The shaded band marks the $100,000 assets-per-client line; the dashed line marks 1,000 clients per advisory employee.

App-native (40) Institutional & TAMP (49) Human RIAs, 10k+ clients (308)
$100$10k$1M$100M101001k10k100kClients per advisory employee (log)Assets per client (log)

The group

Forty firms, 13.2 million clients, 454 advisors

The full result. Declared marks whether the firm checks the SEC’s internet-adviser box — 25 of the 40 do not, and those 25 hold 86% of the group’s clients.

The ten largest of the 40 firms by client count
FirmClientsAdvisory staffClients / advisorAssets / clientDeclared
AcornsCRD 1659264,813,13717283,126$2,157No
Stash InvestCRD 2265502,282,2725456,454$1,825No
BettermentCRD 1491171,200,6733336,384$46,948No
Guideline Investments, LLCCRD 281515792,6533264,218$20,810Yes
MoneyLionCRD 289521679,2531679,253$17Yes
Human Interest Advisors LLCCRD 269875582,6761582,676$11,344No
Wealthfront Advisers LLCCRD 148456536,3402026,817$66,965No
SoFi Wealth LLCCRD 167958279,2161027,922$5,283No
NomadCRD 307764237,3622210,789$1,890No
Albert Investments, LLCCRD 298006230,377376,792$91No

Sorted by client count. Source: Form ADV Part 1A, as filed May 12, 2025. Fields: Items 1.F(5), 2.A(11), 5.A, 5.B(1), 5.F(2)(c), 5.F(2)(f), 5.I.

Two corrections this produced

Findings that change how the filings should be read

Wrap-fee assets are three fields, not one

A firm’s wrap-fee assets split into three separate buckets on Form ADV: assets where the firm is the sponsor, assets where it is the portfolio manager, and assets where it is both. Reading the sponsor field alone returns $0 for Betterment, Wealthfront, Acorns, SoFi, Titan, all of which are flagged wrap sponsors. Summing all three fields instead returns $56,400,000,000, $35,300,000,000, $10,400,000,000, $1,500,000,000, and $1,100,000,000, in that order. Across the group, 108.7 billion dollars is invisible to the single-field read.

Fifteen of these firms report one advisor or fewer

Between them they serve 1.8 million clients MoneyLion 679,253 against 1; Human Interest 582,676 against 1; Webull 136,277 against 1. Any rule that treats a near-empty advisory staff count as evidence a firm is dormant will suppress exactly these.

State-registered firms can’t be answered this way

The internet-adviser box is absent from state filings entirely — state-registered advisers aren’t SEC registrants, so the exemption it tracks doesn’t apply to them. The fields this method relies on are present. State-registered app advisers have to be classified this way, structurally, or not at all.

Implication

Who is winning the advice war

Set this group against every other firm that reports retail-comparable clients, and the answer splits in two directions at once.

Figure 2 — Relationships versus dollars, app-native against human RIAs

Human comparison set: 8,930 firms reporting clients, advisory staff and assets, excluding accounts averaging over $5M (family offices and institutional mandates). App-native: the 40 firms above.

Client relationships

18.00% · 13.2M clients59.9M clients

Assets under management

0.63% · $182B$28.5T

Apps have taken 18% of retail advisory relationships and 0.63% of the assets. The average app account holds $13,800; the average human-advised retail account holds $475,21034× larger.

Read one way, the incumbents are untouched: the money has not moved. Read the other way, the apps have already won the relationship with a generation whose balances are small because they are young, and 454 people are servicing 13.2 million of them at 172× the client load of a human practice.

The apps did not take the assets. They took the customers who do not have assets yet.

Which of those readings matters depends on whether those balances grow in place — a question this data cannot answer, but successive annual filings can. This report is one frame of it.

See the 40 firms, scored.

Every investing app and robo-advisor registered with the SEC, scored on what its own filings disclose.

Browse investing apps and robo-advisors →

Limits

What this does not establish

  • One filing date. Every figure here reflects filings as of May 12, 2025. Client and asset figures are annual and firms restate them freely. No trend is claimed from one snapshot.
  • Self-reported throughout. Every field is the firm's own answer on Form ADV. The method replaces a self-declared category with self-declared quantities, which are harder to shade but not audited.
  • Thresholds are judgment. 1,000 clients per advisor and $100,000 per account are chosen to sit inside thin regions of the distribution, not derived. Both are defensible; neither is unique. The account threshold is the sensitive one: the empty band around it is 26% wide on this filing date and 6.5% on filings dated September 7, 2026, so firms near the line move in and out between filings. This group's largest members sit orders of magnitude clear of it.
  • IA side only. Form ADV covers investment advisers. Broker-dealer-only platforms — Robinhood Financial, Webull's brokerage entity — file with FINRA and are out of scope. Where a brand runs both, only the advisory entity appears.
  • Boundary firms are genuinely mixed. Transamerica Retirement Advisors, Gradient and Signal Advisors sit near the AUM-per-client edge; Guided Choice at $88,187 is inside by $12k. These are recordkeeping and retirement-plan businesses that share the arithmetic of an app without being one.

Revisions

1.0 — September 3, 2026. Initial publication.

1.1 — September 9, 2026. Revised September 9, 2026: the filing date this note is based on is now stated up front, and the account-size boundary is reported as a measured 6.5% gap in a later filing rather than as empty space. No headline figure changed. A later review of filings dated September 7, 2026 reproduces this same group of firms — 40 of them — and the 86% finding lands at 87.4%.

Source: SEC Form ADV Part 1A, as filed, dated May 12, 2025. 21,893 firms reviewed. Fields: Item 1.F(5), 2.A(11), 5.A, 5.B(1), 5.F(2)(c), 5.F(2)(f), 5.I(1), 5.I(2)(a-c), 5.J(1).

Fidelon scores what firms actually disclose, not what they should disclose. Our methodology is public. Our scores are not for sale, and no firm can pay to change one.