The Eternal Fund: How Dragoneer's OpenAI and SpaceX Continuation Vehicle Exposes the Friction Between Private Illiquidity and Public Trust

The machine that manufactures our future is now packaging itself for resale. Dragoneer Investment Group, a San Francisco-based growth equity firm with a portfolio that reads like a who's who of twenty-first-century private dominance, is reportedly spinning up a new continuation fund using its existing stakes in OpenAI and SpaceX, with a target of $2 billion. The most telling phrase in the entire brief is the casual mention of a “minimal discount” on the transferred assets. That is not a detail; it is a confession. It says the firm believes those shares are worth nearly their last private valuation, and it believes the new LPs will accept that without flinching. The ledger does not sleep, it only waits. And this time, it's watching a high-wire act between GP incentives and the cold arithmetic of locked capital. Context is everything, and the context here is a private equity secondaries market that has grown from about $40 billion in annual volume in 2015 to over $100 billion today, with continuation funds now accounting for nearly a third of that activity. Dragoneer is not an outsider; it has been a professional resident in the cult of the tech unicorn for years—with positions in Stripe, ByteDance, Robinhood, and NVIDIA—long before OpenAI and SpaceX became household names. This new fund is not an act of innovation. It is an act of engineering, a deliberate attempt to solve a structural mismatch that the crypto world has been ranting about for a decade: private assets have a ten-year fund lifecycle, but truly transformative companies need twenty years to mature. Instead of letting the clock expire, Dragoneer is simply resetting it. The fund itself is a sort of perpetual motion machine, a financial mechanism that turns a maturing portfolio into a fresh starting line. The irony is not lost that this news surfaces on Crypto Briefing, a publication oriented toward the very decentralized alternatives that this concentrated, opaque, and perfectly frictioned vehicle now stands in opposition to. The core of this story is not that Dragoneer wants to own more of two brilliant companies. It is that the structure itself—the continuation fund, the minimal discount, the $2 billion price tag—exposes a deeper friction between the way private capital claims to value assets and the way that value is actually realized. As someone who spent hundreds of hours backtesting Ethereum liquidity pools against treasury yields during DeFi Summer, I can spot an artificially extended yield when I see one. This is the same game, played with higher legal fees. Let me break down the mechanics, because the details are as instructive as they are cynical. A continuation fund works like this: the General Partner (GP)—here, Dragoneer—creates a new fund vehicle. It then sells the most attractive assets from the old fund (the one approaching its mandated end-of-life) into the new one. The existing LPs are given a choice: they can "roll" their interest into the new fund, receiving shares in the continuation vehicle, or they can take cash, effectively cashing out at the current valuation. The GP, meanwhile, continues to manage the assets, collects management fees on the new fund (typically 1.5-2% of committed capital), and—crucially—starts a fresh carry clock. Any future appreciation on those now-transferred assets will generate a 20% performance fee for the GP, even though that appreciation is a continuation of gains that were already building in the old fund. The GP has found a way to resell the same growth story twice. The "minimal discount" is the price of admission for this trick: Dragoneer will value the assets at or near their last financing round, regardless of the fact that those financing rounds were cherry-picked moments of exuberance. It is a way to offer existing LPs a face-saving exit while tempting new LPs with the prestige of owning a piece of OpenAI and SpaceX—rarefied air that most institutional investors would kill to breathe. The problem is that the entire valuation edifice rests on a foundation of smoke. OpenAI and SpaceX are not publicly traded. Their values are determined by negotiated private rounds, secondary markets on platforms like Forge Global and EquityZen, and the feverish expectations of a market that has decided artificial intelligence and space exploration are the only two growth narratives left on Earth. OpenAI was valued at $157 billion after its October 2024 funding round; SpaceX reached $350 billion by December of that year. These numbers are not the result of continuous price discovery; they are the outputs of a handful of large checks written by VCs and sovereign funds who are themselves betting on future milestones. In my 2022 stablecoin audit, I found a $50 million discrepancy in a proof-of-reserves report that was supposed to be bulletproof. I have learned to distrust any ledger that is not continuously someone's direct, liable ledger. These companies' cap tables are the private ledgers, and they are only updated when a round closes. In between, the billionaire's guess is the only mark. That is not a market; it is an argument. The "minimal discount" makes that argument even more fragile. In classic continuation fund practice, a discount is applied to reflect the illiquidity of the assets, the cost of carrying them, and the risk of no real exit. A 5% to 15% markdown is common. Dragoneer is reportedly offering something much narrower, perhaps signifying that it has enough confidence in its own story—or enough power over its counterparties—to dictate nearly par value. That is a dangerous signal. A minimal discount means the new LPs are paying nearly the last round's price, not a price that embeds a cushion for the 15-20 years of uncertainty ahead. It means if OpenAI's valuation corrects by a third due to a regulatory crackdown on AI, or if SpaceX faces a string of Starship failures, the fund's Net Asset Value (NAV) will bleed, and there will be no liquidity to stanch the wound. Liquidity is a ghost; solvency is the body. These assets might be solvent, but they are not liquid. A continuation fund is an exercise in locking up that illiquidity for yet another decade. Now, let's talk about the elephant in the room: what this means for the crypto world. On its face, this is a traditional finance story, far removed from blockchain news. But the mere fact that Crypto Briefing is covering it—and that a growing number of crypto-native family offices and fund managers are sniffing around PE secondary exposure—signals a convergence that has been brewing since the collapse of Silicon Valley Bank and the rise of real-world asset (RWA) tokenization. The report suggests that Dragoneer may, eventually, consider tokenizing this fund's shares, offering security tokens to a wider audience of accredited crypto investors. That would be the revenge of the on-chain ledger. But do not hold your breath. The SEC's current stance on tokenized securities is a labyrinth of byzantine complexity, and the recent court battles over the Private Fund Rules show that even basic disclosure enhancements are being fought tooth and nail. The probability of a fully tokenized continuation fund for OpenAI/SpaceX shares within two years is low—call it 10% —but the interest in it is a tell. The LPs who are most attracted to these "status assets" are the same investors who have become comfortable with digital assets. They want the same liquidity, transparency, and 24/7 market that Bitcoin gave them, but they are willing to accept the opacity of a private fund because the returns are, on paper, astronomical. This is where my macro-liquidity lens has to zoom out. The Fed and global central banks are in a synchronized race to cut rates, with some terminal rates already being priced in for 2026. Money market funds are yielding 4-5% today, but that will slide as the easing cycle progresses. When the opportunity cost of parking cash drops from 5% to 3%, the relative attractiveness of holding illiquid private assets rises. This is precisely the environment in which a $2 billion continuation fund like Dragoneer's can find buyers. Institutional LPs are redeploying from cash into longer-dated, higher-yield assets. It is a time when "risk appetite" is a beautiful, fragile thing. So, Dragoneer's fund is not an anomaly; it is a predictable response to the end of zero-interest-rate policy and the beginning of a new carry chase. However, this chase is not without a dark side. As rates fall, the pressure to deliver outsized returns increases, which means more continuation funds, more aggressive valuations, and more LPs being asked to roll over their existing commitments rather than receive a check they can spend. There is a blind spot here, one that crypto investors should recognize better than anyone. The entire promise of blockchain was to eliminate the need for trust by imposing verification. Continuation funds are the opposite: they are a monument to unverifiable trust. The GP controls the information; the LP is forced to rely on the GP's fairness; the valuation is a negotiation between the GP's side and itself. The report mentions that the biggest risk is valuation fairness and conflict of interest. How can there be fair value when the seller and buyer are both the same person? Dragoneer will have to disclose the methodology, but in private markets, methodology is a creative writing exercise. In my experience auditing stablecoin reserves, I found that the most damning evidence wasn't in the reports themselves—it was in what they omitted. The same will apply here. LPs will see a summary valuation report from "independent" third parties that were paid by Dragoneer, but they will not see the cap table negotiation detail, the term sheet concessions, or the side letters that affect the true economics. This is not cynicism; it is the structure of the instrument. What makes this even more contrarian is that crypto's answer to this problem is staring everyone in the face. We already have the technology to tokenize private company shares, create continuous secondary markets, and let the public (within regulatory bounds) trade them. The infrastructure is there. What we lack is regulatory approval from the SEC, which has historically been hostile to such innovation. If the SEC ever allowed a compliant security token for a company like SpaceX, priced on a decentralized exchange, it would render the entire continuation fund model obsolete. Why would an investor accept a 10-year lockup and a minimal discount when she could buy a tokenized share on a public market at a real-time price, with fractional ownership and instant settlement? The answer is she would not. The existence of these continuation funds is a symptom of the failure of traditional securities law to adapt to digital liquidity. They are a band-aid on a wound that crypto amputation would clean. But let's be pragmatic. The SEC is unlikely to approve such a token before this fund launches. So, in the immediate term, Dragoneer's fund will likely gather its $2 billion, buy those OpenAI and SpaceX shares, and let its new LPs sit on a thinly traded, concentrated position that is materially undiversified. The concentration risk is severe: the report correctly notes that within a single fund, OpenAI and SpaceX might represent 40-60% of the portfolio. If you are an LP already holding Amazon, Microsoft, and a few AI ETFs, adding a fund that is 60% OpenAI and 20% SpaceX is not diversification; it is doubling-down on the same ten stocks story. In fact, I would argue that the only legitimate diversification benefit comes from the fact that these are pre-IPO companies with a possible 100x upside, but that is not a portfolio construction argument—it is a lottery ticket wrapped in a legal document. The macro environment will determine if this bet pays off. We have an AI arms race accelerating, with policy support from the US government (despite export control talks), and we have a space economy that is still in its primordial stage. SpaceX's Starship success or failure directly impacts its valuation, while OpenAI's future depends on AGI milestones and the corresponding regulatory frameworks. These are binary, fat-tailed events. A continuation fund that holds these assets for 10 years is essentially a long-dated call option on two outcomes: either these companies become the standard utilities of an upgraded civilization, or they stall, and the fund implodes. The "minimal discount" tells you the GP believes the call is in the money. History suggests that when everyone expects those extraordinary results, the exit liquidity is often already gone. Plus, consider the actual exit paths. Even with a continuation fund, the only way for the new LPs to realize returns is if Dragoneer eventually takes these companies public, or finds another secondary buyer. For OpenAI, that would be the largest IPO in history, a prospect that is still distant. For SpaceX, it is a similar story. In the meantime, LPs are asked to watch the NAV fluctuate based on quarterly marks that the GP itself sets. This is not an income-producing asset; it is a belief engine. What does this mean for the crypto market? It means capital is being siphoned from liquid crypto assets into illiquid private instruments. If a sovereign fund or family office has $500 million to allocate to a "technology future" bucket, and it puts $200 million into Dragoneer's continuation fund, that is $200 million that will not flow into Bitcoin or Ethereum or any token. The narrative of crypto as the only shortcut to exposure to AI and space is being challenged by traditional PE's ability to offer direct exposure to the actual companies. As rates continue to fall, this effect will intensify. But there is an equal and opposite force: as institutional LPs become more aware of the opacity and lockup risk, some may turn to crypto assets for true liquidity. The result is a two-way street. For every Dragoneer, there is a pension fund questioning why it cannot hold a diversified basket of AI tokens instead of a single private share with zero daily price transparency. The crypto ecosystem offers continuous pricing, permissionless access, and the ability to exit in seven days. That is an eternal advantage. The continuation fund, by contrast, is a cage designed to hold the bird while it sings. The cage is beautiful, but the bird is still trapped. In my time monitoring the Vietnamese CBDC pilot, I learned that the gap between institutional custodianship and individual autonomy is not a technical gap; it is a philosophical one. The technological revolution that allegedly wanted to democratize finance has given birth to a new aristocracy of lockups, carry recalibrations, and "minimal discounts" reserved for the accredited few. It is a hemorrhage of trust, but not the silent kind. It is loud, structured, and paid for by management fees. The contrarian angle is not that Dragoneer's fund will fail. It might well make its LPs rich—OpenAI and SpaceX are that good. The contrarian angle is that the very existence of this fund is a confession that the traditional private equity model cannot deliver value without resorting to creative financial engineering that is often adversarial to LPs. It is a system that rewards the GP twice for the same work. And while it may be the smartest game in town—and indeed a not-unreasonable allocation for a sophisticated investor—it is a far cry from the transparent, automated, trustless markets that the crypto originalists once promised. The fact that a crypto outlet is reporting on this with a hint of admiration suggests we are witnessing the denouement of the decentralization dream: the traditional world is absorbing crypto's best ideas (liquidity, access, fractional ownership) without adopting its ethos (openness, auditability, self-custody). The new fund will not be on-chain. It will be in the Cayman Islands. It will not have a smart contract; it will have a 300-page LPA. And it will still be called "innovation." The "minimum discount" is the key tell. A true market sets its own discount. A true ledger does not ask permission. The attempt to find a price without an actual market is exactly the type of friction that blockchain was invented to eliminate. So here is my takeaway: do not be envious of this fund. Be cautious. The search for yield has pushed the most sophisticated investors into accepting terms that would be laughed at in a public market. In the next 12 to 24 months, watch how these continuation funds handle the inevitable volatility in AI and space valuations. When a company like OpenAI hits a regulatory wall or SpaceX experiences a public failure, the "minimal discount" will become a maximal loss. And the only sound you will hear is the silent hemorrhage of algorithmic trust as the LPs realize that their liquidity was never there. I would rather hold a tokenized fraction of intelligently managed open-source AI models than a share in a private fund that trades at a price I cannot see. The ledger does not sleep, but it also does not lie. Dragoneer's fund is one of the most sophisticated pieces of financial architecture we have seen in this cycle. That does not mean it is worthy of trust. It means it is worthy of scrutiny. Keep your eyes open, and keep your capital closer to the light. In the end, the question is not whether OpenAI and SpaceX will be worth more. They will be. The question is whether the vehicle designed to profit from them is serving the GP's interest or, eventually, the LPs' interest. As with any complex derivative, the right answer is often found in the footnotes. I would like to read the footnotes. Until then, I will watch from the sidelines. After all, in a world where the asset itself is a story, the fund is merely the cage. And the bird—the future—will fly anyway. The only difference is who gets to watch from where.