The momentum fund sold Nvidia. Not because Nvidia is broken, but because a rulebook said so, and the rulebook does not care that the stock is up roughly 30% over the past year.
At the reconstitution that took effect after the close on Sept. 18, 2026, the S&P 500 Momentum Index dropped 54 names and added 54. Nvidia was one of the exits, ranked just outside the top 100 by momentum score. Broadcom left with it.
Apple arrived as the largest single addition, at 9.38% of the fund. Micron stayed the second-largest position, at 8.99%.
That is the whole lesson of this product. Momentum is not a synonym for owning the winners forever. It is a mechanical screen, refreshed twice a year, that re-picks roughly half its names and hands the money to whatever has run hardest.
In September that meant selling 34.00% of the prior portfolio's weight. The 54 replacements had already averaged +86.45% over the trailing year before they were bought. You are, by construction, buying what has already run and selling what is merely still good.
This is the Invesco S&P 500 Momentum ETF, ticker SPMO, on NYSE Arca since October 2015. It charges 0.13%, holds 102 positions and runs $23.54 billion. What follows is what you own, how the machinery works, what it costs, what it has returned, and the strongest case against it.
What you actually own
Start with the snapshot from Sept. 18, 2026, taken before the swap. Micron was the largest holding at 11.31%, Nvidia second at 8.98%, Broadcom third at 6.12%, Johnson & Johnson fourth at 4.65% and AMD fifth at 4.53%.
Alphabet's Class A shares took 4.46% and its Class C shares 3.52%, because the company issues two classes of stock. Lam Research held 3.23%, Exxon Mobil 3.16% and Intel 2.82%. Those ten positions carried 52.78% of the fund.
The next tier ran SanDisk 2.63%, Caterpillar 2.43%, Cisco 1.85%, Seagate 1.85% and Applied Materials 1.77%. Everything below that is a rounding error with a ticker attached.
Two holdings deserve a second look. Johnson & Johnson, a consumer-healthcare conglomerate, sat in the top five of a momentum fund. Micron, the most brutally cyclical business in technology, was the single biggest position.
More than half the fund in ten names is a statement about intent. The other 92 positions are there for structure, not for return, and a bad quarter at one or two companies lands in your account.
After the swap the arithmetic shifted but the shape did not. Apple and Micron together carry roughly 18% of the fund. It is still about 52% technology, and it still owns the memory complex: Micron, SanDisk, Seagate and Western Digital.
Sector weights as of Sept. 23, 2026, set against the plain S&P 500 tracker VOO, tell the same story. Technology is 52.84% of SPMO against 38.72% of VOO, roughly 14 points of overweight.
Financials are 6.22% against 12.02%, and consumer cyclical is 1.25% against 9.31%. SPMO holds about half the financials and about an eighth of the consumer discretionary of the index it is drawn from.
The remaining gaps are smaller but consistent. Healthcare is 7.40% against 9.29%, communication services 8.01% against 9.50% and industrials 12.28% against 7.75%. Energy is a wash at 3.49% against 3.48%.
How the sausage is made
SPMO is an index fund, which means nobody here is picking stocks. It tracks the S&P 500 Momentum Index: about 100 S&P 500 companies carrying the highest momentum score, weighted by market cap multiplied by that score.
Cap-weighting means a larger company gets a larger slice. Multiplying by the momentum score tilts it further toward whatever has already run. That is how Micron cleared 11% before the swap and how Apple walked in at 9.38%.
The index reconstitutes and rebalances semi-annually, after the close of the third Friday of March and September. The fund is non-diversified, keeps at least 90% in index names and uses full replication, meaning it buys the whole list rather than a statistical sample.
One detail matters more than the rest. The September basket was built from momentum data through July 31, roughly seven weeks stale by the time it was implemented. In a market rotating this fast, you are buying a July snapshot in late September.
Reported portfolio turnover is 44%, which is what happens when the rules re-pick half the book twice a year and nothing caps how much can move. The rival momentum ETF has a rule for exactly that.
MTUM tracks MSCI's USA Momentum SR Variant Index, which draws 125 names from a large and mid-cap parent and scores each one on risk-adjusted momentum. The score compares excess return over the 3-month Treasury bill with the annualized standard deviation of weekly returns, then averages the result over 6- and 12-month windows.
That index caps every issuer at 5% at reconstitution and keeps the aggregate weight of names above 4.5% below 22.5%. It reconstitutes quarterly, caps one-way turnover at 30% and uses representative sampling rather than full replication.
SPMO has no issuer cap that shows in the data and no turnover cap at all. That is how a third of the portfolio can move in a single rebalance. Same factor, two rulebooks, and a gap of 7.62 percentage points a year over five years.
One structural detail is quietly in your favour. ETF shares are created and redeemed in kind, so the fund hands over baskets of stock instead of selling to raise cash. That keeps realized capital gains out of the wrapper, which is a large part of why index ETFs hand shareholders so few taxable surprises.
The label on the box tells you very little about what is inside it. Momentum is the label. Turnover, caps and weighting rules are the product.
The bill
The expense ratio is 0.13%. That is the annual fee the fund takes from your money, expressed as a percentage of assets, and it is charged quietly rather than invoiced.
Put it in hands. Thirteen basis points is $13 a year on a $10,000 position, or about $130 over a decade if the balance never moves. Every basis point is $1 a year per $10,000.
Against VOO's 0.03% you are paying a 10-basis-point premium, roughly $10 a year on every $10,000, for the momentum screen. Against a 1% active fund, SPMO costs about an eighth as much.
Its closest rival, MTUM, charges 0.15%, so the gap between the two momentum funds is two basis points. The larger costs sit elsewhere.
There is the spread between the price you pay to buy and the price you receive to sell. There is also the gap between the fund's market price and the value of its holdings, called the premium or discount to NAV. On a fund this size both are normally small, and neither is zero on a violent day.
Then there is tracking difference, the gap between what the index returns and what the fund returns after fees. SPMO also pays a 0.73% dividend yield quarterly, and in a taxable account that income is taxed whether you spend it or not.
The fee question is settled, though. Thirteen basis points for a rule-based screen is cheap, and the real cost of owning SPMO is the risk rather than the fee.
The scoreboard
These are dividend-adjusted numbers through Sept. 23, 2026. SPMO has returned +24.82% over one year, +37.83% annualized over three, +20.49% over five and +20.36% over ten.
MTUM, the same factor with a tighter rulebook, did +23.11%, +30.89%, +12.87% and +16.71% on the identical windows. VOO, the plain S&P 500 tracker, did +16.32%, +21.56%, +13.00% and +15.60%.
Turn it into money. $10,000 in SPMO five years ago became $25,610. The same $10,000 in MTUM became $18,560 and in VOO $18,579, so the momentum fund finished about 38% ahead of the index fund.
Over ten years the gap widens further: $63,045 against $46,212 and $42,014. A fee difference of a tenth of a percentage point did not produce that spread.
The calendar record is more humbling, and worth reading properly. SPMO beat the S&P 500 in six of the last ten calendar years. It lost to it by 5 to 9 points in the other four: 2016, 2019, 2021 and 2023.
The winning years were rarely modest. It gained +45.82% in 2024 against VOO's +24.98%. So far in 2026 it is up +27.44% against +13.40% for VOO and +26.37% for MTUM.
Then there is 2022, the year momentum was supposed to fall apart. SPMO fell −10.45% while VOO fell −18.17% and MTUM fell −18.27%. It lost less than the market it is drawn from.
Drawdown tells a similar story. The worst peak-to-trough stretch since 2016, measured on monthly closes, was −30.95% for SPMO, −33.99% for VOO and −34.08% for MTUM, all bottoming in March 2020.
One caution on reading that table. Six winning years out of ten, with a 38% edge over five years, is the record of a factor that pays handsomely in a decade of technology leadership.
What you are really betting on
Strip the branding away and SPMO is a bet that the stocks already winning keep winning. That is a real factor with decades of academic work behind it. It is also a bet with a specific failure mode.
First, the mechanic forces you to sell winners. Nvidia was up about 30% over the year and got the boot for ranking just outside the top 100. Broadcom, GE, Palantir and GE Vernova left the same way.
The 54 names that left had averaged +20.40% over the trailing year. The 54 that arrived had averaged +86.45% and were, by definition, already expensive.
Second, this is not a diversified fund. It is a concentrated idea about the leaders of a bull market continuing to lead. Half the fund is technology against 38.72% for the S&P 500, and the top ten carried 52.78% before the swap.
Third, the discipline that produced the 2024 gain is the same discipline that produced the misses. It lagged the S&P 500 by 5 to 9 points in four of the last ten calendar years. MTUM, with the tighter rulebook, trailed the S&P 500 by more than 15 points in 2021 and more than 17 in 2023.
That comparison is uncomfortable for anyone who thinks a better-specified momentum fund is the fix. Momentum goes years without paying, and tighter rules did not spare MTUM the wait.
Fourth, the signal is stale by design. The September basket was built from data through July 31. You are buying a July snapshot in late September while the market rotates beneath you.
Fifth, crash risk is documented rather than hypothetical. Daniel and Moskowitz's "Momentum Crashes" is the canonical evidence. Momentum's worst stretches arrive when beaten-down names snap back hardest, and the losses cluster into a handful of months.
The fund's worst year so far is 2022 at −10.45%. It has not yet lived through a genuine momentum unwind, and the beta it carries into one is 1.30 over five years, down from 1.39 before the swap.
Sixth, here is what would prove the bears right. A sharp rebound in the laggards of 2026 while Apple and the memory complex stall. Or a market-wide risk-off in which a 52%-technology, non-diversified fund does what a 1.30 beta says it does.
One more number to sit with. After the swap the portfolio trades at 18.05 times forward earnings with 17.78% expected growth in next-year earnings per share and a 25.00% weighted-average return on invested capital. Those are growth-stock numbers being bought into a mature bull market.
None of this makes SPMO a bad fund. It makes it a specific bet, and specificity is exactly what an investor is paying 0.13% to purchase.
How to use it
SPMO suits an investor with a long horizon who wants an explicit tilt toward large-cap leadership and who will read a rulebook before buying. It works as a satellite position beside a broad index fund, sized small enough that a lag year does not change your plan.
It does not suit anyone already tech-heavy. If your S&P 500 fund is the core holding, SPMO mostly adds a second helping of the same sector at a 14-point overweight.
It does not suit anyone who needs income. A 0.73% yield paid quarterly will not fund much. And it does not suit anyone who will sell after a year when the screen lags by five points. That is the specific pain this rulebook keeps handing out.
If you want the factor without the concentration, the honest comparison is MTUM, which runs the same idea with an issuer cap and a turnover limit. You pay two basis points more and accept a lower ceiling in exchange for a rulebook built to blunt the crowding.
Worth saying plainly: nothing in this structure is broken. The fund does what it says, cheaply and tax-efficiently. The only question left is whether the bet fits you.
The verdict
SPMO is a well-built, cheap wrapper around a bet that has worked for a decade and is now the consensus view. Own it deliberately, and at a size you can hold through a lag year.
What to watch next. The March 2027 reconstitution, where the index re-picks from a fresh dataset and may decide the memory complex no longer qualifies. Whether Apple's 9.38% arrival gets trimmed at the next semi-annual review, or grows from there.
And whether the laggards of 2026 start ranking back in, which is the moment the same mechanic quietly reverses and the fund starts buying the stocks it sold.
One line to finish. This is a momentum label wrapped around a concentrated portfolio, and the gap between those two things is worth 14 points of technology exposure.
Analysis, not advice. We may own these names and we may sell them at any time.
