Where the Ruler and ARK Part Ways

We set ARK's three published funds against every listed company on the Blitz Index. ARK holds the better-scored half of our universe — no D grades, no BlitzFailing watch names, and 11 of the 19 companies at or above 65. The eight high scorers it does not hold are not random.

The comparison

Our company index lists 121 scores. Removing the three private companies (scored as ranges) and the historical WeWork file leaves 117 listed companies scored today. We split them by whether they appear in ARKK, ARKW or ARKF as of 11/04/2026. Alphabet counts as held: ARK owns its Class C shares (GOOG).

The held group sits about ten points above the rest on the mean and seven on the median.

Three things ARK's funds do not own

1. No D grades. The eleven listed D-grade companies — among them Talen Energy, Centrus, AeroVironment, EVgo, Dropbox, Rivian, Cipher Digital, Fermi, Liberty Energy, Eos and Voyager — appear in none of the three published funds. The lowest-scored ARK holding is 10x Genomics at 41, near the floor of C.

2. No BlitzFailing watch names. The three companies currently flagged for demerits of −15 or worse — Super Micro Computer, Carvana and Eos — are not held.

3. Eight of the nineteen 65+ companies.

The pattern is visible in the scorecard language: harvest mode, graduated, moat mature. Most of these are companies the Blitz Index scores highly because the blitz worked — they hold the category and earn the margins. ARK describes its mandate as disruptive innovation, and these companies are largely past the disruption stage. The two exceptions are AppLovin and Reddit, fast growers whose scorecards flag questions of trust and traffic source rather than maturity. For the rest, the ruler and the manager are answering different questions about the same companies.

Where ARK holds what the ruler grades lower

The other side of the divergence is the low-scored names ARK holds at size:

Each scorecard explains its score in terms of structure rather than ambition: reserve-interest revenue (Circle), demerits on governance and related-party revenue (Tempus), product revenue routed through a partner (CRISPR), leverage (CoreWeave), three flat years (10x). These are the kinds of bets a disruptive-innovation manager makes early, before the structure is proven — which is exactly when this ruler grades conservatively.

When the lens widened to six funds

The comparison above covers the three funds published on 5 November. The Lens has since published the other three — ARKQ, ARKX and ARKG — so the same split can be recomputed on holdings of 11/06/2026 against the current published score record. The headline gap narrows sharply, and one of the three findings does not survive.

The held group’s mean falls from 61.2 to 55.0, and the gap over the not-held group shrinks from about ten points to about two and a half. The median gap shrinks to a single point, 55 against 54. The three new funds add 32 scored names to the held group, and they are, on this ruler, ordinary-to-weak names: the specialist sleeves — autonomy and defense, space, genomics — are where the ruler’s conservatism about early-stage businesses bites hardest. The three-fund picture flattered ARK’s held group; the six-fund picture is closer to a tie.

Two of the three findings survive the widening. First, the eight high scorers ARK did not hold — Uber, Arista, Microsoft, AppLovin, Reddit, ServiceNow, Salesforce and MongoDB — are still held by none of the six funds. A ninth joins them: Nu Holdings (65), also held by none of the six funds. The harvest-mode pattern is unchanged: the ruler’s highest-scored non-holdings are still the companies whose blitz already worked. Second, the watch-list names — Super Micro Computer, Carvana, Eos and Voyager — still appear in none of the six funds on 11/06/2026.

The first finding does not survive. “No D grades” was true of the published record on 5 November; it is no longer true. Seven D-grade companies are now held across the six funds: AUR (34) in ARKQ, AeroVironment (35) in ARKQ and ARKX, HONA (37) in ARKX, KDK (36) in ARKQ, Intuitive Machines (38) in ARKQ and ARKX, Pacific Biosciences (29) in ARKG and ARKK, and Personalis (39) in ARKG. Six of the seven sit only in the three newer funds, where the specialist sleeves hold early-stage names the ruler grades in the 30s. The seventh, Pacific Biosciences, sits in ARKK itself — its score entered the published record after this article’s snapshot, which is why the original table did not show it. The finding should now be read with a date on it: no D grades in the three published funds as scored on 5 November.

Why the gap narrowed

The narrowing has a mechanical cause and a substantive one. Mechanically, the held group grew from 40 names to 72, and the new names score lower: the three new funds’ scored holdings average 48.8 (ARKG) to 56.4 (ARKX) on an equal-weighted basis, against 58.4 to 64.6 for the first three funds. A mean over 72 names moves more slowly than a mean over 40, but it also moves toward the new names.

Substantively, the three new funds hold the parts of ARK’s mandate where the Blitz Index is most skeptical by design. Genomics names are scored against a G1 gate that demands product revenue; space and defense names against a G2 gate that demands a ten-fold market; autonomy names against both. The ruler was built for companies with proven demand and large headroom, and the specialist sleeves are, by mandate, earlier than that. The six-fund comparison therefore measures something different from the three-fund one: not just whether ARK holds better-scored businesses, but whether the ruler’s definition of “better” fits every theme ARK runs. On the evidence here, it fits the generalist funds better than the specialist ones.

The selection-effect caveat from the original comparison gets stronger, not weaker, in the wider lens. Of the 154 listed companies with a published score, 72 — nearly half — are now held by at least one ARK fund, because coverage has kept following ARK’s holdings as the Lens expanded. The “not held” group is still not a random sample of the market; it is the remainder of a list built to test blitzscaling candidates, including names scored specifically to test the model’s failure side. The six-fund gap of two and a half points is measured on a universe that ARK itself helped select.

What still holds

Three things are unchanged by the wider lens. The harvest-mode pattern among the unheld 65+ names is intact — none of the eight has entered any ARK fund. The low-score, large-weight pattern is intact: Circle (52), Tempus AI (55), CRISPR (48), CoreWeave (53) and 10x Genomics (41) are still held at size in the first three funds, for the structural reasons their scorecards give. And the core limitation is intact: this is a descriptive comparison of two lists on one date, and our pre-registered tests still find no score-based signal that times returns. What the six-fund view adds is a boundary on the original claim — the held group’s advantage was a three-fund phenomenon, driven by the generalist funds, and it mostly disappears once the specialist sleeves are included.

Read this with the selection effect in mind

The comparison is not clean, and the biggest caveat is ours. BlitzLab chose which companies to score, and the ARK Lens itself pushed coverage toward ARK's holdings. A good part of the "held" group exists in our universe because ARK holds it. The universe is not a random sample of the market — it is a list of blitzscaling candidates, so the "not held" group includes many companies we scored specifically to test the model's failure side.

Two further limits:

• Only three of ARK's six funds are included. ARKQ, ARKX and ARKG holdings are not yet published on the Lens, so some "not held" companies — space, defense, robotics or genomics names among them — may be held elsewhere at ARK.
• One day's holdings. A position bought or sold tomorrow changes the split.

What the comparison does and does not show

It shows that, on 11/04/2026, ARK's three published funds lean toward the better-scored part of our universe and avoid its weakest tail entirely. That is consistent with our ARK backtest, where the model's most reproducible strength was identifying the D-grade failures. It does not show that ARK uses anything like our ruler, and it does not show that holding higher-scored businesses produces better returns — our pre-registered tests found no score-based signal that did.

The boundary

A descriptive comparison of two lists on one date. Scores measure business quality, not price; holdings reflect a manager's judgments, including on price, that the Blitz Index deliberately does not make.

BlitzLab scores businesses against a scholar's standard — not investment advice. No price targets, no buy/sell recommendations. Holdings come from ARK Invest's daily public disclosures; BlitzLab is not affiliated with ARK Invest and this research is not endorsed by them.

Sources

1. BlitzLab, "All Company Scores" — https://blitzlab.app/companies/ — checked 2026-11-05
2. BlitzLab, ARK Lens fund pages and overlap view — https://blitzlab.app/ark/overlap/ — holdings as of 11/04/2026 — checked 2026-11-05
3. BlitzLab, "About BlitzLab" (grades, BlitzFailing watch list) — https://blitzlab.app/about/ — checked 2026-11-05
4. BlitzLab, "Can Our Model Tell ARK's Winners from Its Blowups?" — https://blitzlab.app/research/ark-backtest — 2026-10-15 — checked 2026-11-05
5. BlitzLab, "What the Blitz Index Measures — and What It Doesn't" — https://blitzlab.app/research/synthesis — 2026-10-15 — checked 2026-11-05