CHAT vs. FTEC: The "Broad" Fund Is the Real Concentration Play

CHAT vs. FTEC: The "Broad" Fund Is the Real Concentration Play

Same bell, same clock, one question: which wrapper buys the next twelve months of technology better — CHAT, sold as concentrated generative-AI exposure, or FTEC, sold as broad tech diversification? Both convert to 100 paper points at Friday's close and run to the same Friday a year from now. Total return decides, distributions are reinvested, no substitutions are allowed. FTEC opens a 53/47 favorite.

Here is the twist that makes the matchup worth staging in the first place: the marketing has the two labels backwards. The "concentrated AI" fund holds roughly four dozen names and spreads its top ten across just over a third of the portfolio. The "diversified" fund is built so that three mega-caps weigh as much as all ten of those names combined. Before the race, one false assumption has to die: diversification is not the same thing as diversification headline.

The Frozen Card

• The contestants. CHAT, the Roundhill fund that was billed at launch as the world's first Generative AI ETF, in the red corner. FTEC, Fidelity's mechanical tracker of the tech sector, in the blue.
• The question. Do you capture the next leg of the AI trade better through a human-curated screen that owns the builders and pays ten times the fee, or through a cap-weighted basket that charges almost nothing but lets Apple sit second in the fund?
• The starting line. Friday's official close: CHAT at $87.80, FTEC at $282.35. Both normalize to 100.00 paper points in U.S. dollars.
• The finish. The last Friday in August 2027, twelve months from the bell.
• The score. Total return measured on net asset value, with quarterly distributions reinvested. Both funds pay out; CHAT's trailing distribution yield near 1.91% versus roughly 0.35% for FTEC means that yield kick is inside the score, not on the side. Neither side gets a fee haircut beyond what it already pays — at 0.75% versus 0.08%, the cost gap is embedded in observed performance and compounds for whoever carries the cheaper wrapper.
• The cadence. Marks at the end of each quarter, plus an immediate write-up after any material corporate action — a liquidation, an objective change, an index reconstruction. No mid-quarter re-marking, no peak picking.
• The exits. If either fund liquidates, the protocol freezes its value at the last official NAV, that cash sits in the account earning the prevailing rate, and the match continues on the surviving leg. No substitutions, ever.
• The source. Prices, returns, yields, drawdowns and assets are live market-data snapshots as of the last completed session; fund weightings come from sponsor and index-publisher documents.

The Mechanism Board

The score decides the winner; this board explains the score. Both funds answer the same economic question — how to own mega-cap technology with as little friction as possible — so the fair-opponent test passes, but they answer it with opposite construction logic.

Variable	CHAT (AI tilt)	FTEC (broad tech)
Expense ratio	0.75%	0.08%
Portfolio	~49 names; top 10 = 38.25% of assets	Cap-weighted; top three alone ≈ 39%
Distribution yield, trailing	1.91%	0.35%
Move over the last five sessions	−5.9%	−3.4%
Distance below 52-week high	≈ 16%	≈ 6%
Assets under management	$1.85 billion	$20.8 billion

Read the first two rows against each other and the inversion is complete. FTEC's three largest holdings — Nvidia at 16.12%, Apple at 14.35%, Microsoft at 8.29% — put roughly 39% of the fund into three tickers. That single fact describes CHAT's entire top ten. The "broad" fund concentrates its eggs into a mega-cap trio; the "concentrated" fund spreads across 49 names with nothing close to a 16% single-stock position. What CHAT is concentrated in is not name-count but thesis: every major holding is there because of generative-AI revenue, which is a different risk than single-stock blowup, but a risk all the same.

The middle rows are where FTEC's case lives. Nvidia dominates both funds and Microsoft is a top holding in both, so the genuine point of disagreement collapses to a few positions: Apple, missing from the AI-revenue construction, sits at roughly 14% of FTEC as a hardware ballast; CHAT instead leans into the silicon-and-software fliers — Micron, Broadcom and the rest of the AI supply chain that rode this year's run. And the beta is real: CHAT has already given back about 16% from its 52-week high while FTEC sits 6% below its own, and in the latest five-session pullback the concentrated fund fell 5.9% against FTEC's 3.4%.

Round One: The Case for Paying for Purity

Forget who sounds smarter; the match turns on what counts. CHAT enters with the better season behind it — up 49.25% year to date versus FTEC's 25.79%, and 64.92% on the rolling twelve months against 37.18%. That is not a small edge; it is the difference between the revenue pools the AI build-out is flooding and the broad index that still drags a hardware laggard's weight through it.

The mechanism argues this can run further. A revenue-purity screen is a bet that the fastest-growing revenue, not the biggest market cap, is the better index for a technological shift. As long as hyperscaler and enterprise AI spending keeps compounding, the names at the tilt — the memory and networking suppliers feeding new data centers — should keep compounding harder than the median S&P tech name. The distribution kick adds a real point and a half a year to the total-return score, which is not nothing in a race this tight. And the top-ten cap means no single holding can sink the portfolio the way a 16% Apple weight can be a drag in a flat tape.

The bill for all that is a 0.75% expense ratio, nearly ten times FTEC's, plus the higher beta that has already cost CHAT roughly 10 more points of drawdown from its high.

Round Two: The Case for the Boring Wrapper

Give FTEC its full argument before declaring the lean. A 0.08% expense ratio is not a rounding error; over a full cycle it is a locked transfer from the fund to the holder, paid every year regardless of whether the AI thesis works. The cap-weighted basket might be concentrated at the top, but it is concentrated at the top for a reason: Nvidia and Microsoft are compounding earnings machines with pricing power, and Apple at a lower multiple than both is exactly the kind of ballast that holds up when the momentum complex rolls over.

That ballast is doing work right now. In the last five sessions FTEC lost 3.4% while CHAT lost 5.9%; from peak, the broad fund is 6% off its high against CHAT's 16%. A total-return race rewards the side that draws down less, because every percentage point protected in a correction is a percentage point that does not have to be re-earned on the way back. Add the structural point: FTEC tracks an index bearing the 25/50 label, the cap rule that prevents any single issuer from exceeding a quarter of the fund, which is the only reason Nvidia is not even bigger here.

The weakness is the flip side of the ballast. The same Apple weight that cushions a sell-off is dead weight in an AI bull — the fund is forced to hold participation in the story's slowest major digitizer — and broad tech's median constituent is a drag on the exact names where the upside concentrates.

The Reveal, and the Odds

Here is where the odds come from, and they are published at kickoff, not after the tape clarifies. The honest reading of this pairing is not "concentrated risk versus safety" — both funds are exposed to Nvidia, both carry mega-cap weights, and the concentration math runs opposite to the marketing. The real choice is whether to pay 0.75% for an AI-revenue tilt with a yield kick and a sharper drawdown, or 0.08% for a mega-cap trio with an Apple brake.

That choice is close on a coin-flip basis, but the structure argues for a narrow FTEC lean. The fee gap is the only guaranteed line on the card, compounding against CHAT annually. The drawdown gap is measurable and already visible. And the current tape shows relative strength rotating toward the cap-weighted large caps just as the momentum complex corrects. Set it at FTEC 53, CHAT 47 — an editorial probability about a hypothetical paper race, not a sportsbook line and not a promise.

The condition that flips the match is stated up front, the better to test it: if AI capex keeps compounding through 2027 without a sustained air pocket, CHAT's revenue-purity tilt should carry the scoreboard because it owns the marginal winners outright. If the build-out stalls or rotates to slower monetizers, FTEC's Apple weight and fee advantage should grind out a compounding win. That is the lead-change variable to watch at the first checkpoint at the end of September.

Final Bell, Early Call

The official score is 100.00 to 100.00 at the bell — day zero, nothing yet decided. The form entering the tape is not: CHAT ran the trailing year and the year-to-date far ahead of FTEC, and the mechanism board says that lead came from owning the AI supply chain and skipping the Apple drag. The mechanism board also says the lead carries a cost — a tenfold fee and a sharper beta — and the start-of-race edge, on balance, belongs to the cheaper, flatter wrapper by a hair.

Scoreboard verdict on the race we just started: undecided. Mechanism verdict: the labels both fail a fair test. The broad fund is the concentrated one, and the "concentrated AI" fund is the diversifier with a single-thesis bias. The design lesson travels further than either fund: when an ETF calls itself diversified, read the holdings, not the headline — cap-weighting concentrates, and revenue-purity diversifies even while it narrows the bet. Next checkpoint: the last trading day of September, with the September distribution already reinvested into both scores.