The $2 Billion Illusion: Dragoneer's Continuation Fund and the Liquidity Paradox of AI's Crown Jewels

What if the most coveted assets in private markets are also the most dangerous ones to hold? That's the question Dragoneer Investment Group just forced the market to confront. The San Francisco-based firm is reportedly raising a $2 billion continuation fund backed by its stakes in OpenAI and SpaceX. The pitch, according to the initial report, emphasizes a "minimal discount" on the transfer price. Let me translate that from fund-speak into plain English: Dragoneer believes its existing limited partners should accept near-par pricing for assets that have no public market, no objective valuation, and no exit mechanism other than the goodwill of a handful of secondary buyers. Tracing the fault lines before the quake hits — this is where the structural tension begins. I've spent the last decade watching private market structures evolve from simple venture funds into increasingly complex vehicles designed to solve one problem: the mismatch between a 10-year fund lifecycle and the 20-year maturation curve of modern technology companies. OpenAI and SpaceX are not your father's portfolio companies. They're capital-intensive, technology-driven, and — critically — they're taking far longer to reach their terminal value than traditional fund structures can accommodate. The continuation fund is the industry's answer to this temporal arbitrage. Instead of forcing a sale at year 10, the GP transfers the asset into a new vehicle, resets the clock, and gives both existing and new LPs a fresh 5-8 year window. It's elegant in theory. In practice, it's a minefield of conflicting incentives, valuation gymnastics, and what I'd call "liquidity theater" — the appearance of exit optionality where none truly exists. Dragoneer's position is uniquely interesting because of what it holds. OpenAI, post-restructuring, is arguably the most closely-watched private company on Earth. Its valuation trajectory — from $29 billion in early 2023 to $157 billion in late 2024 — has been nothing short of parabolic. SpaceX, meanwhile, has quietly become the most valuable private company in the world at approximately $350 billion, driven by Starlink's revenue generation and the Starship program's technical milestones. These are not ordinary assets. They're strategic national infrastructure disguised as venture investments. And that's precisely where the analytical framework breaks down. Let me walk through the valuation mechanics, because this is where the "minimal discount" language becomes most revealing. In a standard continuation fund transaction, the GP typically offers existing LPs a choice: roll into the new vehicle at a negotiated price, or take cash at a discount to that price. The discount serves as compensation for the liquidity premium the GP is extracting — the GP gets to keep managing the asset, and the LP gets to exit without waiting for an IPO that may never come. Dragoneer's "minimal discount" approach signals something important: they believe the assets are so scarce, so in-demand, that they don't need to offer traditional compensation for the liquidity lock-up. And they might be right. The secondary market for OpenAI and SpaceX shares has been running hot for years, with buyers consistently outnumbering sellers. Forge Global and EquityZen have built entire platforms around facilitating these trades, and the premiums over primary market valuations have routinely hit 10-30%. But here's the problem with that logic: it assumes the current pricing environment is rational. It assumes that the AI hype cycle — which has driven valuations to levels that would make 2021's crypto mania look conservative — will persist indefinitely. It assumes that SpaceX's Starship program will hit its milestones on schedule, that Starlink's subscriber growth will continue unabated, and that no regulatory body will step in to constrain either company's trajectory. Code never lies, but it does omit. And what the valuation models omit is the tail risk. Let me be specific about what I mean. I audited failed ICO tokens back in 2018, and I saw the same pattern then that I see now: a narrative-driven market that rewards optimism over analysis, where the fundamental question — "what is this asset actually worth?" — gets buried under a mountain of momentum and FOMO. The names have changed, but the mechanics haven't. In the crypto winter of 2018, I watched projects with real technology and real teams collapse because their tokenomics were structurally unsound. The vesting schedules were misaligned, the incentive structures were broken, and the founders had no path to sustainability. The market didn't care about fundamentals during the bull run. It cared about narrative. And when the narrative shifted, everything fell apart. I'm not saying OpenAI and SpaceX are ICO scams. Far from it. These are real companies with real revenue, real technology, and real competitive moats. But the valuation question remains: at what price does the risk become unacceptable? Consider the concentration risk. A $2 billion fund backed primarily by two assets means that each position likely represents 40-60% of the fund's NAV. That's not diversification; that's a leveraged bet on two companies. For institutional LPs with strict mandates around portfolio construction, this should be a red flag. For family offices and sovereign wealth funds chasing the "status good" of owning OpenAI shares, it's apparently a feature, not a bug. The narrative shifts, but the leverage remains. And in this case, the leverage is structural rather than financial. The fund's performance will be determined almost entirely by the trajectory of two companies, neither of which has any obligation to maximize shareholder value in the traditional sense. OpenAI's governance structure — even post-restructuring — retains elements of its nonprofit origins. SpaceX is controlled by a founder with a 20-year vision that may not align with LP expectations for quarterly mark-to-market performance. This brings me to the deeper issue: the continuation fund model itself. The industry has embraced this structure with remarkable speed — Preqin data suggests continuation funds now represent over 30% of secondary market volume, up from less than 10% a decade ago. But the model has a fundamental flaw that few are willing to acknowledge: it allows GPs to avoid the hard decision of actually selling assets. In the traditional model, a fund has a 10-year life. At year 10, the GP must sell the assets, return the capital, and face the market's judgment. This discipline forces honest pricing. It forces the GP to confront whether the asset is actually worth what they think it is. The continuation fund removes this discipline. It allows the GP to say, "The asset is too valuable to sell now. Let's extend the timeline." And because the GP controls the information flow, the LPs have limited ability to challenge this assessment. I've seen this dynamic play out in crypto markets repeatedly. Projects that should have died in the 2018 bear market were kept alive through token swaps, rebranding, and new narratives. The underlying technology never improved, but the story did. And each time, the insiders — the founders, the early investors, the VCs — found ways to extract value while the retail holders absorbed the losses. I'm not suggesting Dragoneer is running a scam. But the structural incentives are worth examining. The GP earns management fees on the new fund. The GP earns carry on the new fund's performance. The GP gets to maintain its relationship with OpenAI and SpaceX — relationships that have enormous strategic value beyond the financial returns. The existing LPs get a choice between rolling over or taking cash at a "minimal discount." The new LPs get access to assets they couldn't otherwise access. Everyone wins on paper. But the question is: what happens when the music stops? Let me run through a scenario. Suppose OpenAI's next model release underperforms expectations. Suppose the AGI timeline slips by 18 months. Suppose a major regulatory action in the EU or the US constrains the company's ability to monetize its technology. The valuation — currently based on aggressive assumptions about future revenue growth — would need to be revised downward. But here's the catch: in a continuation fund, there's no public market to force the mark. The GP can maintain the valuation, citing "long-term fundamentals," while the LPs have no mechanism to challenge it. This is the liquidity paradox at the heart of the model. The fund promises liquidity — the ability to exit an illiquid asset — but it actually delivers the opposite. It locks LPs into a vehicle with no exit mechanism other than the GP's willingness to facilitate a secondary sale. And in a downturn, that willingness evaporates. I've been through enough market cycles to know that liquidity is just patience disguised as capital. The funds that survive are the ones that can wait out the storm. The funds that fail are the ones that need to sell at the worst possible moment. Dragoneer's continuation fund is designed to be the former. But the design only works if the underlying assets maintain their value. And that's a bet, not a certainty. Let me also address the regulatory dimension, because it's more significant than most observers realize. OpenAI and SpaceX are not ordinary portfolio companies. They're strategic assets in the US technology ecosystem. OpenAI is at the center of the AI competition with China. SpaceX is integral to US national security and space policy. When Dragoneer raises a fund backed by these assets, it's not just a financial transaction — it's a statement about who gets to own a piece of America's technological future. The Committee on Foreign Investment in the United States (CFIUS) has been increasingly active in reviewing transactions involving sensitive technology assets. If Dragoneer's new fund attracts capital from sovereign wealth funds in the Middle East or Asia — which is likely, given the "status good" appeal of these assets — the transaction could trigger CFIUS review. Even passive financial investments can raise national security concerns when the underlying assets are this sensitive. This creates an interesting tension. Dragoneer wants to raise $2 billion, and the most likely sources of that capital are institutional investors who want exposure to AI and space technology. But the more attractive the fund becomes to international investors, the more regulatory scrutiny it attracts. The fund's success could become its own obstacle. There's also the SEC angle. The agency has been scrutinizing private fund practices under Chair Gensler, with particular attention to fee disclosures and conflicts of interest. Continuation funds are a natural target for this scrutiny because they involve the GP transacting with itself — selling assets from one fund it manages to another fund it manages. The potential for self-dealing is obvious, and the SEC has signaled that it's watching. Dragoneer's "minimal discount" strategy is, in part, a response to this scrutiny. By pricing the assets at near-market levels, the firm can argue that it's treating its LPs fairly. But this argument cuts both ways. If the discount is too small, new LPs are overpaying. If it's too large, existing LPs are being shortchanged. The "minimal discount" is a political statement as much as a financial one — a signal to regulators that Dragoneer is playing by the rules. But here's what the regulatory framework doesn't capture: the information asymmetry. Dragoneer knows more about OpenAI and SpaceX than any LP possibly could. It has board seats, or at least observer rights. It has access to management. It has visibility into the companies' strategic plans. When it sets the transfer price for the continuation fund, it's using information that the LPs don't have. This is not illegal — it's the nature of private markets. But it's worth acknowledging that the "minimal discount" is not a market price. It's a negotiated price, set by a party with a vested interest in the outcome. Let me step back and think about what this means for the broader market. Dragoneer's continuation fund is not an isolated event. It's part of a trend that's reshaping the private equity landscape. As more GPs embrace continuation funds, the traditional 10-year fund lifecycle is becoming obsolete. Assets are being held longer, valuations are being maintained at higher levels, and the discipline of forced selling is being removed from the system. This has implications for the crypto market as well. The same dynamics that drive continuation funds — the desire to hold assets longer, the belief that current valuations are justified, the willingness to defer liquidity — are present in crypto markets. The difference is that crypto has no GP to control the narrative. The market is the market, and prices adjust in real-time. I've often said that chaos is the only constant variable. In crypto, that chaos is visible — it's in the volatility charts, the liquidation cascades, the protocol failures. In private markets, the chaos is hidden. It's in the valuation models that no one can verify, the fund structures that no one can fully understand, and the conflicts of interest that no one wants to discuss. Dragoneer's continuation fund is a bet on the continued growth of AI and space technology. It's a bet that OpenAI will continue to dominate the AI landscape, that SpaceX will continue to push the boundaries of space exploration, and that the current valuation environment will persist. These are reasonable bets. But they're not certainties. What happens if OpenAI's next model fails to impress? What happens if Starship's next launch ends in disaster? What happens if the AI bubble — and I use that word deliberately — finally bursts? The LPs in Dragoneer's continuation fund will have no exit. They'll be locked in, watching their NAV decline, with no mechanism to sell and no one to blame but themselves for not reading the fine print. I'm not saying this will happen. I'm saying it could. And the "minimal discount" that Dragoneer is offering doesn't compensate for that risk. It's a signal of confidence, not a hedge against uncertainty. Let me offer a different perspective. Maybe the "minimal discount" is actually the right call. Maybe OpenAI and SpaceX are so uniquely positioned that any discount would be leaving money on the table. Maybe the continuation fund is the only way to capture the full value of these assets, and the LPs who roll over will be rewarded handsomely for their patience. This is the bull case, and it's not without merit. OpenAI's technology is genuinely transformative. SpaceX's achievements are historically significant. Both companies have demonstrated an ability to execute on ambitious visions. If they continue on their current trajectories, the current valuations will look cheap in hindsight. But here's the thing about hindsight: it's always 20/20. The challenge is making decisions in the present, with incomplete information, under conditions of uncertainty. And the present looks a lot like 2021, when everyone was convinced that crypto was the future, that DeFi would replace traditional finance, and that the only risk was not being invested enough. We all know how that turned out. I'm not predicting a crash. I'm not saying Dragoneer's fund will fail. I'm saying that the structural risks are real, that the "minimal discount" doesn't adequately compensate for those risks, and that the continuation fund model — while innovative — has yet to be tested in a severe downturn. Arbitrage is the market's way of correcting itself. The arbitrage here is between the narrative of AI and space exceptionalism and the reality of valuation uncertainty. Eventually, the market will correct this arbitrage. The question is whether Dragoneer's LPs will be on the right side of that correction. Let me also consider the competitive dynamics. Dragoneer is not the only firm pursuing this strategy. Blackstone, KKR, and Carlyle have all been active in the continuation fund space. Ardian and Lexington Partners have built entire businesses around secondary transactions. The market is becoming crowded, and the competition for quality assets is intensifying. This competition has a predictable effect: it drives up prices. As more GPs compete for the same assets, the discounts shrink, and the valuations rise. Dragoneer's "minimal discount" is partly a reflection of this competitive pressure — the firm can't offer a large discount because other buyers would snap up the assets at a higher price. But this dynamic also creates a bubble. When everyone is competing for the same assets, the prices become detached from fundamentals. The "minimal discount" becomes the norm, and the risk-adjusted returns become increasingly unattractive. The LPs who enter at these prices are betting on continued appreciation, not on value creation. I've seen this movie before. In 2017, ICOs were the hottest asset class. Everyone wanted in, and the prices reflected the demand, not the fundamentals. When the music stopped, the losses were catastrophic. The same dynamics are at play in the private markets today, just with better suits and more sophisticated language. Collapse is a feature, not a bug. Markets need to correct. They need to purge the excesses and reset the expectations. The continuation fund model is not immune to this process. When the correction comes — and it will come — the funds that are most exposed to overvalued assets will suffer the most. Dragoneer's fund is particularly exposed because of its concentration. Two assets, two bets, two companies that are at the center of the most hyped narratives in the market. If either company stumbles, the fund's performance will suffer. If both stumble, the fund could be wiped out. This is not a prediction. It's a risk assessment. And it's the kind of assessment that LPs should be doing before they commit capital to a continuation fund with a "minimal discount." Let me end with a question: what is the actual value of OpenAI and SpaceX? Not the narrative value, not the hype value, but the fundamental value based on discounted cash flows, competitive positioning, and realistic growth assumptions? If you can't answer that question with confidence, you shouldn't be investing in a fund that's concentrated in these assets. Reading the silence between the block heights — that's what I do. I look for the gaps in the narrative, the places where the story doesn't quite add up. And in Dragoneer's continuation fund, I see a gap. The "minimal discount" is a signal of confidence, but it's also a signal of risk. The LPs who roll over are accepting that risk without adequate compensation. Maybe they'll be rewarded. Maybe OpenAI and SpaceX will continue to grow at their current pace, and the "minimal discount" will look like a bargain. But maybe not. And the asymmetry of the outcome — the potential for significant upside versus the potential for significant downside — is not adequately reflected in the fund's structure. I'll be watching this fund closely. Not because I think it will fail, but because it represents a test case for the continuation fund model. If Dragoneer succeeds, we'll see more funds like it. If it fails, we'll see a reckoning. Either way, the market will learn something. And that's the only certainty in this business. The narrative shifts, but the leverage remains. And in this case, the leverage is the belief that AI and space technology are the future. That belief may be correct. But the price of that belief — the "minimal discount" — may be too high. Time will tell. It always does.