Disruptive innovation ETFs share a label, not a portfolio

Two U.S. funds that both promise disruptive innovation share nine stocks and 15% of their weight. That mismatch, not the marketing line, is the 2026 story.

The ARK Innovation ETF CARKK, an actively managed cross-sector book, is led by Tesla, Space Exploration Technologies, Circle Internet Group, Tempus AI, and CRISPR Therapeutics. The Fidelity Disruptive Technology ETF BFDTX is led by Micron Technology, Marvell Technology, and Taiwan Semiconductor. One portfolio is a high-conviction mix of electric vehicles, private space, stablecoins, gene editing, and crypto plumbing. The other is a semiconductor and software stack. They are sold as the same idea.

etf.net groups 30 funds under Disruptive Innovation. Returns inside that file, and in the adjacent robotics, quantum, and chip sleeves readers often treat as substitutes, have split with the holdings. As of Friday, September 11, ARK's genomic revolution fund is up 58% year to date. Its blockchain and fintech sibling is down 5.5%. Fidelity's chip-heavy version has returned 34%. The Nasdaq-100 ETF AQQQ is up 17%. The theme did not have a year. Specific books did.

ARK's family is not one factor

CARKK remains the brand. It launched in October 2014, charges 0.75%, and holds $5.56 billion. Forty-seven names sit in the portfolio. The top ten are 49% of assets. Tesla alone is 9.7%; Tesla and SpaceX together are 16%. Healthcare is the largest sector at 29%, ahead of technology at 27% and financials at 17%. That is not an AI-infrastructure fund wearing an innovation costume. It is a manager's list of companies it ties to autonomous mobility, multiomics, digital wallets, and next-generation software, including OpenAI Group PBC Series C, a private line, at 2.6%.

The rest of the ARK complex makes the construction point sharper, because the funds share an issuer, a fee, and a research culture, and still do not share a return.

The ARK Genomic Revolution ETF DARKG ($1.56 billion, 0.75%) must keep at least 80% of assets in companies tied to genomics. It is 97% healthcare. Its two largest holdings, 10x Genomics and Twist Bioscience, have each returned more than 300% this year. 10x is 10% of the fund; Twist is 8.1%. On August 6, 10x reported fiscal-quarter revenue of $151 million, above the $147 million estimate, and beat on earnings. On August 3, Twist reported revenue of $118.4 million, above the $114.5 million estimate, while still missing on per-share earnings. Those prints sit on top of a move that was already in place: Twist is up 197% over six months, 10x 260%. Put those weights on those stocks and a 58% fund return is not a mystery. It is concentration doing what concentration does.

The ARK Blockchain & Fintech Innovation ETF DARKF ($701 million, 0.75%) has to keep at least 80% in financial-technology names. Block is 7.9%. Circle is 6.9%. An in-house bitcoin ETF holdco is 6.8%. Then come Shopify, Robinhood, and Coinbase. Coinbase is down 21% year to date and 43% over 12 months. Circle is up 20% this year and down 17% over 12 months. Tesla, the other large public name that still dominates the flagship, is down 18% year to date. The fintech book is not a failed innovation thesis. It is a different set of businesses, and this year those businesses did not pay.

The spread between DARKG and DARKF is 64 percentage points year to date. Over 12 months into September, the genomics fund took in $163 million of net creations; the fintech fund saw $351 million of net redemptions. That is the cost of treating "ARK" or "disruptive" as a single allocation.

The flagship sits in the middle, which is what a blended book does. CARKK is up 10% year to date and 12% over 12 months. Over three years it has almost doubled, in line with AQQQ. Over five years it is still down 31%, against a 94% gain for the Nasdaq-100 fund, and its annualized volatility is 46%, about twice AQQQ's. The 2022 drawdown, at a peak-to-trough of 76%, has not been earned back. A three-year rebound that matches large-cap growth is not the same thing as a five-year compounding engine.

The chip book hiding in "disruptive technology"

Fidelity's version of the same phrase owns a different economy. BFDTX is $288 million, so it is not a substitute for CARKK on size, liquidity, or history. It launched in April 2020, charges 0.50%, and holds about 50 names. Technology is 86% of the portfolio. Micron is 8.5%, Marvell 7.9%, TSMC 6.0%. Palantir and NVIDIA follow. The top ten are 52% of assets.

Micron has returned 242% year to date and 598% over 12 months. It is a $1.10 trillion company. AMD has returned 139% year to date and 221% over 12 months. A fund whose largest bets sit in memory, custom silicon, and foundries did not need a new definition of disruption to have a 34% year. It needed those stocks.

The VanEck Semiconductor ETF ASMH, a dedicated chip fund rather than an innovation fund, is up 58% year to date and 90% over 12 months. That is the same neighborhood as DARKG, from a completely different industry. The economic engine in public markets this year has been physical AI infrastructure (memory, foundries, equipment) and a handful of biology-tools companies. Software platforms and crypto rails have not been that engine.

Here is the overlap that makes the Fidelity fund a different decision than it looks on a fact sheet. BFDTX shares 20 names and 33% of its weight with AQQQ. Micron is already 4.9% of the Nasdaq-100 fund, NVIDIA 8.5%, Microsoft 5.9%. Adding Fidelity's disruptive-technology book on top of a Nasdaq-100 core is, in large part, a decision to own more Micron, more Palantir, more Palo Alto Networks, and more TSMC than AQQQ already delivers. It is not a way to buy the names CARKK owns that AQQQ does not.

CARKK and BFDTX, the two products whose names sound interchangeable, overlap by 15% of weight. Palantir, NVIDIA, Amazon, Meta, and Alphabet show up in both. So does SpaceX: 6.37% of CARKK, its second-largest holding, and 0.3% of Fidelity, a 21-times gap between a core position and a rounding error. Tesla, Circle, Tempus, CRISPR, Coinbase, and Robinhood do not appear in the Fidelity book. That is the whole comparison.

AKOMP's 500 stocks and AQTUM's $5.57 billion

Index construction produces a third book that the innovation label also fails to describe.

The State Street SPDR S&P Kensho New Economies Composite ETF AKOMP ($2.63 billion) tracks a 500-stock index that uses artificial intelligence and a systematic, data-driven methodology to identify companies Kensho tags to a "new economic era." It charges 0.20%, the lowest fee among the major products in this group. The top ten names are 10% of assets. Twist Bioscience is the largest holding at 1.3%. NVIDIA is 0.81%. Tesla is 0.72%. AKOMP and CARKK share 24 stocks, but only 7% of weight: Tesla is more than 13 times as large in CARKK. AKOMP is up 12% year to date and 15% over 12 months, close to the S&P 500 ETF ASPY at 13% and 19%. A 500-stock new-economy composite behaves like a diversified mid-and-small-tilt growth fund, because that is what it is.

The surprise of scale sits outside the ARK-Fidelity-State Street triangle. The Defiance Quantum ETF AQTUM now holds $5.57 billion, matching CARKK almost dollar for dollar. It tracks an equal-weighted machine-learning and quantum-computing index, charges 0.40%, and holds 89 names. The largest, Arqit Quantum, is 1.9%. The top ten are 17%. It is up 35% year to date and 53% over 12 months. The fund seeks companies tied to quantum computers, machine-learning hardware and software, chip-packaging machinery, and raw materials for quantum computing. In practice the book is a diversified basket of chip designers, cloud software, and a long tail of smaller computing names. Over 12 months through September 9 it took in $2.03 billion of net creations. Tesla and Coinbase are not among its largest holdings.

Three other books in the same aisle split along the same construction lines: the iShares Future Exponential Technologies ETF AXT, a global mega-cap growth portfolio with extra healthcare; the ALPS Disruptive Technologies ETF BDTEC, an equal-weight of about 100 names across ten themes; and the robotics pair of ARK's autonomous-technology fund CARKQ and Global X's robotics and AI fund CBOTZ. Their holdings and returns sit in the table.

The comparison below uses the same Friday session for every return. Size and fee sit next to the numbers because a $288 million active fund and a $2.6 billion 500-stock index are not interchangeable just because both say innovation.

The year-to-date column is a holdings report. It is not a verdict on which manager to trust.

Defiance's foundry ETF, and the stack you already own

On September 10, Defiance launched the Defiance Global Foundries ETF —AIFR, a 0.71% fund dedicated to semiconductor foundries, tracking the MarketVector Global Foundries Index. On September 9 the same issuer listed the Defiance AI Magnificent 10 ETF AIMG, a 10-stock concentrated AI fund. Defiance already runs AQTUM. Two products in two days, from the house that built a $5.57 billion quantum-and-machine-learning book without a Tesla-sized bet. The pattern is the same one the returns have been teaching all year: you are no longer being asked to buy "innovation." You are being asked to pick a link in a chain (memory, foundry, equipment, genomics tools, crypto rails) and to accept the cycle that comes with it.

That cycle is not theoretical. Semiconductor funds can drop several percent in a session without the long-run demand story changing; they did so in late June, and they did so again in early September, when ASMH fell 2.4% on Wednesday after a strong run. Circle's own 12-month range runs from $50 to $159. Equal-weighting, as AQTUM and BDTEC do, reduces the chance that one name defines the year. It does not remove the chance that the whole sleeve is the same trade.

The Global X Artificial Intelligence & Technology ETF BAIQ is larger than any of the disruption-labeled funds, at $10.1 billion. Palantir, Oracle, Microsoft, SpaceX, and Tesla lead. It is up 26% year to date. Combined with AQQQ's 8.5% NVIDIA weight and 4.9% Micron weight, that is the reminder the category is built to ignore: a holder of a Nasdaq-100 or broad AI value-chain fund already owns the listed companies that turned AI capex into 2026 stock returns.

What those core funds do not own in size is the rest of CARKK. They do not own a 10% sleeve of 10x Genomics. They do not own Circle or OpenAI Group PBC Series C. They do not own Twist at 8%. They also do not own a 500-stock Kensho mash of 3D-printing, lidar, and defense-electronics names at 20 basis points. Those are the incremental decisions. The innovation label collapses them into one.

CARKK still carries a 76% peak-to-trough drawdown that has not been earned back. What 2026 added is a shelf of the pieces the bundle mixed together: genomics tools, foundries, an equal-weight quantum book that already matches the flagship in assets. The label is still one product. It is no longer the only way to own the parts.