BTCI: how the ETF Goldman is buying manufactures its 26.73%

The NEOS Bitcoin High Income ETF (BTCI), the fund for which Goldman Sachs is set to pay up to $2.25 billion, warned on August 19, 2026, that 94% of its monthly distribution was a return of capital. This is the estimate from its 19a-1 notice, the document that U.S. asset managers must publish when they distribute something other than profit: of the $0.6289 per share paid on August 21, $0.5940 came from the shareholder's capital and $0.0349 from investment income, pending final tax classification. The fund's fact sheet advertises an annualized distribution rate of 26.73% and, a few lines below, a 30-day SEC yield—the metric required by the SEC, the U.S. market regulator, which only counts interest and dividends—of 1.62%. Between the two figures lies a twenty-five-point gap, and those twenty-five points are a machine with identifiable parts: U.S. Treasury bills, a synthetic Bitcoin position, and sold calls (call options) on a Cboe index. Here it is dismantled piece by piece, using the fund's actual holdings and its mechanics applied to the week of August 17 to 21, 2026, when Bitcoin rose by 22%.

What exactly is Goldman Sachs buying by paying $2.25 billion for NEOS?

On August 12, 2026, Goldman Sachs announced the acquisition of NEOS Investments in a cash-and-stock deal valued at up to $2.25 billion, expected to close in the first quarter of 2027 subject to regulatory approvals. What is changing hands are 19 option-based exchange-traded funds totaling approximately $30 billion, three of which are cryptocurrency-focused: BTCI itself, the Boosted Bitcoin High Income ETF (XBCI), and the Ethereum High Income ETF (NEHI). Goldman Sachs Asset Management already managed about $40 billion in income and defined-outcome (buffer) funds using options on traditional assets; with NEOS, its active ETF management business approaches $80 billion, the eighth largest among active ETF managers according to Morningstar.

BTCI is the crypto jewel of the package: $1,272,931,745 in net assets and 39,750,000 shares outstanding as of August 21, 2026, according to the manager's fact sheet. That is, a net asset value (NAV) of $32.02 per share. The total annual fee is 0.99% (0.98% management fee plus 0.01% in acquired fund fees), which on that asset base equates to approximately $12.6 million per year from this fund alone.

There is a detail in EDGAR, the SEC's public registry: Goldman did not enter this business without trying on its own. It registered its own Goldman Sachs Bitcoin Premium Income ETF on April 14, 2026, and has postponed its effective date four times, the last time being eight days after announcing the purchase of its competitor.

A proprietary product registered and postponed four times, and a $1.27 billion competitor bought in the interim. Whether the price is expensive or cheap will be determined by the 2027 closing; what follows is what the machine Goldman decided to buy ready-made, rather than build, actually does.

What is a covered call and why does a Bitcoin income ETF distribute every month?

Imagine you own an apartment worth 200,000 euros and you don't plan to sell it this month. Someone offers you 2,000 euros for a specific right: if in thirty days they want to buy it from you for 220,000, you are obligated to sell it at that price. If they don't want to, nothing happens and the 2,000 is yours regardless. You collect the money on day one, no matter what you do later.

The three month-end outcomes are asymmetrical. If the apartment stays at 200,000, you keep the apartment and the 2,000. If it goes up to 260,000, you must sell it at 220,000 and the extra 40,000 goes to the buyer of the right: you gain 22,000 instead of 60,000. If it crashes to 160,000, no one exercises anything and you are left with an apartment worth 160,000 and 2,000 in your pocket: you lose 38,000.

The mapping to the financial product is direct. The right you sold is called a call, a purchase option. The 2,000 euros are the premium, and that premium is what goes out the door every month in the form of a distribution. The 220,000 is the exercise price or strike, which acts as a ceiling on your participation in the upside. And the apartment is still yours through the entire downside: no one has covered you for the fall; they have only paid you to renounce a part of the future. When the option seller owns the underlying asset, the operation is called a covered call; why a fully collateralized covered call immobilizes the entire underlying is explained in our on-chain options market size analysis.

The limit of the analogy is what matters: the apartment has produced nothing, and what has produced money is the sale of a part of its future. The BTCI prospectus states it bluntly: "this strategy effectively converts a portion of the potential upside of Bitcoin price growth into current income." Converting is not the same as adding. In BTCI, on August 14, 2026, the "apartment" was worth 1,482.61 points of the Cboe CBTX index and the sold ceilings were at 1,570 and 1,630.

What is actually inside BTCI?

The manager's holdings sheet, consulted on August 23 and dated by them as August 24, contains seven lines. There is no Bitcoin among them.

Almost 60% of the portfolio consists of Treasury bills. Exposure to the Bitcoin price comes via two paths: 25.54% in spot ETFs—IBIT and HODL—and a synthetic position built by buying a call and selling a put at the same strike price (1,390) and with the same expiration (October 16), a combination that replicates the upside and downside of the underlying without paying the full cost of buying it. The prospectus explains why spot funds do not exceed that amount: they are held through a Cayman Islands subsidiary limited to 25% of total assets at the time of investment, a cap imposed by U.S. tax regulations for regulated investment companies. The 25.54% on the sheet is pinned to that limit.

The last three lines manufacture the distribution: two sold calls expiring September 18 and one put. The underlying on which they are written deserves precision, as the fund's website blurs it. Its strategy headline says the fund "distributes monthly income generated by selling calls on Bitcoin futures ETFs," but the position identifier is CBTX: the Cboe Bitcoin U.S. ETF Index, an equal-weighted basket of U.S. spot Bitcoin ETFs—market cap only determines the eligibility of each fund, according to §3.4 of its methodology—with cash-settled options and European exercise (only exercisable at expiration). The prospectus does cover this: it defines "Bitcoin-related instrument" as Bitcoin, a futures fund, a spot fund, or an index that uses any of them as a reference. The practical effect is that the ceiling is placed one step above the asset: on a basket of funds that replicate Bitcoin with their own fees.

Where does BTCI's 26.73% come from if its SEC yield is 1.62%?

The BTCI fact sheet publishes three figures that are daily confused and measure different things. The distribution rate, 26.73% as of July 31, annualizes the last distribution: the $0.6458 from July times twelve, divided by the NAV. The trailing twelve-month rate, 39.81%, sums what was distributed in a year over the current price, which is higher because the denominator has fallen. The 30-day SEC yield, 1.62%, is the only one with a regulated definition: interest and dividends accrued by the portfolio, net of expenses, excluding option premiums because they are not income.

The 1.62% aligns with what is in the portfolio. The three-month U.S. bill yielded 3.71% on August 20, 2026, according to the St. Louis Federal Reserve series; applied to 59.85% of the assets, it gives just over two gross points, minus the 0.99% fee. That is all the income the regulated metric recognizes, and the manager documents every month where the rest comes from in its 19a-1 notice.

The series of notices says the same thing month after month: the percentage of the distribution classified as return of capital was 96% in January 2025, 96% in August 2025, 95% in December 2025, 96% in January 2026, 93% in April, 92% in July, and 94% in August. Between those dates, the CBTX index went from 2,473.09 points (January 2025 notice) to its peak of 2,967.82 on October 6, 2025, up 20%, and from that peak it fell 50.0% until the close of August 14, 2026, and 38.8% until the close of the 21st; the distribution percentage barely moved throughout the journey. So far in the 2026 fiscal year, $5.7023 of the $6.1027 distributed per share—93%—is estimated return of capital.

19a-1 notices are accounting estimates, and the manager itself warns that the final tax classification may change and will arrive in the 1099-DIV form. The audited accounts, which are final, confirm the direction and lower the proportion: in the seven months ended December 31, 2025 (fiscal year change from May to December), the fund distributed $8.94 per share, of which $3.97 was return of capital—44%—, $4.22 realized gains, and $0.75 net investment income; in the previous period, ended May 31, 2025, it distributed $11.10 and $6.51 was return of capital, 59%. The difference with the monthly estimates, which range from 92% to 96%, is the subsequent tax reclassification, which is why the manager warns that the 1099-DIV may change the label. The auditor is Cohen & Company.

Returning capital is legal and common in this family of products, and it should be stated as clearly as the rest: in the United States, it is not taxed at the time of collection; it reduces the acquisition price and defers the tax until the sale, which is the tax efficiency argument the manager links on its own page. The label describes the source of the money, not its legality. The design contrast is provided by the competitor: BlackRock's iShares Bitcoin Premium Income ETF (BITA, on Nasdaq) began selling options on June 9, 2026, and its first distribution, declared on July 1, was $457,924.72 in total—what the option sales had collected in those three weeks, with no annualized target to meet. BTCI distributed in August $0.6289 for each of its 39.75 million shares: about $25 million in one month.

How much of Bitcoin's rise from August 17 to 21, 2026, did BTCI lose due to the call ceiling?

That week saw the largest weekly rise in Bitcoin since 2024, and we have it documented session by session in our rally anatomy: Friday's close near $78,335, 22% above the start, driven by two cascades of short liquidations and five sessions of inflows into spot funds.

The index on which BTCI writes its options followed the same path: the Cboe Bitcoin U.S. ETF Index closed at 1,482.61 points on Friday, August 14, and at 1,816.61 on Friday, the 21st: 334 points higher, a 22.53% increase. Each index point was equivalent to $43.12 of Bitcoin's price at the close of the 21st, and the equivalence holds throughout the episode (on August 19, with Bitcoin at $69,564, the index closed at 1,613.85: $43.10 per point).

Of the 334 points the CBTX index rose that week, 246.61 were above BTCI's low ceiling and 186.61 above the high ceiling. As a percentage of the advance: 73.8% of the rise was out of the fund's reach in the portion covered by the 1,570 calls, and 55.9% in the portion covered by the 1,630 calls. The premium collected for selling those options is the agreed price for yielding that segment, fixed before the rise occurred.

Three caveats. The holdings sheet is published after the movement: we know what ceilings were in place after the rally, but not on which day those calls were sold, although both strikes were above the August 14 market, which is how a call is sold to collect premium. The contracts expire on September 18, so they are mark-to-market in the meantime, and the prospectus allows the manager to manage positions before expiration. And the notional covered by each line is not published: this measures the ceiling on the covered portion, not the result of the entire fund. For that, there are the published returns.

Did BTCI cushion Bitcoin's fall in 2026?

The manager publishes its NAV returns—total return with distributions reinvested—alongside those of its benchmark, the S&P Bitcoin Index in dollars, as of July 31, 2026. The difference between the two columns is everything the option machinery contributed versus simply following the asset.

The cushion exists and can be measured: between 0.44 and 4.46 percentage points depending on the window, after fees. In the only bullish month in the table—July, with the index up 7.13%—the fund lagged by 0.27 points. An advertised 26.73% that translates into 2.61 points of cushioning year-to-date describes the distance between the distribution rate and the economic result quite well.

The pending test is the opposite, and it must be said: 2026 has been a down year, and in a down year, a sold-call strategy usually beats the asset. What remains to be seen is a full year of strong upside with this portfolio, where the ceiling acts month after month. The week of August 17 to 21 is only five sessions; the answer requires a full fiscal year.

What does someone who bought BTCI on day one have as of August 21, 2026?

BTCI was launched on October 16, 2024—first trading day October 17, according to audited accounts—with a NAV of $50.00 per share. As of August 21, 2026, that value is $32.02. Along the way, it has distributed $26.15 per share in twenty-three monthly payments, with a maximum of $1.5718 in January 2025 and a minimum of $0.6289 in August 2026.

The rate is calculated on the NAV at the time, so when the NAV falls, the check falls with it: the monthly payment has dropped 60% since January 2025 without the advertised rate moving from the 26% band, because the denominator shrinks at the same time as the numerator.

The two ways of counting the same investment yield different results, and both are correct. Someone who collected the distributions and spent them had, as of July 31, $25.52 collected plus a share worth around $29—derived figure: the July distribution equated to 2.23% of the NAV—: about $54.5 for every $50 invested, near a 9% nominal return in twenty-one months, excluding taxes. Someone who reinvested every distribution into more shares of the fund itself obtained the −6.68% published by the manager, because each reinvestment bought an asset that continued to fall. Just over fifteen points separate spending the check from reinvesting it, and in a bullish cycle, the arithmetic reverses: it is worth knowing which of the two figures is being shown by whoever speaks of "return since launch."

There is a third calculation, the one the manager publishes in its standardized table after taxes, using maximum U.S. marginal rates and without state taxation. As of June 30, 2026, the return since inception was −12.68% before taxes and −22.76% after taxes on distributions: a 10.08 point cumulative difference. Return of capital defers the tax; the rest of the distribution does not. Outside the United States, the treatment changes—we break it down jurisdiction by jurisdiction in our crypto tax guide by country—and the non-resident usually faces withholding at the source on each distribution.

Meanwhile, money has continued to flow in: from $987.6 million and about 22.5 million shares as of December 31, 2025, to $1,272.9 million and 39.75 million shares on August 21, 2026. 77% more shares with the NAV per share 27% lower.

When does a Bitcoin income ETF with covered calls win and when does it lose?

The performance profile of a covered call can be tabulated scenario by scenario, compared to the alternative of holding the asset outright—a spot ETF like those BTCI holds, whose mechanics are in our guide to what a crypto ETF is.

The structure is designed for sideways and moderately bearish markets, where a portfolio of sold calls collects premiums without anyone claiming the asset. What decides the outcome is the shape of the rise rather than its size: twenty flat months with two violent rally months is the worst possible environment for this design, and it is a frequent pattern in this asset. July 2026—index +7.13%, fund +6.86%—already shows the ceiling working in a single month.

What three questions should you ask a product advertising 26%?

1. Where does the flow come from? In BTCI, from three named sources—Treasury bill interest, option premiums, and shareholder capital—and the manager publishes the proportion every month in a mandatory document.
2. What has been sold in exchange? The upside segment above the strike, and the premium is its price. The useful figure is the distance from the ceiling to the market: 5.9% and 9.9% in BTCI's September calls, because a closer ceiling pays more premium but gives away more upside.
3. Is the advertised figure yield or distribution? The 26.73% says how much goes out the door; the 1.62% SEC yield says how much the portfolio generates according to the regulated definition. The distance between the two measures the conversion described in the prospectus: upside potential transformed into current income.

What Goldman is buying is a machine for converting Bitcoin volatility into a monthly distribution: $1,272.9 million within BTCI—the $30 billion deal includes all 19 NEOS funds combined—and a 0.99% fee on an asset base that has grown 29% so far this year despite the fall of the underlying. Where the twenty-five points that the regulated metric doesn't explain come from has been published since 2024, month by month, on the fund's own page.

Sources and links: NEOS — official BTCI fact sheet · August 2026 19a-1 Notice · BTCI Prospectus · SEC EDGAR — Goldman Sachs ETF Trust Amendment 727 (Aug-20-2026) · SEC EDGAR — iShares Bitcoin Premium Income ETF 8-K · Cboe — Bitcoin U.S. ETF Index (CBTX) · CoinDesk — NEOS acquisition · The Block — deal details · FRED — 3-month Treasury Bill