The Continuation Fund Shell Game: Dragoneer's $2B Bet on OpenAI and SpaceX

The mechanics are elegant. Dragoneer Investment Group, a San Francisco-based growth equity firm, is raising a $2 billion continuation fund using its existing stakes in OpenAI and SpaceX as the underlying assets. The pitch to limited partners is simple: acquire shares in two of the most sought-after private companies on Earth without the hassle of sourcing them yourself. The structure, however, deserves closer scrutiny. Continuation funds are not new. But the specific combination of assets here—OpenAI and SpaceX—creates a unique set of failure modes that institutional investors should map before committing capital. Let me establish the context. Dragoneer has been a quiet but consistent presence in the tech investing landscape for over a decade. Its portfolio includes Stripe, ByteDance, Robinhood, and Nvidia. The firm operates as a financial investor, not a strategic one. It does not seek board seats or operational influence. This positioning has historically given it access to founders who prefer passive capital. The continuation fund model, popularized by firms like Carlyle and Hellman & Friedman, allows a GP to transfer assets from a maturing fund into a new vehicle, effectively resetting the clock on management fees and carried interest. The old fund's LPs get liquidity; the new fund's LPs get exposure to proven assets. Dragoneer's version of this play is notable for two reasons: the quality of the assets and the reported "minimal discount" at which they will be transferred. Here is where the structural analysis begins. The core tension in any continuation fund is valuation. The GP sits on both sides of the trade—selling for the old fund, buying for the new fund. This is an inherent conflict of interest. Dragoneer's "minimal discount" strategy signals confidence in the current mark-to-market of OpenAI and SpaceX shares. But it also means the new fund's LPs are entering at a high NAV with compressed upside. Based on my experience auditing private market transactions, I can tell you that Level 3 assets—which is what unlisted OpenAI and SpaceX shares are—carry significant model risk. The valuation methodologies for these companies rely on forward-looking assumptions about AGI commercialization and Starship launch cadence. These are not stable inputs. A single missed technical milestone could trigger a 30% repricing. The "minimal discount" approach removes the margin of safety that typically protects new entrants in secondary transactions. The second structural issue is concentration. A $2 billion fund anchored by two assets means each position likely represents 40-60% of the portfolio. This is not diversification; it is a leveraged bet on two companies. For institutional LPs with existing AI or aerospace exposure, this fund amplifies hidden correlation. The fund documents will disclose this, but the disclosure does not mitigate the risk. The third issue is liquidity. Continuation funds typically have 5-8 year lockups. The underlying assets are highly illiquid. If OpenAI's valuation contracts due to regulatory pressure on AI exports—a real possibility given the current geopolitical climate—LPs have no exit. They are locked into a declining NAV with no mechanism to reduce exposure. This is the structural weakness of the model: it converts liquid wealth into illiquid promises. Now, the contrarian angle. The bulls have a point. The "minimal discount" strategy is not merely a GP-friendly maneuver. It reflects genuine scarcity. OpenAI and SpaceX shares are in high demand and low supply. Secondary market platforms like Forge Global and EquityZen consistently show buyers willing to pay premiums of 10-30% over the latest primary round valuations. Dragoneer's access to these shares at any price is a testament to its relationship network. For LPs, the fund offers a curated entry into two companies that are otherwise nearly impossible to access. The brand value alone—being able to claim ownership in OpenAI and SpaceX—carries weight with sovereign wealth funds and family offices. The continuation fund structure also addresses a real market inefficiency: the traditional 10-year fund lifecycle does not fit the 15-20 year growth trajectory of companies like SpaceX. The model is a rational institutional response to a duration mismatch. But here is the accountability call. The "minimal discount" framing obscures a critical question: who bears the risk if the assets are overvalued at transfer? The answer is the new fund's LPs. They are paying top dollar for assets that may face headwinds. The SEC has been increasing scrutiny on continuation fund fee disclosures and conflict of interest management. Dragoneer should expect regulatory attention. The CFIUS angle is also relevant—foreign LPs, particularly from the Middle East, may face additional scrutiny given the strategic nature of OpenAI and SpaceX. The fund's success will depend on its ability to navigate these regulatory currents while maintaining the trust of both old and new LPs. The next 12-24 months will reveal whether the "minimal discount" was a fair price or a structural flaw. The market will vote with its capital. Code is law until it isn't. The same applies to fund structures. s heart. The question is not whether Dragoneer can raise the $2 billion. It can. The question is whether the LPs who commit will understand the risk profile they are inheriting. The answer, based on the current structure, is probably not.