Held Everywhere: The Thirteen Names in All Three ARK Funds
Thirteen listed companies appear in ARKK, ARKW and ARKF at the same time. Their average score is the highest of any group in the overlap view. Their median is not — and that difference is the finding. (Updated: Cloudflare’s score is 65, not 62, on the current record; the core mean is 66.9 and five of nine scored core names sit at or above the 65 line. The median is unchanged at 62.)
The question
A manager running several thematic funds reveals something when the same company shows up in all of them: the name fits more than one thesis. The overlap view on the ARK Lens deduplicates every holding across the three published funds — 69 unique listed tickers, 40 of them scored — and sorts by how many funds hold each one.
We asked a simple descriptive question: do the names ARK holds most broadly score differently on the Blitz Index from the names it holds in only one fund?
The thirteen
| Company | ARKK | ARKW | ARKF | Blitz Index |
|---|---|---|---|---|
| NVIDIA (NVDA) | 3.03% | 4.30% | 3.10% | 87 A |
| Meta (META) | 2.21% | 5.54% | 4.84% | 78 B |
| Shopify (SHOP) | 2.98% | 4.11% | 8.20% | 71 B |
| Palantir (PLTR) | 2.61% | 2.59% | 4.43% | 65 B |
| AMD (AMD) | 1.65% | 2.69% | 1.70% | 62 B |
| Cloudflare (NET) | 1.05% | 1.92% | 1.12% | 65 B |
| Coinbase (COIN) | 3.92% | 3.38% | 5.14% | 61 B |
| Robinhood (HOOD) | 2.88% | 3.73% | 4.95% | 61 B |
| Circle (CRCL) | 4.75% | 4.57% | 6.12% | 52 C |
| Block (XYZ) | 2.43% | 3.55% | 8.64% | Gate failed — G2 |
| Amazon (AMZN) | 2.40% | 4.64% | 3.93% | Gate failed — G2 |
| Bitmine Immersion (BMNR) | 1.70% | 1.80% | 2.02% | Gate failed — G2 |
| Bullish (BLSH) | 1.40% | 1.02% | 1.97% | Gate failed — G2 |
Two private positions — OpenAI Group PBC Series C and Kalshi — are also held in all three funds but carry no ticker, so they do not appear in the overlap view. OpenAI is 3.75–4.41% of each fund.
By breadth
| Held in | Listed names | Scored | Mean score | Median score | Share of scored at 65+ | Gate failed |
|---|---|---|---|---|---|---|
| All three funds | 13 | 9 | 66.9 | 62 | 56% (5 of 9) | 4 |
| Two funds | 17 | 11 | 65.0 | 62 | 36% (4 of 11) | 2 |
| One fund | 39 | 20 | 56.8 | 57 | 15% (3 of 20) | 6 |
Read the mean column alone and the story writes itself: the broader the holding, the higher the score. Read the median column and the story narrows. The three-fund core and the two-fund names share the same median, 62. The core's higher mean comes from its top two — NVIDIA (87) and Meta (78) — which is exactly the kind of tail our research program has learned to distrust in small samples. Our standing rule from the hypothesis tests applies here too: means lie in winner-take-all distributions; medians tell the truth.
What the medians do show is a step down to the single-fund names: 57, five points lower. That step is not a property of breadth itself. It is mostly one fund's specialist sleeve. Of the 20 scored single-fund names, nine are ARKK's genomics, diagnostics and therapeutics holdings, with scores between 40 and 61.
Three observations
1. The core is where the gates bite. Four of the thirteen core names are gate failures, all at G2 — Block, Amazon, Bitmine and Bullish. No other breadth group has a higher share of gate failures. The names ARK holds most broadly include large incumbents and crypto-market infrastructure for which a ten-fold market ceiling is out of reach — a statement about blitzscaling headroom, not about quality.
2. The core's lowest score carries the most weight. Circle (52) is the lowest-scored core name and sits at 4.6–6.1% of each fund — summed across the three funds, more than any other core name. Its scorecard explains the gap between conviction and score: USDC circulation is blitzing (+67% in 18 months), but more than 95% of reported revenue is interest on reserves, so the score reads a rate-dependent business, not a usage-dependent one.
3. The core is a cross-section of themes, not a theme. Semiconductors (NVIDIA, AMD), consumer platforms (Meta, Shopify), enterprise software (Palantir, Cloudflare), and crypto rails (Coinbase, Robinhood, Circle, Bitmine, Bullish) all sit in all three funds. Shared holdings look less like a single thesis and more like the names that fit several.
The core, one week later
The thirteen survived the week intact. On holdings of 11/06/2026, all thirteen names still appear in ARKK, ARKW and ARKF, at weights close to the 11/04 snapshot. But their combined weight is worth stating, because the overlap table shows each name separately and never adds them up. Together, the thirteen account for 35.5% of ARKK, 44.8% of ARKW and 55.5% of ARKF — more than a third of the first fund, nearly half of the second, and over half of the third. The scored nine of the thirteen account for 26.6%, 32.9% and 38.0% respectively; the rest is the four gate failures, worth 8.8%, 11.9% and 17.6% of each fund.
That concentration reframes the first observation above. The core is not only where the gates bite; it is where a large share of each fund’s capital sits outside the ruler entirely. In ARKF, more than a sixth of the fund is in four names — Block, Amazon, Bitmine and Bullish — that carry no score at all, and they sit alongside the core’s lowest-scored name, Circle (52), at 4.6–6.1% per fund. A reader who knows only the core’s mean of 66.9 knows the average of the nine names the ruler can grade; the weights say that nearly a third of the core’s capital in ARKF is in names it cannot grade.
Four names go wider
The thirteen were defined by breadth across three funds. With six funds published, breadth has a longer scale — and four of the thirteen stretch further. NVIDIA, Palantir, AMD and Amazon now appear in five funds, having been added to ARKQ and ARKX: NVIDIA at 5.73% in ARKQ and 5.02% in ARKX, Palantir at 4.84% and 3.71%, AMD at 2.38% and 2.31%, Amazon at 4.09% and 5.45%. The other nine core names appear in no additional fund.
No ticker reaches all six funds. The blocker is ARKG, which holds none of the thirteen — its mandate does not overlap with the semiconductor, platform, software and crypto names that make up the core. The five-fund core is therefore a narrower, more specific claim than the three-fund one: it is the set of names that fit the innovation, next-generation-technology, fintech, autonomy and space theses at once, but not genomics. Its scored members are NVIDIA (87), Palantir (65) and AMD (62), with a median of 65; Amazon, the fourth, is a G2 gate failure in all five funds that hold it. The pattern from the three-fund analysis repeats at the wider scale: the broader the holding, the more the group is pulled up by its top name — here NVIDIA, 22 points above the next scored member.
What breadth still doesn’t mean
The Cloudflare correction above is a small demonstration of the article’s own warning. A single rescore — Cloudflare from 62 to 65 — moved the core mean by 0.3 points and the 65+ share from four of nine to five of nine, while the median did not move at all. In a nine-name group, the mean is a hostage to one company’s quarter; the median is the stabler description, and it still says 62. The same caution applies to the new five-fund core: three scored names, one of them 22 points clear of the rest, is not a distribution but a trio with an outlier.
Breadth also remains a count, not an exposure, and the new funds sharpen the reason. NVIDIA at 5.73% of ARKQ and 3.03% of ARKK is the same breadth — one fund each — at very different weights, and the funds themselves are different sizes, so no cross-fund sum is meaningful. What breadth does show, at three funds and at five, is fit: the names that clear more than one thesis. NVIDIA, Palantir and AMD clear five; Meta, Shopify, Cloudflare, Coinbase, Robinhood, Circle, Block, Bitmine and Bullish clear three. Whether that fit predicts anything is still untested — no hypothesis was registered before looking, and the sample is still nine scored names at the core.
What this is not
Summing a company's weights across three funds does not produce a portfolio weight; the funds have different sizes and different investors. Breadth here is a count, not an exposure. And the sample is small — nine scored names in the core, eleven in the two-fund group — so a difference of a few points in the mean is well inside what a single company can move.
The boundary
This is a snapshot of one day's holdings, described against one ruler. It does not test whether breadth predicts anything, and we registered no hypothesis before looking. Scores measure business quality; they do not measure price, and our pre-registered tests found no score-based signal that timed returns.
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
- BlitzLab, "ARK Funds Overlap: Every Holding, Every Fund." — https://blitzlab.app/ark/overlap/ — holdings as of 11/04/2026 — checked 2026-11-05
- BlitzLab, ARK Lens fund pages (private positions not shown in overlap) — https://blitzlab.app/ark/arkk/ · https://blitzlab.app/ark/arkw/ · https://blitzlab.app/ark/arkf/ — checked 2026-11-05
- BlitzLab, "Circle — Blitz Index 52 (C)" — https://blitzlab.app/companies/crcl/ — Oct 2026 scorecard — checked 2026-11-05
- BlitzLab, "What the Blitz Index Measures — and What It Doesn't" (medians vs means) — https://blitzlab.app/research/synthesis — 2026-10-15 — checked 2026-11-05