Early-stage startups need to raise small amounts of money quickly, cheaply, and repeatedly, often when they have very little to show to prospective investors. At this stage, it is far too early to negotiate the terms traditional venture deals are built around: how much the startup is worth, how control should be allocated between founders and investors, and what downside protections investors should have.
Until about a decade ago, most founders and investors maneuvered around this problem with the convertible note: an instrument that looked like a loan but was never meant to be one. Investors would write checks that technically accrued interest and came due on some maturity date, while everyone knew that the fledgling startup would never be able to repay it in cash. The “loan” was really an advance payment for preferred stock that the parties hoped would be when the startup raised “real” financing. The deal was structured to allow the founders and investors to defer negotiations over stock price and investor rights so that early financings could close quickly and cheaply. However, the debt-based structure proved complex to operate and created undesirable friction where startups were not ready to issue stock when the note matured.
In 2013, Y Combinator, one of Silicon Valley’s most successful and influential tech accelerators, created a substitute for convertible notes, an instrument that was essentially a convertible note stripped of its debt-like mechanics. The instrument was still based on providing the investment amount first and converting it into preferred stock in a future financing round. However, unlike the note, it had no maturity date by which it had to convert or be repaid, and it did not accrue interest. The new instrument was branded as the “Simple Agreement for Future Equity,” or “SAFE.” And the SAFE did exactly what every Y Combinator startup hopes its own product will do: It gained market share fast, won over founders and investors alike, and became the undeniable market standard for very-early-stage financing.
As SAFEs proliferated, courts encountered them with increasing frequency. In a new paper, I argue that emerging case law suggests that SAFEs may have become victims of their clever branding. Courts generally seem to grasp that SAFEs are not stock but may only provide for “future equity,” such that shareholder status only follows from a conversion event. What they tend to miss is that this is only part of the picture. Before conversion, SAFEs are designed to deliver preferred stock economics without preferred stock governance. SAFEs give their investors a well-defined claim on the startup’s assets, broadly similar to that held by later startup investors holding preferred stock, while allowing both sides to defer or avoid the control rights and fiduciary duties that come with being a shareholder. Courts that fail to appreciate this design choice risk either stripping SAFEs of aspects of their economic substance or saddling holders with governance rights they never asked for. Either way, the result risks eroding the qualities that made SAFEs such an elegant solution to the challenges of early-stage finance.
The paper begins by laying out the problem that the SAFE was created to solve. Investing in early-stage startups is an inherently risky business. Founders almost always have a significant information advantage over investors, outcomes are wildly unpredictable, and once money is in the door, the interests of founders and investors start to pull in different directions. Traditional VC deals address these tensions by giving investors preferred stock bundled with a set of protective rights. This works well enough once a company has reached a certain level of maturity, but it falls apart at the earliest stages. Negotiating and executing a full venture deal demands time and resources which are hard to justify when the investment amounts involved are still small. Pricing the startup’s stock is another obstacle: At such an early point, any valuation is little more than a guess. Lastly, a standard venture deal would hand investors governance rights that early investors do not need or want when the startup is still in experimentation mode.
The paper then traces how these constraints pushed the market away from common-stock seed rounds toward deferred-equity instruments. Common-stock deals were leaner and less complex than traditional preferred stock deals but still forced founders and investors to settle on a valuation and still burdened investors with shareholder status, which meant a statutory rights and protections package that served no purpose at that stage. Convertible notes came next, offering a way to get deals done quickly while kicking the valuation question down the road but achieved this by wrapping the investment in a debt instrument, which introduced its own complications. The success of SAFEs reflects the market’s frustration with those complications. By dropping the debt features of notes while holding onto their business logic, SAFEs managed to accomplish the same goals of deferring valuation, keeping transaction costs low, and avoiding premature governance structures without the baggage that came with a debt-based structure.
The paper then takes a close look at the SAFE’s conversion and payout mechanics, since this is where courts tend to get things wrong. It walks through the process by which SAFEs convert into shares and how SAFE holders get paid out at exit, drawing distinctions between discount-based SAFEs and the two versions of valuation-capped SAFEs that Y Combinator has put out over the years. It shows that modern SAFEs give their holders a well-defined slice of the startup’s economic upside, along with liquidation preferences and antidilution protections that are broadly comparable to those enjoyed by preferred stockholders, even if by other means. SAFEs are, therefore, more than just placeholders waiting to become future equity. A better way to think about them is as a kind of stockless preferred stock.
Finally, the paper traces how courts have begun to reason about SAFEs, and where that reasoning has gone off course. By emphasizing what SAFEs are not (not yet stock, not conferring shareholder status) rather than what they are (embodying preferred stock-like economic claims unbundled from governance rights), courts have either discounted their economic substance or treated their minimal-governance structure as a gap to be filled rather than a design choice to be respected.
In LifeVoxel, a U.S. District Court for the Southern District of California held that a SAFE holder making a Rule 10b-5 fraud claim must show economic loss and loss causation by showing that the SAFE will never convert, presumably based on the assumption that SAFEs have no trackable economic value prior to their conversion. SAFEs, however, certainly have economic value while outstanding, the components of which are essentially similar to those comprising the economic value of preferred stock. Estimates of that value may be contestable, but that does not make them conceptually impossible.
A related error involves classifying SAFEs as debt. In Rhodium Encore, a U.S. Bankruptcy Court for the Southern District of Texas pointed to the priority that certain SAFE payouts receive over common stock as grounds for treating SAFE investors as holding claims in bankruptcy. But the court glossed over the absence of anything uniquely debt-like in the SAFE’s liquidation preferences. In fact, the SAFE was designed to mirror the liquidation preference of standard, VC-style preferred stock. And preferred stock, even the kind with features that are significantly more debt-like than what VC-style preferred stock offers, have been treated as equity in bankruptcy.
On fiduciary duties, courts have been broadly correct to rule that SAFE holders are not owed such duties before conversion, since they are not shareholders, and deny offering such protections under a “prospective shareholder” theory. That result preserves an important feature of early-stage finance: offering founders and investors a menu of governance choices. Investors who want fiduciary protections can insist on a stock-based deal, while those who place less weight on fiduciary oversight can choose non-stock instruments such as notes or SAFEs. Problems start where courts deny fiduciary duties at the front end but end up recreating them through the back door by stretching contractual doctrines like the implied covenant of good faith to police management discretion. Doing so, they collapse the menu of choices and ignore how the founders and investors have chosen, of all available early-stage investment instruments, to use precisely the alternative that provides its holders with the fewest governance rights.
The paper makes three contributions to the literature. First, it offers a theory of the problems SAFEs were built to solve in early-stage finance and why they provided such an effective solution. Second, based on a close analysis of how SAFEs convert and pay out, the paper develops a descriptive model of SAFEs as “stockless preferred stock.” Finally, it uses these insights to assess, for the first time in the legal literature, the growing body of SAFE case law.
Gad Weiss is a Wagner Fellow at NYU’s Pollack Center for Law & Business. It is based on his recent working paper, “Safeonomics,” forthcoming in the Yale Journal on Regulation and available here.