Spotify's $26.5 Billion Direct Listing That Skipped the IPO

On April 3, 2018 Spotify Technology S.A. became a public company without selling a single share. No bank underwrote the listing, no order book was built, no offer price was struck the night before, no stock was allocated to favored institutions, and no shareholder was contractually barred from selling. The New York Stock Exchange published a reference price of $132.00, the shares opened at about 12:45 p.m. Eastern at $165.90, and the session closed at $149.01, valuing the company near $26.5 billion.

A direct listing was not itself new. Spun-off subsidiaries, companies emerging from Chapter 11, and foreign issuers already quoted at home had all arrived on US exchanges without an underwritten offering. What had never happened was a private company of Spotify's size and profile choosing the route deliberately, and the reason it had never happened is that the machinery of going public assumes an underwriter at every step: the price on the prospectus cover, the lockup, the stabilizing bid, the allocation list, the research restriction. Removing the underwriter meant persuading the SEC and the NYSE to rebuild several of those steps from scratch, in public, on a deadline set by a convertible note.

The question the transaction poses is narrower than the coverage suggested at the time. Spotify's own case was never that IPOs are broken for everyone; it was that the traditional model did not fit a company that needed no money, had a global consumer brand, and had shareholders who wanted to sell at market prices rather than at a price a syndicate had set for them. Whether the deal fixed the first-day transfer of value that has bothered issuers for four decades, or simply proved that about a dozen unusually well-placed companies can route around it, is still argued. This account is built from Spotify's Form F-1 and final prospectus, the SEC's Regulation M no-action relief, the amended NYSE listing rules, a case study co-written by Spotify's own general counsel, contemporaneous reporting, a University of Chicago Law Review analysis, and Jay Ritter's direct listing dataset maintained through 2026.

A business already funded in private

Spotify arrived at the exchange with a balance sheet that made the usual reason for an IPO irrelevant. For the year ended December 31, 2017 it reported revenue of €4,090 million, up about 39%, against a net loss of €1,235 million, and it disclosed 157 million monthly active users and 71 million Premium Subscribers, growing 28% and 46% respectively. It had also turned free cash flow positive, which is the fact that matters for this deal: the company was not funding its own operations out of investor capital.

The headline loss makes the "no capital need" claim look strange until it is decomposed. Most of it was not operating: the operating loss was €378 million, and the gap to the bottom line came overwhelmingly from finance costs on a convertible instrument that was accounted for at fair value and was, by the time of the listing, no longer outstanding. A company can post a billion-euro accounting loss and still have no use for primary proceeds, and Spotify was that company.

That single fact removes the strongest argument for hiring underwriters. A firm commitment underwriting exists to convert an uncertain market into a certain sum of money in the issuer's account on a known date. If there is no sum of money to certify, the insurance is being bought against a risk the issuer does not run, and the fee paying for it becomes very hard to justify.

Direct listing

A listing in which a company's existing shares begin trading on an exchange without any underwritten offering, primary or secondary. The company issues nothing and raises nothing; existing holders become free to sell into the open market but are not obliged to. Because there is no underwriting syndicate, the features that syndicate builds around an IPO fall away with it: no fixed offer price, no allocation, no over-allotment option, no stabilization, and no lockup agreements.

The clock TPG and Dragoneer had set

Spotify's freedom from needing money came with a deadline attached. On April 1, 2016 it had issued $1,000 million of convertible notes to investors led by TPG and Dragoneer Investment Group, on terms that grew more expensive the longer the company stayed private. The coupon started at 5% and stepped up by a full percentage point every six months after two years, the conversion discount to an eventual offering price started at 20% and widened by 2.5 percentage points every six months after the first year, and the holders could sell 90 days after a listing rather than the customary 180.

The provision that actually mattered arrived with the exchanges that followed: under the January 2018 exchange agreement, if Spotify had not listed its ordinary shares on or before July 2, 2018, each exchanging noteholder could elect to unwind: Spotify would buy back the shares and reissue notes on materially identical terms, resetting the whole ratchet. Recode reported in February 2018 that this deadline was the reason the listing had to happen in 2018 at all, which is a fair reading of the incentive even though Spotify never framed it that way.

Spotify cleared the instrument off its balance sheet before the bell rather than at it. On December 15, 2017 holders exchanged $301 million of principal plus $27 million of accrued interest for 4,800,000 ordinary shares, which they then sold to an affiliate of Tencent. A second December exchange took out a further $110 million of principal plus $10 million of interest for 1,754,960 shares, and in January 2018 the remaining $628 million, plus interest, was exchanged for a further 9,431,960 shares. The mechanics of convertible instruments are worth holding here, because the terms were written for an offering price that a direct listing would never produce, and settling in advance removed a live question about how a conversion formula anchored to an IPO could work when there was no IPO.

Two founders who could not be diluted

The last piece of the pre-listing picture is governance, and Spotify's is unusual even by technology standards. As a Luxembourg company it issues beneficiary certificates, and Daniel Ek and Martin Lorentzon held ten of them for every ordinary share they owned of record. Cleary Gottlieb calculated in April 2018 that this gave the two founders roughly 80% of the voting power on record ownership of about 21% of the shares.

Two readings of that structure sat side by side in the commentary and both are defensible. Cleary noted that, unlike Snap's non-voting public class, every Spotify shareholder does have a vote, and that the certificates expire on transfer and lapse if the founders' record holding falls below roughly 4% of the shares, making the extra power personal rather than permanent capital. Against that, there is no time-based sunset, which the Council of Institutional Investors and other governance investors had been pressing for across the whole dual-class debate.

Either way, the structure removes the question a capital raise would have forced. There was no dilution to negotiate, no new class to create, and no reason for the founders to care what a syndicate thought a share was worth. As a foreign private issuer Spotify would also report in euros under IFRS, file annually on Form 20-F rather than quarterly, and sit outside the US proxy rules and Section 16 insider reporting, which later mattered a great deal for anyone trying to work out who was actually selling.

The listing standard that had to change first

The obstacle was not the SEC's registration regime; it was a paragraph of the NYSE Listed Company Manual. Section 102.01B required a company that had never been SEC-registered to demonstrate a market value of publicly held shares of at least $100 million, established through a combination of an independent third-party valuation and a sustained recent trading history in a private placement market. Spotify's shares had changed hands privately, but not on the kind of venue, and not with the kind of continuity, the rule contemplated.

The exchange began the formal rule filing process with the SEC in March 2017 and amended the proposal three times, and the SEC approved the change on February 2, 2018, two months before Spotify traded. The amended rule allows the NYSE to waive the private-market trading test where a company has a recent independent valuation showing at least $250 million of publicly held shares and, critically, engages a financial adviser to be consulted by the exchange's designated market maker in setting the opening price.

Nobody at the NYSE named Spotify, and the rulemaking was justified by the broader phenomenon of large private companies delaying listings because they had no need of primary capital. The timing is hard to read as coincidence, and the second limb of the new test matters most: the mandatory adviser consultation is the clause that put an investment bank back inside a transaction designed to do without one.

Reference price

The price an exchange publishes before a direct listing opens, derived from recent private transactions in the stock and from the financial adviser's read of buying and selling interest. It is not an offer price and not a valuation: no shares change hands at it and no one is committed to trade there. Its function is procedural, anchoring the opening auction and the exchange's volatility bands, which is why a "first-day pop" measured from it describes nothing anyone paid.

A registration statement nobody had drafted before

Because Spotify was selling nothing, its Form F-1 was not an offering document in the ordinary sense but a resale shelf registration statement: it registered 55,731,480 ordinary shares, about 31% of the 178,112,840 then outstanding, held by affiliates, by the founders, and by employees and other holders who could not yet rely on Rule 144. Roughly 60% of the company was held by non-affiliates who had owned their shares for over a year and were free to sell without registration at all. Taken together, about 91% of the stock could legally be sold on day one, against a float that a conventional IPO would have capped at whatever the syndicate chose to sell.

Two drafting problems had no precedent. Item 501(b)(3) of Regulation S-K requires a bona fide estimate of the price range on the cover of a preliminary prospectus, and there was no range because management would play no part in pricing; Spotify substituted an explanation of how the opening price would be determined plus the high and low prices of recent private transactions, the highest of which was $132.50 between January 1 and March 14, 2018. The underwriting section was replaced by a plan of distribution limited to ordinary brokerage transactions, deliberately narrow so the listing could not be characterized as an organized sale.

The SEC's review reflected how new all of this was. Latham & Watkins, which acted for Spotify alongside general counsel Horacio Gutierrez, reported that roughly a third of the staff's comments on the F-1 concerned the structure itself, the opening-price procedure, and the roles of the advisers and the market maker. The registration statement went effective on March 23, 2018 and was withdrawn on June 20, once the 90-day period that unlocked Rule 144 for the remaining holders had run.

Lockup agreement

A contract, customarily 180 days, in which insiders and pre-IPO investors promise the underwriters not to sell after an offering. Its purpose is to manage post-deal supply so the syndicate can distribute stock into a market it is not competing with. Because a direct listing has no syndicate to protect, there is nothing for a lockup to serve, and Spotify imposed none; the trade-off is that all of the supply that an IPO defers arrives on the first morning.

The Regulation M problem and the SEC's answer

The hardest legal question was whether any of this was a distribution. Regulation M restricts issuers, selling holders and distribution participants from bidding for or buying the security during a defined restricted period around a pricing, and in an IPO the boundaries of that period are obvious because there is an offer price and a fixed number of shares. In a direct listing there is neither, so the rule's timing had no anchor.

Spotify sought a no-action letter from the SEC's Division of Trading and Markets rather than conceding the point. The submission represented that participants would observe a restricted period beginning five business days before the designated market maker determined the opening price, the customary pre-pricing window in a traditional IPO, and ending when secondary trading began; the Division said it would not recommend enforcement. The relief was procedural, but it told shareholders and brokers exactly when normal trading behavior could resume.

Underneath sat a more revealing fact. The NYSE had originally proposed a third change, allowing a company to list on effectiveness of an Exchange Act registration alone with no Securities Act registration at all, and withdrew it in December 2017 after the SEC sought comment on whether such a listing raised unique issues around distribution participants, pricing information and the supply of shares. As Cleary Gottlieb observed in April 2018, the practical result was that the regulatory process had to unfold as though there were an IPO even when there was not, which is the most economical description of the whole exercise.

An Investor Day streamed to anyone with a browser

A conventional IPO markets itself through a one or two-week roadshow: group meetings with institutions, one-on-ones for the largest accounts, and a recorded version posted for retail. That process exists to build a book, and Spotify had no book to build, so it replaced the whole apparatus with a single event. On March 15, 2018 it held an Investor Day that ran just over two hours, roughly twice the length of a standard IPO group meeting, presented by the entire leadership team rather than the chief executive and chief financial officer, streamed live to about 10,000 unique viewers and left online afterward.

The event still counted as a road show under SEC rules, which is why the timing was fixed rather than chosen: a company that submits its registration statement confidentially must file publicly at least 15 days before beginning one, and Spotify's first public filing landed on February 28. The presentation was treated as a free writing prospectus and filed with the SEC, as were four additional videos with filed transcripts on specific topics.

Daniel Ek framed the choice as a temperamental one rather than a financial one, in a blog post published the day before trading.

The claim to equal access is strong but not total, and the prospectus itself shows where the seam is. Spotify's advisers were expressly not engaged to participate in investor meetings or coordinate price discovery, and the internal investor relations team led the education effort; but the advisers did form their own view of demand separately from the company's marketing, and one of them would carry that view into the opening auction. The difference from a bookbuild is real, and it is a difference of degree.

Guidance ten days before there was a price

Spotify then did something no company in registration does. On March 26, 2018, three days after effectiveness and eight days before trading, it issued a press release with its financial outlook for the first quarter and full year 2018, furnished to the SEC on Form 6-K, which is the behavior of a seasoned public company rather than an issuer in a quiet period.

The eleven-day gap between effectiveness and the first trade was engineered for exactly this. In an IPO, effectiveness means pricing that evening and trading the next morning; Spotify wanted its guidance to season with investors before a price existed, and it also needed time for thousands of existing holders to move shares into brokerage accounts through the Depository Trust Company so they could actually trade from the opening bell if they chose.

The approach carried a real legal question, and Cleary flagged it plainly: publishing forward-looking guidance days before a first trade is only clearly not an offer if the issuer can lean on Rule 168, which contemplates a company that has previously released information of the same type in the ordinary course. A first-ever guidance release from a company that had never reported publicly is a harder fit. Spotify took the risk in the service of the transparency argument, and no enforcement followed.

The three banks that were not underwriters

Spotify retained Goldman Sachs, Morgan Stanley and Allen & Company as financial advisers, with defined and deliberately limited mandates: help set objectives for the listing, advise on the registration statement, and help prepare the Investor Day and other public communications. They ran no bookbuilding, took no orders, made no allocations, provided no price support, and committed no capital. Morgan Stanley alone took the additional role the amended NYSE rule required, consulting with the designated market maker on the opening price.

The prospectus disclosed roughly $35 million of fees for other advisers, which is understood to cover the three banks, within total listing expenses of about $46 million. Writing in the University of Chicago Law Review, Benjamin Nickerson estimated that a firm commitment underwriting for a company of Spotify's size would have cost $80 million to $120 million, a gross spread that in midsized deals runs to exactly 7% of proceeds and in the largest ones is negotiated down toward 1%.

Nickerson's argument is that the exhibit above understates what the banks did. Because the advisers helped draft the registration statement and the investor materials and, through Morgan Stanley, performed the pricing function the SEC itself described as an underwriter's, he contends they should be treated as statutory underwriters under Section 11 of the Securities Act, with strict liability for material misstatements and a due diligence defense. Spotify structured expressly against that conclusion, and the point has never been litigated; what is not in dispute is that the same banks did a materially narrower job for roughly a third of the money and none of the exposure.

Master the mechanics behind a deal answer: practice 1,000+ technical questions on equity capital markets, valuation, and deal structuring, download our iOS app for the full toolkit.

A price that was published, not paid

The NYSE put out a reference price of $132.00 ahead of trading, in line with the $132.50 high of the private transactions Spotify had disclosed on the prospectus cover for the period from January 1 to March 14, 2018. Not one share traded at it, and none was ever meant to.

That distinction was lost almost immediately in the coverage, and it is the most common error made about direct listings. Because the opening print came at $165.90, headlines computed a 25.7% first-day gain, using the arithmetic that measures an IPO pop from an offer price at which money genuinely changed hands. Here nothing did, which is why the only defensible first-day return runs from the open to the close.

The auction, and the bank inside it

Spotify appointed Citadel Securities as its designated market maker, one of a handful of firms holding that role on the NYSE floor. Its job in any listing is to run the opening auction, gathering pre-open interest until a single print clears; here it did that without an underwriter's price to work toward, which is why price discovery took more than three hours after the opening bell rather than minutes.

The amended rule required a financial adviser in the room, and Morgan Stanley filled it. Its contribution, as the prospectus described it, was to give the market maker an understanding of the ownership of Spotify's outstanding shares and of the pre-listing buying and selling interest it was aware of from potential investors and holders. The prospectus was equally careful to state that this consultation happened without coordination with Spotify, and that the company would play no part in setting the opening price.

Designated market maker (DMM)

A firm assigned to a listed stock on the New York Stock Exchange floor with obligations to quote continuously, dampen volatility, and open and close the security through manual auctions. At an IPO the DMM works alongside the lead underwriter, which knows the offer price and holds the book. In Spotify's listing there was no book and no offer price, so the DMM balanced live orders against an adviser's read of who owned the stock and who wanted to trade it.

The uncomfortable symmetry is that this is a recognizable description of what a syndicate desk does on the morning of an IPO: aggregate what it knows about holders and demand, and help produce an opening level. The differences are that Morgan Stanley controlled no allocation, took no principal risk, and was not paid a spread on proceeds. Whether those differences are the substance of underwriting or merely its plumbing is precisely the question Nickerson raised and no court has answered.

What the tape actually said

The opening print of $165.90 turned out to be the day's high, and the stock spent the session drifting away from it, reaching $148.26 before closing at $149.01, down 10.2% from the open and 12.9% above the reference price. Volume was 30,526,500 shares, about 17% of the shares outstanding, which is heavy for a first session and unsurprising given that roughly 91% of the company was legally free to sell.

The number that settled the immediate argument was volatility, not direction. Latham & Watkins reported intraday volatility of 12.3% on the first day against 15.6%, 34.7% and 48.1% for three other large technology IPOs, which is the opposite of what critics had predicted for a stock with no lockup and none of the greenshoe and stabilization machinery that steadies an ordinary debut. Barbara Gray of Brady Capital Research had warned before the listing that a flood of sellers could hit a market with none of an IPO's supports, and the flood did not arrive.

Some of the selling was visible after the fact. Sony Music sold 17.2% of its 5.71% stake into the first day, and Sony guided to roughly $986 million of non-operating profit in the quarter from the sale and from appraisal gains on what it still held. Most of it was not visible at all: as a foreign private issuer Spotify's insiders were outside Section 16, so there was no prompt public record of who sold, which is a genuine cost of the structure that had nothing to do with price.

The selling wave that did not come

The bear case for a direct listing was mechanical rather than fundamental: with every holder free on day one and no bank obliged to bid, supply would overwhelm a thin market and the price would find a level far below where an orderly IPO would have left it. Over the following weeks the stock traded in a relatively narrow range on unremarkable volume, and CNBC's assessment seven weeks in was that the feared dumping had not materialized.

The calendar then did the rest. Spotify reported first-quarter results on May 2, 2018 and filed a prospectus supplement the following day to keep the resale registration current, behaving from the start like a company that had been listed for years. On June 20, 2018 it filed to deregister the shelf, effective the next day, because by then it had been an Exchange Act reporting company for 90 days and its remaining holders could sell under Rule 144 without it.

Sony Music sold half of its original 5.71% stake by the end of the second calendar quarter for a confirmed $768 million, leaving it about 2.85%, which is the clearest evidence available that large holders did use the liquidity the structure was built to provide, and used it in size, without breaking the stock. That is the strongest single piece of support for the design: the thing an IPO's lockup exists to prevent happened anyway, in the open, and the market absorbed it.

The arithmetic of what was not left on the table

Measured the way Jay Ritter of the University of Florida measures direct listings, from the open where shares actually change hands to the close, Spotify's first-day return was negative 10.2%. Across the twelve direct listings Ritter counts between 2018 and 2021, the average first-day return on that basis was negative 1.8%. Across 3,404 other US IPOs from 1999 to 2021, the average first-day return measured from the offer price was 30.2%.

Those two numbers are not directly comparable, and that incomparability is the analytical heart of the deal rather than a footnote to it. An IPO's pop is a measurable transfer, the money left on the table, because the issuer and its selling holders received the offer price and someone else received the close. A direct listing has no offer price, so there is nothing to measure: whatever the market decided at the open is what selling shareholders received. Spotify's holders sold at $165.90 rather than at a number a syndicate had set, and the fact that the stock then fell 10% is a cost borne by the buyers, not by the sellers.

Money left on the table

The number of shares sold in an IPO multiplied by the gap between the first-day close and the offer price. It measures value that moved from the issuer and its selling holders to the investors who received allocations. Because a direct listing sells no shares at a set price, the measure does not exist for one, which is either the point of the structure or a way of making the comparison unanswerable, depending on who is arguing.

The counterfactual is worth stating carefully because it is where advocates overreach. In 2017 the average US IPO rose 11.8% on its first day, and through late March 2018 the figure was 13.2%; a conventional Spotify IPO of comparable size would plausibly have delivered a pop in that range to a few dozen institutions. What the direct listing did not do is prove that pop was avoidable in general. It proved it was avoidable for a company that did not need to sell anything to anyone.

The liability question the Supreme Court answered

The structure's most consequential legacy for investors was decided in a case about Slack, not Spotify, and it runs against buyers. Because registered and unregistered shares begin trading together on day one of a direct listing, a purchaser generally cannot tell which she bought. Fiyyaz Pirani bought 30,000 Slack shares on June 20, 2019, sued under Section 11 of the Securities Act over the registration statement, and won in the Ninth Circuit, which relaxed the long-standing requirement that a plaintiff trace her shares to the challenged filing.

On June 1, 2023 the Supreme Court reversed unanimously, restoring the strict tracing requirement and leaving the parallel question under Section 12 to the lower courts. The practical consequence is blunt: in a direct listing, the strict-liability provision that disciplines disclosure in a conventional IPO is very hard to invoke, because the shares are commingled from the first trade.

Go deeper before an interview: our 160-page PDF covers capital markets mechanics, deal frameworks, and behavioral prep end to end, access the IB Interview Guide and pair it with these case studies.

Slack, Palantir, Coinbase, and then almost nobody

The template held for three years and for a specific type of company. Slack listed on June 20, 2019 against a $26.00 reference price, opened at $38.50 and closed at $38.62. Palantir and Asana both listed on September 30, 2020. Roblox followed in March 2021, and Coinbase went to Nasdaq in April 2021 with a $250.00 reference price, an open at $381.00 and a close at $328.28. Squarespace, ZipRecruiter, Amplitude and Warby Parker completed the 2021 cohort.

The pattern in the exhibit is the argument. Every company on it was a consumer or enterprise brand that a retail investor could describe in a sentence, every one had raised heavily in private markets, and none needed the money. Ritter's dataset records that from 2022 through 2026 the companies choosing direct listings have generally been microcaps, most of them with violently volatile debuts, which is a different phenomenon wearing the same name. Twelve listings in four years, only ten of them companies of consequence, then a long tail of small ones, is not the displacement of the IPO that was forecast.

The rule that let companies raise money, and the companies that did not

The obvious fix was to let a direct listing raise primary capital, and the exchanges built it. The SEC approved the NYSE's primary direct floor listing rule on August 26, 2020, stayed it days later after the Council of Institutional Investors signaled a petition for review, conducted a fresh review, and approved it again on December 22, 2020; Nasdaq followed, and both exchanges relaxed the pricing limits in 2022 after issuers found the original bands too tight.

Gurley, the venture capitalist who had campaigned hardest against IPO underpricing, told CNBC on the day of the second approval that the change would unquestionably end traditional IPOs, pointing at Airbnb and DoorDash, which had risen 112% and 86% on their December 2020 debuts. More than five years later, no company has completed a primary direct listing on either exchange.

That gap between the rule and its use is the most informative single fact in the aftermath. A company raising primary capital wants a known amount on a known day, and an auction that must clear inside a price band offers neither; the underwriter's firm commitment, expensive as it is, is the product being bought. Once capital is genuinely needed, most issuers conclude they want the insurance after all, which is why the structural alternatives to a bookbuilt offering have stayed niche through two full cycles rather than displacing it.

What the record settles

Three things are settled by the evidence and should be stated plainly. The transaction worked: it opened in an orderly auction, traded 17% of the company on day one with lower intraday volatility than comparable technology IPOs, absorbed genuine institutional selling including Sony's, and left a stock that has compounded from a $149.01 first close to $543.62 at the August 31, 2026 close, a valuation near $112 billion.

It was also dramatically cheaper, and the saving was real rather than a reallocation. About $35 million in advisory fees and $46 million of total expenses against a plausible $80 million to $120 million underwriting cost is money that stayed with the company, and it bought a process that did what the company wanted it to do.

And it forced permanent institutional change. The NYSE rewrote a listing standard, the SEC granted Regulation M relief it had never granted before, and both exchanges eventually built a primary capital-raising version. The route exists now because Spotify walked it, and the conventional IPO process has an acknowledged alternative it did not have in 2017.

What it does not settle

The claim that direct listings solve underpricing does not survive contact with the arithmetic. Spotify did not eliminate the first-day transfer; it eliminated the transaction that makes that transfer measurable. Its shareholders sold at the open rather than at an offer price, which is better for them, and the buyers who paid $165.90 were down 10% by the close, which is the same outcome an overpriced IPO produces for the same people. What changed is who captured the difference and who bore it, not whether a difference existed.

The comparison with the underwritten model is genuinely two-sided. Facebook's 2012 offering showed an issuer capturing almost all of the value by pricing to the top, and the lead bank then spending the day buying stock to defend the price, a support Spotify's holders did not have and did not need. Alibaba's 2014 listing showed the opposite: a 38% first-day gain concentrated in about fifty institutions, precisely the transfer Barry McCarthy, Spotify's chief financial officer and the architect of its listing, attacked when he wrote in the Financial Times that Spotify had chosen a free-market approach instead.

The liability point cuts the other way and is rarely conceded by advocates. After Slack v. Pirani, a buyer in a direct listing is in a materially weaker position to bring a Securities Act claim than a buyer in an IPO, and the diligence incentive that underwriter liability creates is absent. A structure that is cheaper for issuers and harder for plaintiffs is not obviously a net gain for markets, and reasonable people read that trade-off differently.

The narrow class problem

The strongest evidence about who this works for is who has done it. Ten companies of consequence across twelve listings in four years, all with recognizable brands, no capital need, and large diverse shareholder registers capable of supplying stock into an opening auction; zero primary direct listings in more than five years of the rule existing; and a recent cohort of microcaps whose debuts look nothing like Spotify's.

The honest verdict is therefore split, and the evidence supports both halves firmly. Spotify proved that the underwriter is not a structural necessity for a listing, that a market can discover a price without one, and that the process can be run for a third of the cost; that was doubted in 2018 and it is not doubted now. It did not prove that the model generalizes, and the record since is fairly read as evidence that it does not: when a company actually needs money, or lacks a brand, or has a concentrated register, the machinery it dispensed with turns out to be doing something. The direct listing did not replace the IPO. It carved out the narrow set of cases where the IPO was never the right instrument, and made the price of the alternative visible to everyone else.

1. 1Spotify Technology S.A., Form F-1 registration statement, filed with the SEC on February 28, 2018.
2. 2Spotify Technology S.A., final prospectus (Form 424B4), filed with the SEC on April 3, 2018.
3. 3US Securities and Exchange Commission, Division of Trading and Markets, Spotify Technology S.A. no-action letter on Regulation M (March 23, 2018).
4. 4Marc D. Jaffe, Greg Rodgers and Horacio Gutierrez, "Spotify Case Study: Structuring and Executing a Direct Listing", Harvard Law School Forum on Corporate Governance (July 5, 2018).
5. 5Cleary Gottlieb Steen & Hamilton LLP, "Spotify's Direct Listing: A Look Under the Hood" (April 17, 2018).
6. 6Paul, Weiss, "SEC Approves NYSE Rule Change to Facilitate Listing Without an Initial Public Offering" (February 2018).
7. 7Benjamin J. Nickerson, "The Underlying Underwriter: An Analysis of the Spotify Direct Listing", University of Chicago Law Review (2019).
8. 8Jay R. Ritter, "Table 13a: Direct Listings in the U.S., 2018-2026", University of Florida (updated 2026).
9. 9CNBC, "Spotify closes up 13 percent after falling from highs on first day of trading" (April 3, 2018).
10. 10Reuters, "Spotify listing to land a $986 million windfall for Sony in the first quarter," carried by Business Standard (April 4, 2018).
11. 11Variety, "Spotify's Daniel Ek Quotes Daft Punk in Day-Before-IPO Blog Post" (April 2, 2018).
12. 12Recode (Vox), "Here's why Spotify had to go public this year" (February 28, 2018).
13. 13CNBC, "Spotify's IPO disrupted Wall Street. What lies ahead now for unicorns looking to go public" (May 22, 2018).
14. 14Wilson Sonsini, "SEC Approves NYSE Proposal for Primary Direct Listings (Again)" (December 2020).
15. 15CNBC, "IPO process has gotten worse in last five years, says Bill Gurley" (December 22, 2020).
16. 16Davis Polk, "Supreme Court confirms the scope of Section 11's tracing requirement" (June 2023).
17. 17Harvard Law School Forum on Corporate Governance, "Evolving Perspectives on Direct Listings After Spotify and Slack" (December 17, 2019).
18. 18Billboard, "Sony Sells Half Its Spotify Stock, Signaling Payday for Label's Acts & Indies" (July 2018).