DXYZ Review: Paying a Premium for SpaceX, OpenAI and Anthropic

Almost every conflict here runs back to one person, so start with the entities and who owns them.

The legal shape

Destiny Tech100 Inc. is a Maryland corporation formed on November 18, 2020, registered under the Investment Company Act of 1940 as a non-diversified, closed-end management investment company, with its office in Austin, Texas (Form N-2 filings, 2022 to 2026). A closed-end fund is a pooled vehicle with a fixed share count that trades on an exchange like a stock. It does not create or redeem shares when you buy and sell, the way an ETF does. The consequence, which is the whole story of this review, is that the market price is set by supply and demand for the shares, and can sit far above or far below the value of what the fund owns.

The fund is not a broker, not a marketplace and not an issuer of private company shares. It is a buyer of them. Buy DXYZ and you own a share of a corporation that owns interests in special purpose vehicles that in turn hold shares of private companies. You do not hold SpaceX stock; you hold a claim on a fund that holds a claim on a vehicle that holds it, and the layers matter for fees, for voting and in a wind-up.

The adviser and the owner

The investment adviser is Destiny Advisors LLC, a Delaware limited liability company and a wholly owned subsidiary of Destiny XYZ Inc., which is controlled by Sohail Prasad, the fund’s chairman, president and chief executive (Form N-2/A, 2024 and 2025). Prasad co-founded Equidate, later renamed Forge Global, with Samvit Ramadurgam after the pair met in Y Combinator’s summer 2012 batch. They hired Forge’s chief executive in 2018 and were still nonvoting directors with stakes there in June 2021, when they raised $5M for the stealth venture that became Destiny (TechCrunch, June 22, 2021). Destiny does not publish the adviser’s total regulatory assets under management.

The ownership chain is short and it is the source of the conflicts. The adviser is paid on the size of the fund. It is also the Valuation Designee appointed by the board under Rule 2a-5 of the 1940 Act, so it determines the fair value of the private holdings the fee is charged on, with quarterly input from an independent third-party valuation agent (Form N-CSR, 2025 and 2026). An independent agent in the loop is the right structure and standard for this kind of fund. It does not change the fact that the party whose pay rises with the marks is the party responsible for the marks.

How it raised money, and from whom

Destiny did not start as a listed fund. In January 2021 it began a private offering of Simple Agreements for Future Equity under Rule 506(b) to qualified purchasers. SAFE holders approved a mandatory conversion at $10.00 a share, and on May 11, 2022 each received shares equal to their investment divided by $10.00. After the conversion and a reverse stock split the fund had 10,879,905 shares outstanding (Form N-2/A, 2022). Early backers reported in the press included Quora co-founder Charlie Cheever and Coinbase co-founder Fred Ehrsam.

On March 26, 2024 the shares listed on the NYSE, a direct listing of existing stock that raised the fund no new money. The reference price was $4.84, the NAV per share at December 31, 2023; the stock opened at $8.25, up 70.5%, traded as high as $30.48 and closed at $9.00 on 540,000 shares (BusinessWire, March 21, 2024; IPO Edge, March 2024). Anyone who put in $10.00 a share in 2022 was, on paper, underwater on NAV and above water on price within an hour.

The money came later, through a shelf. On August 8, 2025 the fund entered an Open Market Sale Agreement with Jefferies LLC for an at-the-market programme of up to $1,000,000,000 (Form 424B5, August 8, 2025). In 2025 it sold 11,096,400 shares for net proceeds of $324,015,375. Between April 1 and June 30, 2026 it sold 17,191,674 shares at a weighted average price of $41.82 for net proceeds of $715,442,732 (Form 424B3, August 28, 2026). Jefferies takes its commission out of the proceeds: on that quarter’s figures the gap between $718.9M of gross sales and $715.4M of net proceeds was $3.4M, about 0.5%. We could not verify the contractual cap in the sale agreement. Those two filed numbers also give the share count this review leans on: 30,465,664 shares at March 31, 2026 plus 17,191,674 sold in the quarter is 47,657,338 at June 30, 2026, which is exactly the figure a market analysis published on September 14, 2026 read off the schedule of investments (ElliottWaveTrader, September 14, 2026).

The size, at the last reported dates

Net assets were $438,044,192 at December 31, 2025 on 21,976,305 shares, which is $19.93 a share; the press release rounded it to $19.97 (Form N-CSR for 2025; company release, February 11, 2026). Six months later the portfolio had an aggregate fair value of about $1.64B including short-term investments, and NAV per share was $34.30 (company release, August 29, 2026). On 47,657,338 shares that is roughly $1.635B of net assets, a 3.7 times increase in six months, most of it new money rather than investment gain. The track record section does that arithmetic, because it is the most misunderstood number about this fund.

Here is the full path of a dollar, from your brokerage app to a private company’s cap table and back.

1. You buy a share on the exchange

There is no application, no subscription document, no accreditation test and no minimum. DXYZ trades on the NYSE and can be bought in a taxable account, a traditional IRA or a Roth IRA at any broker offering listed US equities (Destiny FAQ, retrieved September 2026). The minimum is the price of one share, about $31 in mid-September 2026. Most retail brokers charge no commission, so your entry cost is the spread plus, crucially, whatever premium to NAV the market is charging that day. Your money does not go to the fund; it goes to the person selling you the share. The only time your money reaches the portfolio is when the fund issues new shares through the Jefferies programme, and that has a legal limit we come to below.

2. The fund sources a private position

Destiny buys exposure two ways. Occasionally it buys directly into a company’s preferred round. More often it buys units in a special purpose vehicle organised by a third party that already holds, or is buying, the underlying shares. An SPV is a single-deal fund: one company, one pool of investors, usually a fixed fee to the organiser. The schedule of investments at September 30, 2025 showed the fund’s entire SpaceX position held through three of them, DXYZ SpaceX I, MWAM VC SpaceX-II and a series of Celadon Technology Fund VIII. The 2026 Anthropic position sits in Magnitude ANC III, LLC and the 2026 OpenAI position in Goanna Capital 26E LLC (Form 424B3 filings, February and August 2026).

This matters three ways. The SPV organiser is usually paid, so there is a fee layer beneath the fund’s own fee. The SPV, not the fund, is on the company’s register, so the fund has no direct information rights and no vote. And the underlying company often has consent rights over transfers, which is why an SPV interest is frequently the only way in.

3. The adviser marks the book every quarter

Each quarter the adviser, as Valuation Designee, values every Level 3 investment, meaning every holding without an observable market price, under Rule 2a-5 and ASC Topic 820, with input from an independent third-party valuation agent (Form N-CSR, 2025). Inputs are the most recent financing round, secondary transactions, comparables and discounts for lack of marketability. The output is NAV per share, published quarterly, roughly two months after quarter end.

Two months is a long lag for assets whose comparables move daily. When you read that DXYZ trades at a given premium, check which NAV date it is measured against. On May 21, 2026 the fund’s own prospectus supplement described the shares at $61.66 as a 151% premium to the March 31, 2026 NAV of $24.56, a NAV already seven weeks old when it was filed.

4. The fund gets paid, quarterly

The adviser takes 2.50% a year of average gross assets, computed on the average of the two most recently completed calendar quarters and payable quarterly (Form N-2 and Form 424B5 filings). Gross assets, not net assets: the fee applies to cash as well. There is no incentive fee or carried interest at the fund level, better than the interval funds it competes with. Operating costs, custody at U.S. Bank National Association, audit, legal, directors and exchange fees come out on top.

5. You exit by selling to somebody else

There is no redemption right. None. A closed-end fund does not buy your shares back on demand, and DXYZ has no interval or tender structure. Your exit is a market order at whatever the marginal buyer will pay. In August 2026 the board approved a share repurchase programme allowing the fund to buy back stock at prices below the then-current NAV, in amounts and at prices the adviser deems appropriate (Form N-CSRS for the period ended June 30, 2026). That is a support mechanism, not a redemption right, and the adviser decides whether to use it.

Destiny sells one product: the common stock of one fund. There is no menu, no deal-by-deal selection and no share class choice, a real simplification against every marketplace competitor. What varies is what is inside the fund, and by mid-2026 that had changed a great deal.

The portfolio as of June 30, 2026

The fund’s stated ambition is a portfolio of 100 top venture-backed private technology companies, which is where the name comes from. It has never been close. Morningstar counted fewer than 25 holdings in April 2024; the fund’s filings put the count at 32 at December 31, 2025 and 36 at March 31, 2026, when about 68.6% of the $742.5M portfolio sat in private technology issuers (Morningstar, April 18, 2024; Form 424B3, 2026). We could not verify a count at June 30, 2026, because direct access to EDGAR was blocked from our network.

What we can verify is the shape, and the shape is unusual. Reported positions at June 30, 2026, as percentages of net assets: a money market fund at $939.7M, 57.5%; Anthropic, via Magnitude ANC III, at $235.7M, 14.4%; SpaceX, across three vehicles, at $173.1M, 10.6%; Shield AI 4.1%; Databricks about 4.0%; Beast Industries 3.5%; xAI 3.5%; Revolut 2.9%; and OpenAI about 2% (ElliottWaveTrader, September 14, 2026, reading the June 30, 2026 schedule of investments). The fund’s own prospectus supplement of August 28, 2026 confirms the two largest lines at 57.3% and 14.4% of the $1.64B portfolio, a slightly larger denominator than net assets; the rest of the list rests on that single third-party reading.

That is a private-technology fund with a majority of its net assets in Treasuries. The cause is mechanical: the fund raised $715M in one quarter through the at-the-market programme and had not deployed it. The cost is not. At 2.50% of average gross assets, holding $939.7M in a money market fund costs shareholders about $23.5M a year in management fee alone, against a money market yield the fund does not disclose separately.

What changed after the quarter end

Deployment followed. After June 30, 2026 the fund closed three investments totalling about $169.0M: $150.0M into Goanna Capital 26E LLC, an OpenAI Class A vehicle, fully closed August 27, 2026; $15.0M into Magnitude FSTK, LLC for Fluidstack Ltd; and $4.0M into Boom Technology, Inc. (Form 424B3, August 28, 2026). That lifts the private book by roughly $3.55 a share and cuts cash to roughly 47% of net assets on an unchanged NAV, still the fund’s largest single line.

What is gone

The 2024 and 2025 portfolio was an aerospace and consumer-internet book: SpaceX was 52.4% of the fund’s economic exposure at December 31, 2024 (company release, April 4, 2025), with Relativity Space, Boom Supersonic, Epic Games, Discord, Stripe, Klarna, Plaid and Flexport in small weights. SpaceX went public on June 12, 2026, pricing at $135 on Nasdaq under SPCX and closing its first day at $160.95 for a market value near $2.1T (CNBC, June 11 and 12, 2026). The fund’s largest historical position became a public stock inside private vehicles, subject to lockup releases running through 2026. The 2026 fund is an artificial-intelligence book instead: Anthropic and OpenAI together were about 16.5% of net assets at June 30, 2026, and more after the August purchase.

The fee page is short and the real cost stack is not. This section counts everything, including the cost that appears in no fee table at all.

The stated fees

• Management fee: 2.50% a year of average gross assets, payable quarterly on the average of the two most recently completed calendar quarters. It was 2.00% before the NYSE listing (Form N-2, 2024).
• No incentive fee and no carried interest at the fund level. This is a genuine advantage over Robinhood Ventures Fund II, which charges 20% of gains, and over most private funds.
• Other fund expenses: administration, custody at U.S. Bank, audit, legal, transfer agency, directors’ fees and listing costs. The 2024 Form N-2 fee table estimates other expenses at 3.62% of net assets, which with a management fee shown at 2.66% gives a total of 6.28%. An earlier estimate from a 2024 SEC filing put the all-in figure at 4.98%, the number ARK used against Destiny in April 2024 (Reuters, April 2024). Both are the fund’s own disclosures on different bases; 6.28% is the later and more conservative, and it is an estimate rather than an audited outturn.
• Acquired fund fees and expenses: 0.00% in the same table. For a fund holding most of its private exposure through third-party special purpose vehicles, a zero here is worth reading twice. It is defensible on the technical definition of an acquired fund, and it means the SPV-level economics sit outside the headline number.
• At-the-market selling commission to Jefferies on newly issued shares, paid out of the proceeds and so borne by every shareholder. In the second quarter of 2026 it ran at about 0.5% of gross sales, $3.4M on $718.9M, and it reduces the accretion new issuance produces.
• Your broker: typically $0 commission at US retail brokers, plus the spread.

The fee nobody prints: the premium

Every figure above is small next to the premium. Pay $61.66 for $24.56 of assets and you have paid a 150% upfront load that no prospectus calls a load. It is recoverable only if the next buyer also pays a premium, or if the assets grow by 151% before you sell. The fee table measures the drag. The premium measures the hole you start in.

Worked example one: the premium buyer

Take $10,000 invested on May 21, 2026 at $61.66, the price and date in the fund’s own prospectus supplement. That buys 162.18 shares. NAV at the most recent reported date, March 31, 2026, was $24.56, so your $10,000 bought $3,983 of net asset value and $6,017 of premium.

Now hold five years. Assume the private book compounds at 15% a year gross, aggressive but not absurd for late-stage AI and space names, and that expenses run at 6.28% a year of net assets, so net compounding is 8.72%. Over five years that is a factor of 1.519: NAV per share reaches $37.31 and your 162.18 shares are worth $6,051 if the premium has gone to zero. You are down $3,949, a loss of 39.5%, while net asset value rose 51.9%. Getting back to $61.66 at 8.72% net needs NAV to rise 2.51 times, or about 11 years. The same $10,000 in a Nasdaq-100 index fund at an assumed 12% gross less a 0.20% expense ratio is $17,467 after five years. The gap is $11,416, and it is created on the day you buy, not by the fee.

Worked example two: the buyer in September 2026

The arithmetic reverses when the premium does. At roughly $32 on September 14, 2026 against a June 30, 2026 NAV of $34.30, you are paying a small discount, and you can see through it. Cash was $939.7M across 47,657,338 shares, about $19.72 a share, so the private book was about $14.58 a share at the adviser’s marks. Paying $32 means paying $12.28 for private assets the adviser values at $14.58, a 15.8% discount, plus $19.72 for cash you will be charged 2.50% a year to hold.

At $31.20, the quote we retrieved on September 17, 2026, the discount on the private book widens to 21%. That is a different investment from the one people bought in April 2024, and the first time in the fund’s listed life that the arithmetic has favoured the buyer. It is also why the board authorised buybacks below NAV in August 2026 and why the at-the-market machine stopped: Section 23(b) of the Investment Company Act generally bars a closed-end fund from selling common stock below net asset value without shareholder approval. The issuance engine runs only at a premium, so the fund’s ability to grow is a function of the market’s enthusiasm, not of the opportunity set.

The fund’s communications blend three things: the change in net asset value per share, the change in the price of the stock, and the return an investor earned. Here they have almost nothing to do with one another.

What the fund claims

The 2025 annual report states that for the twelve months ended December 31, 2025 the fund’s net asset value increased by 209.59%, against 20.36% for the Nasdaq Composite (Form N-CSR, filed 2026). For 2024 the fund reported that its investments generated a return of 33.12%, against 28.64% for the Nasdaq Composite, and that NAV per share ended the year at $6.44, up from $5.32 at the end of the third quarter (BusinessWire, April 4, 2025). These are claimed figures on unrealised marks, and the 2025 number in particular is not what it appears to be.

The 2024 figure is the control that proves the point. The fund issued no shares that year: the count stood at 10,879,905 at both ends of it. NAV per share went from $4.84 to $6.44, a rise of 33.1%, the investment return the fund reported, to the decimal. When Destiny does not sell stock, NAV growth is performance. When it does, it is not.

Where the NAV growth actually came from

A closed-end fund that sells new shares above net asset value raises NAV per share for everybody, instantly, with no investment skill involved. Buy a share at $41.82 when NAV is $24.56 and the $17.26 of excess is spread across all existing shares. This is accretion, and it did most of the work. Take the second quarter of 2026, where every input is public. Shares outstanding at March 31, 2026 were 30,465,664 and NAV was $24.56, so net assets were $748.2M. The fund sold 17,191,674 shares for net proceeds of $715,442,732, an average of $41.615 after commissions. Shares reached 47,657,338 and NAV $34.30, so net assets were $1,634.6M. Net assets rose $886.4M, of which $715.4M was new money, leaving $171.0M from operations, or $3.59 a share on the ending count. Issuing 17,191,674 shares at $17.055 above NAV added $293.2M, or $6.15 a share.

$6.15 plus $3.59 is $9.74, and NAV per share rose exactly $9.74, from $24.56 to $34.30. Sixty-three percent of the quarter’s NAV growth was selling stock above NAV, not investing.

The same identity closes 2025, where the effect is larger still. The fund began the year with 10,879,905 shares at a NAV of $6.44, so net assets were $70.1M. It sold 11,096,400 shares for net proceeds of $324,015,375, an average of $29.20 a share. It ended with 21,976,305 shares and net assets of $438,044,192, a NAV of $19.93. Operations contributed $43.96M, or $2.00 a share; issuing stock at $29.20 against an opening NAV of $6.44 contributed $11.49 a share. That is 85% of the year’s $13.49 increase, and it is not an artefact of the reference date, because the average issue price beat even the NAV the fund closed the year at.

This is not a trick, and it is not unique to Destiny. It is standard behaviour for a closed-end fund with a large premium and a shelf. But a NAV-per-share series is not a return series for this fund, and anyone quoting 209.59% as performance is quoting the wrong number.

What a shareholder actually earned

The price series is brutal and public.

A buyer at the April 8, 2024 close of $99.79 who held to mid-September 2026 was down roughly 68% over two and a half years, during which net asset value per share rose more than seven times. Even the $43.50 close eight days later, after the first collapse, was a 900% premium (Morningstar, April 18, 2024). A buyer at $61.66 on May 21, 2026 was down about 48% four months later. A buyer at the $8.25 listing-day open was up roughly 290%. The fund’s assets did well; its buyers, in aggregate, did not, because most of the volume happened at premiums that have since collapsed.

The realised record

There is almost none, and that is the honest answer. The fund has never paid a distribution since inception (Nasdaq and Zacks dividend histories, September 2026), and its stated policy is that it does not anticipate paying quarterly dividends or becoming a predictable distributor, because it makes capital-gains-oriented equity investments (Form N-2). The one genuine liquidity event was the SpaceX IPO on June 12, 2026, and the fund’s SpaceX exposure sits inside SPVs subject to lockups rather than in shares it could sell freely on day one. Every other gain in the NAV series is an unrealised mark set quarterly by the adviser as Valuation Designee, with an independent valuation agent’s input.

The Anthropic position is the cleanest illustration of how those marks are built. The fund committed $100M on January 26, 2026 to a Series B preferred stake through Magnitude ANC III, when Anthropic had a term sheet for a $10B round at a $350B valuation, later expanded to about $20B at the same price (CNBC, January 7, 2026). At March 31, 2026 it was already marked at 18.1% of the $742.5M portfolio, about $134.4M, a 1.34 times mark nine weeks after purchase and before any new round. By June 30, 2026 it was carried at $235.7M, a 2.36 times mark in five months. In May 2026 Anthropic closed a $65B round at a $965B valuation (CNBC, May 28, 2026), 2.76 times the January level. The fund’s mark is consistent with that round and slightly conservative against it. That is the right way to mark a position. It is still a mark, not money.

The liquidity story is genuinely good in one direction and absent in the other, and confusing the two is the most common mistake with listed private-market funds.

What you get

Daily liquidity at a market price, in any brokerage account, with no lockup, no accreditation and no subscription documents. Against every alternative in this category that is a real advantage. A Forge or EquityZen buyer waits 45 to 60 days for a trade to close and then has no exit until an IPO. An ARK Venture Fund holder can redeem only in a quarterly repurchase offer capped at 5% of shares. A DXYZ holder can be out in a second.

What you do not get

You do not get liquidity at net asset value, and you never will unless the market gives it to you. There is no redemption right, no tender offer schedule and no interval structure. The August 2026 buyback authorisation lets the fund repurchase below NAV at the adviser’s discretion, which may narrow a discount but confers no right on any shareholder.

You also do not get deep liquidity in size, or at least we cannot show that you do. Reported average daily volume varies wildly across providers, from about 172,000 shares on one to about 2.19 million on another’s 30-day average, which at roughly $31 is $5.3M to $68M a day (aggregator data retrieved September 2026; we could not reconcile them and treat both as unverified). Ownership is reported as 93.00% retail, 6.54% insider and 0.16% institutional (TipRanks, September 2026, unverified third-party data). A retail-dominated register is precisely the shareholder base that produces the price behaviour this fund has had, and the 52-week range of $19.71 to $72.87 is the result.

Why the premium did not get arbitraged away

The textbook trade is to sell the overpriced fund and buy the underlying. Neither leg works here: the underlying was not investable for most of the fund’s life, so there is no hedge, and the borrow was priced accordingly. Acadian Asset Management reported about 1.78 million shares sold short in early 2024 and an annualised borrow cost above 100% by mid-April 2024 (Acadian, 2024; we could not independently confirm the borrow figures). A 2,000% premium with a 100% cost of carry can persist for a long time, and it did.

What happens if the platform fails

This is the one place where the closed-end fund structure protects you better than any competitor in the category. The assets are held by a custodian, U.S. Bank National Association, not by the adviser, and the fund is a separate registered investment company with its own board. If Destiny Advisors went under, the board would replace the adviser or liquidate the fund; shareholder money is not commingled with the operating company. Compare that with Linqto, where customers held interests in the platform’s own vehicles rather than title to the shares they thought they owned, and which filed Chapter 11 on July 8, 2025. Structure is the single best thing about DXYZ.

The tax wrapper is one of the two genuinely good things about DXYZ and the most underrated reason to prefer it to the private alternatives. It also carries a caveat the fund discloses and most coverage does not.

The form you get

DXYZ elected to be treated as a regulated investment company under Subchapter M of the Internal Revenue Code beginning with the taxable year ended December 31, 2023, and intends to qualify annually (Form N-2 filings). A RIC that distributes essentially all of its net taxable income pays no entity-level federal income tax; the income is taxed to shareholders. Shareholders receive a Form 1099-DIV if distributions are paid and a Form 1099-B from their broker when they sell. There is no Schedule K-1, because the fund is a corporation rather than a partnership or LLC.

That is a meaningful simplification. A K-1 from a private fund can arrive in August, force an extension, and carry state filing obligations in every state where the fund has activity. None of that applies here.

The RIC status is not unconditional

To stay a RIC a fund must pass the diversification tests of IRC section 851, one of which caps a single issuer at 25% of total assets. Destiny’s filings state that the SPVs holding its SpaceX exposure came to about 25.9% of total assets at March 31, 2023, that management identified the issue with its tax and legal advisers, and that remediation was contemplated through raising capital or selling investments, with the reasonable-cause exception as a fallback. Losing RIC status would mean corporate tax at the fund level before anything reached you. No failure has been reported, and the concentration that caused the scare has since been diluted. Read the tax note in the next annual report before you treat the question as closed.

What you will actually owe

Because the fund has never paid a distribution, the only tax event for most holders to date has been a sale. Gains on shares held more than one year are long-term capital gains under IRC section 1222, taxed at 0%, 15% or 20% by income, plus the 3.8% net investment income tax under IRC section 1411 for higher earners. Shares held a year or less produce ordinary income. The 28% collectibles rate under IRC section 1(h)(4) does not apply; these are corporate shares, not collectibles.

If the fund ever realises large gains, by selling SpaceX stock after lockups for instance, a RIC must distribute them, and they arrive as capital gain dividends on a 1099-DIV whether or not you sold anything. To avoid a 4% excise tax the fund intends to declare at least 98% of net investment income and 98.2% of net realised capital gains each calendar year. Plan for a lumpy distribution year at some point; the fund says so itself.

IRAs

DXYZ is an ordinary listed security and can be held in a traditional IRA, a Roth IRA or a 401(k) brokerage window with no special custodian. There is no unrelated business taxable income problem, because a RIC’s dividends are not UBTI, unlike some partnership-structured vehicles. If you intend to own this at all, a Roth is the right home for a position whose thesis is a small chance of a very large capital gain.

This section starts with the risk that actually ends the investor, then gives the dated record.

The risk that ends you is the premium, not the portfolio

Investors in this category worry about issuer fraud, custody and platform insolvency. Those are the right worries for a marketplace. They are not the main risk here: the fund is a registered investment company with an independent custodian, an audited annual report and a board. The risk that has destroyed capital in DXYZ is arithmetic. People paid two to twenty times what the assets were worth. A buyer at the April 8, 2024 close of $99.79 has lost about 68% while the underlying assets rose more than sevenfold, and that recurs every time enthusiasm returns.

Concentration and single-name risk

SpaceX was 52.4% of the fund’s economic exposure at December 31, 2024, held through three SPVs. Anthropic was 14.4% of net assets at June 30, 2026, with OpenAI added on top in August 2026. The name on the tin says 100 companies; the count has never exceeded the 36 reported at March 31, 2026, and the top two positions have repeatedly been a quarter or more of net assets. If Anthropic’s next round prices down, the mark falls, the fee base falls, and the price usually falls further than the NAV.

Valuation risk

The adviser is the Valuation Designee. Marks are quarterly, published roughly two months in arrears, and built from rounds and comparables rather than trades. In a down market the lag cuts the other way: NAV looks stale and high while the price falls. Morningstar’s 2024 review found that of the holdings then disclosed, only SpaceX had gained since purchase, more than half were down over 30% and more than a third were down over 50% (Jack Shannon, Morningstar, April 18, 2024). The marks are not automatically generous. They are, however, the adviser’s.

The governance record

In January 2024, weeks before the NYSE listing, co-founder Samvit Ramadurgam sued Sohail Prasad, Destiny XYZ Inc. and two Destiny XYZ directors in the Delaware Court of Chancery. The fund disclosed the suit in a prospectus amendment. On July 23, 2026 the court issued its post-trial decision in Ramadurgam v. Destiny XYZ Inc., C.A. No. 2024-0057-PAF. The court found that the controlling stockholder breached his duty of loyalty and that the two directors he had appointed days before the vote acted in bad faith, in a scheme that used a reverse-forward stock split to cash out a roughly one-third holder weeks before the Tech100 listing. The remedy was a constructive trust restoring the equity interest Ramadurgam would have held, rather than rescission, plus an order that the individual defendants pay his attorneys’ fees under the bad-faith exception to the American Rule; the opinion describes the defendants’ pre-litigation conduct as “glaringly egregious and the product of unusually deplorable behavior” (Delaware Court of Chancery opinion, July 23, 2026; Gibson Dunn, 2026; National Law Review, 2026).

Read this precisely. The defendants were the adviser’s parent and its controller, not the fund, and the harmed party was a co-founder, not a DXYZ shareholder. No money left the fund. But the person who controls the adviser, sets the marks, runs the issuance programme and decides on buybacks carries a judicial finding of disloyalty in the entity that owns the adviser. That is a governance fact with a date and a docket number, and it belongs in your position sizing.

Public criticism from a direct competitor

In April 2024 ARK Investment Management publicly attacked the fund, with Cathie Wood telling Reuters that Destiny’s structure and fees meant investors faced a much higher price point in exchange for daily liquidity, and pointing to an estimated 4.98% fee from Destiny’s SEC filing (Bloomberg, April 18, 2024; Reuters, April 2024). ARK is not disinterested: it runs ARKVX, the interval fund competitor, which carries a 3.49% gross expense ratio capped at 2.90% by a contractual waiver. The criticism was true and it came from a rival. Morningstar and Axios made the same points that month without a competing product (Axios, April 10, 2024; Morningstar, April 18, 2024), which is why we weight those more heavily.

Regulatory and complaint record

We found no SEC or FINRA enforcement action against Destiny Tech100 Inc. or Destiny Advisors LLC, no state securities order and no securities class action against the fund, as of September 17, 2026. We found no meaningful BBB or Trustpilot complaint file either, which is expected: DXYZ is a listed security bought through third-party brokers, so there is no customer-service relationship to complain about. The recurring retail complaint in forum and social discussion is about the premium and about dilution from the at-the-market programme, and we treat it as unverified customer sentiment rather than evidence.

Dilution risk, in both directions

The fund can issue up to $1.0B of stock at the market whenever it trades above NAV. That dilutes nobody in NAV terms, since it is accretive, but it caps the premium by supplying stock into every rally, which is what holders experienced in 2026. When the stock falls below NAV the issuance stops by law, the fee base stops growing, and the only lever left is the buyback. You are short the premium whether you like it or not.

Buy it if you fit all of these:

• You want exposure to Anthropic, OpenAI, SpaceX or xAI in a brokerage account or a Roth IRA with no accreditation and no paperwork, and you accept that you are two layers away from the shares.
• You will buy only at or below net asset value, after doing the cash-adjusted arithmetic in the fees section.
• The position is small enough that a total loss changes nothing. We would cap it at 1% to 2% of liquid net worth and treat it as a lottery ticket with a 6.28% carrying cost.
• You can hold five years or more, and tolerate a stock that has fallen 68% from a peak while its assets rose sevenfold.

Skip it if any of these is true:

• You are buying at a premium because the underlying companies are exciting. The excitement is already in the price.
• You need income. There have been no distributions since inception and the fund says not to expect regular ones.
• You are accredited and writing $100,000 or more per name. Direct secondaries through a Schwab-owned Forge or through EquityZen cost a commission once, not 6.28% a year, and you own the actual shares.
• You want a diversified private-technology portfolio. Two names were roughly a quarter of net assets at June 30, 2026 and cash was most of the rest.
• Governance findings against a controller matter to you in absolute terms rather than as a discount you demand.

Five competitors and one plain liquid option. The relevant question is not which is cheapest, it is which layer of the private-markets problem you are willing to pay for: access, pricing or liquidity.

Table: Listed and unlisted routes into private technology, as of September 17, 2026

If you are accredited and large, buy the shares directly through Forge or EquityZen and pay a one-time commission. If you are not accredited, the listed funds are the only door, and the ranking among them depends on the price you can get relative to NAV on the day, not on the fee schedule. VCX charges 2.50% since its listing, up from 1.85% before it. RVI charges 2.00% with no carry and listed at a 1.2% premium, the smallest day-one dislocation of the three. ARKVX is the only one that lets you redeem at NAV, quarterly and capped, which is worth more than its higher fee to anyone who does not want to be a price-taker. DXYZ has the best tax form, the largest Anthropic weight and the worst history of investors overpaying for it. If your goal is exposure to AI economics rather than to a specific cap table, a Nasdaq-100 fund at 0.20% owns the companies buying most of the compute.

There is no Destiny account to open; the sequence is a brokerage sequence.

1. Use any US broker that offers NYSE-listed stocks. Commission is typically $0. A Roth IRA is the best home for the position if you have room: the payoff is skewed to a large capital gain and there is no UBTI issue.
2. Find the most recent NAV per share before you order. It comes from the quarterly press release and the semi-annual and annual reports, and it is up to two months stale by the time you read it.
3. Compute the premium or discount yourself against that NAV. Do not rely on a quote page; providers were measuring premiums against different NAV dates during 2026.
4. Subtract cash per share from NAV per share and from the market price, then compare the two remainders. That is what you are paying for the private book.
5. Use a limit order. Volume estimates for DXYZ ranged from about 172,000 to about 2.19 million shares in September 2026.
6. Size it at 1% to 2% of liquid net worth and write down the price at which you will sell.

Six documents to read before you wire anything: the latest quarterly results press release, for NAV and portfolio value; the latest Form N-CSR or N-CSRS, for the schedule of investments, the share count, the expense ratio and the section 851 tax note; the current Form 424B3, for the at-the-market programme and the shares actually sold; the Form N-2 fee table, for the management fee basis and the acquired-fund line; the valuation procedures in the notes, for who marks the book; and the Delaware Chancery opinion in C.A. No. 2024-0057-PAF, for the governance record of the people who control the adviser.

Destiny Tech100 solved a real problem and then let the solution get priced like a meme. A non-accredited investor with $5,000 had no legal route into Anthropic, SpaceX or OpenAI, and the marketplaces serving accredited buyers start at $5,000 to $100,000 per name with 45-day closes and no exit. A listed, audited, RIC-taxed closed-end fund with an independent custodian is a good answer. You get a 1099 instead of a K-1, IRA eligibility, daily liquidity, no subscription documents and no accreditation test. The structure also protects you in the way that matters most in this sector. Linqto customers discovered they held interests in the platform’s own vehicles rather than title to their shares, filed Chapter 11 on July 8, 2025, and waited until the plan was confirmed on February 6, 2026 for a targeted recovery of about 95% of fair market value. A DXYZ holder owns a share of a registered fund whose assets sit at U.S. Bank.

What ruins it is the price and the fee, in that order. Estimated total expenses of 6.28% of net assets would be disqualifying on their own for most portfolios, and the 2.50% management fee is charged on gross assets, so it applies in full to the money market fund that was 57.5% of net assets at June 30, 2026. Then comes the premium, which has decided outcomes: buyers at the April 8, 2024 close of $99.79 are down about 68% while net asset value per share rose from $4.84 to $34.30. And the NAV series that looks so good is not a return series: 85% of 2025’s NAV growth and 63% of the second quarter of 2026’s came from selling stock above NAV rather than from investing. Layer on a Delaware Chancery finding of disloyalty and bad faith against the person who controls the adviser, sets the marks and runs the issuance programme, and 2.5 out of 5 is the honest number: the wrapper is sound, the fee is punitive, the governance record is poor, and almost everyone who has owned it has paid too much.

We would move the rating to 3.5 if three things happened together: total expenses fell below 4% of net assets, cash fell below 25% and stayed there, and the fund began publishing a quarterly decomposition of NAV change into investment return and issuance accretion. To 4 if it added a tender or repurchase mechanism letting holders exit at a stated fraction of NAV. To 2 if the premium returned above 50% and the fund used it to issue another $500M, or if cash rose back above 50% of net assets with the full fee attached. To 1.5 on any restatement of a prior NAV, any SEC action against the adviser, or a change of auditor or valuation agent without an explanation.

What to watch, with dates. The third-quarter 2026 results, due around late November 2026, for NAV against $34.30, cash against 57.5% of net assets, and whether the share count moved above 47,657,338, which would mean the at-the-market programme restarted and therefore that the stock traded above NAV. The next Form N-CSR, for the audited expense ratio against 6.28%, the holdings count against the 100 in the fund’s name, and the section 851 tax note. The buyback: whether the fund repurchased stock under the August 2026 authorisation and at what average discount. Anthropic’s next primary round against the $965B May 2026 level, since that one position was 14.4% of net assets. And the SpaceX lockup calendar into 2027, because the fund’s oldest and most successful position is now a public stock whose eventual sale would trigger the first real distribution in its history.

Nothing here is investment advice; it is our reading of the public record as of September 17, 2026.