Most people who own BST think they own a technology fund. What they own is one of the few listed ways to hold Anthropic, and the second largest position in the whole portfolio is exactly that.
Here is the top of the fund, read from BlackRock's own page and dated 31 August 2026:
| Holding | Weight | Trades on an exchange |
|---|---|---|
| NVIDIA | 7.88% | yes |
| Anthropic | 7.07% | no |
| Broadcom | 6.01% | yes |
| Databricks | 5.87% | no |
| Microsoft | 4.98% | yes |
| Apple | 4.60% | yes |
| Anduril Industries | 4.35% | no |
| Lam Research | 3.74% | yes |
| Taiwan Semiconductor | 3.11% | yes |
| Alphabet | 2.93% | yes |
Three of the seven biggest holdings do not have a share price. Between them, Anthropic, Databricks and Anduril are 17.29% of the fund.
It is not an ETF, and that is the whole point
This gets called an ETF constantly. It is not one. BST is a closed-end fund listed on the NYSE with a fixed 34,720,019 shares outstanding, and the distinction is not pedantry, it is the reason the portfolio above is possible.
An exchange traded fund creates and redeems shares every day. That machinery is what keeps an ETF's price pinned to the value of its holdings, and it is also what stops a fund being built on assets it cannot sell on demand. A closed-end fund never redeems. The share count is fixed, the manager is never forced to sell to meet a withdrawal, and so the portfolio can carry positions that have no market at all.
That is why a listed wrapper can hold 17% in companies you cannot buy.
Compared with the fund actually sold for this
The obvious comparison is the ARK Venture Fund, which is marketed as democratised venture capital and exists to give ordinary investors private exposure. Its ten largest holdings, same date, are all private: SpaceX 7.54%, Kalshi 5.81%, Ayar Labs 5.65%, OpenAI 5.26%, Stripe 4.16%, Anthropic 3.86%, Lila Sciences 3.71%, Crusoe 3.66%, Tenstorrent 3.36%, Figure AI 2.35%. About 45% of the fund in the top ten alone.
So the dedicated venture vehicle holds Anthropic at 3.86%, and the technology fund holds it at 7.07%. BST carries 1.8 times the Anthropic weight of the fund built to deliver exactly that.
They arrive from opposite directions. ARKVX is a venture portfolio that happens to include Anthropic. BST is a large-cap technology portfolio with three private positions bolted into the top seven. Note also that ARKVX is an interval fund bought at net asset value rather than traded on an exchange, so the two are not interchangeable as things you can actually buy.
Both sides of the same relationship
There is a quieter feature of the holdings list. BST holds Anthropic directly at 7.07% and Alphabet at 2.93%. Alphabet is among Anthropic's investors, so a holder owns the private company and a public shareholder in it at the same time, and owns Google's search business alongside one of the companies most often named as a threat to it.
We are not going to put a number on how much Anthropic exposure arrives through Alphabet. Doing that properly needs the carrying value of Alphabet's stake from its own filings, and we have not read them. What we can say is that the two sit in the same fund, near the top, and they are not independent bets.
The record, across every window
This is where it gets awkward for anyone who wants a simple answer. All figures are annualised total return with distributions reinvested, every window ending 31 August 2026, each read from the relevant manager's own published table.
| 1 year | 3 year | 5 year | 10 year | |
|---|---|---|---|---|
| BST | +46.83% | +24.45% | +6.12% | +21.06% |
| QQQ | +26.30% | +24.47% | +14.20% | +20.78% |
| SCHD | +29.53% | +16.19% | +10.01% | +13.17% |
| S&P 500 total return | +20.38% | +21.04% | +12.79% | +15.37% |
Read it honestly and it says four different things. Over one year BST is far ahead of everything. Over three it is level with QQQ and ahead of the rest. Over ten it is level with QQQ again and well ahead of SCHD and the index. Over five it loses to all of them, and not narrowly.
One window contradicts the other three, so anyone can prove whatever they already believed by choosing where to start. That is the most important sentence in this piece.
The five-year hole has a cause. BST's calendar year returns on net asset value run +9.21% in 2021, -39.56% in 2022, +30.63% in 2023, +24.11% in 2024 and +19.15% in 2025. The five-year window opens immediately before a 40% down year. On market price the 2022 figure is -38.23%.
A note on which series to read: BlackRock publishes BST on both market price and net asset value, and over ten years they differ enough to change a conclusion. Market price gives 21.06% a year, net asset value gives 20.34%. QQQ did 20.78%. So on one series BST edges QQQ and on the other it trails. That is why the honest word is level, not beat.
The discount, and why it exists at all
Because a closed-end fund never redeems, nothing forces its share price to equal the value of its holdings. BST's price sat 6.79% below net asset value on 28 September 2026, at $47.94 against a stated NAV of $51.43.
Two things follow. A buyer at a discount pays less than the assets are carried at, which no ETF offers. And a discount can widen as easily as it narrows, so it is a risk as much as an opportunity. BlackRock has a discount management programme it has publicly renewed, which tells you the firm regards the gap as a problem worth addressing.
Do not treat any single day's discount as a standing fact. BST rose 4.17% on 29 September, the day after the figure above was struck, and because NAV is published with a lag there is no way to compute that day's true discount. Price and NAV have to be measured on the same date or the number is meaningless.
The part we could not verify, and it matters
BST shows a distribution rate of 6.26%. That number is the single most common reason people buy funds like this, and it is the number this article can say least about.
A closed-end fund's distribution can be paid out of income, realised gains, or return of capital, and return of capital is, in plain terms, some of your own money coming back. The split is what determines whether a 6.26% distribution rate is income or partly an illusion. BlackRock's page has the table with the right columns, including return of capital, but the rows did not load, and we are not going to characterise the payout from a headline rate alone.
So: the rate is verified at 6.26%, and the composition is not. Anyone buying this for the income should read the fund's Section 19(a) notices and annual report before treating that figure as yield. We will update this piece when we have read them.
Costs and the plain facts
BST charges a 1.11% gross expense ratio, of which 1.00% is the management fee. For comparison SCHD charges 0.060%. The fund holds 89 positions, has about $1.79bn in net assets, is listed on the NYSE and launched on 28 October 2014. Geographically it is 88.36% United States, 5.52% Taiwan and 3.83% Korea.
What this article does not say
It does not say BST is a good or bad holding, and it makes no comparison of the two funds as investments beyond the weights and returns printed above. It does not say the private positions are correctly valued. In fact the opposite deserves emphasis: a fund with 17% in unlisted companies has a net asset value that is partly BlackRock's estimate rather than a market price, and every discount, premium and total return figure quoted here inherits that estimate. The bigger the private weight, the bigger the estimate.
And it does not say these are the only listed routes to Anthropic. They are the two we verified.
All figures read from BlackRock, ARK, Schwab and Invesco's own published pages, plus the S&P 500 total return index computed at matching dates. Holdings and returns as of 31 August 2026; fund facts as of 28 September 2026. Educational only, not investment advice.