Direct Listings: What Spotify and Coinbase Actually Proved

In April 2018, Spotify became the largest company ever to go public without an offering. No shares were sold by the company, no underwriters priced a book, no allocations were handed out and no lockup applied. The exchange published a reference price and the market opened.

Three years later Coinbase did the same thing at greater scale, on a different exchange, in a far more volatile asset class.

Both were reported as validations of a structure that would displace the traditional IPO. That has not happened, and the reason is more interesting than the prediction. What these two listings actually demonstrated is narrower and considerably more useful: a direct listing solves the pricing problem completely and solves almost nothing else.

This piece covers what happened, what the numbers mean, what the structure fixes, what it does not, and — the part almost universally omitted — what it changes about legal liability.

1. What Actually Happened

Spotify, 3 April 2018, NYSE

• Reference price set at $132 by the exchange.
• Opened at $165.90, about 26% above the reference.
• Closed at $149.01 — up roughly 13% on the reference, but well below where it opened, valuing the company around $26.5 billion.
• Goldman Sachs, Morgan Stanley and Allen & Company acted as financial advisers rather than underwriters, on materially smaller fees than a conventional spread.

Coinbase, 14 April 2021, Nasdaq

• Reference price set at $250, implying roughly $49.8 billion.
• Opened at $381, ran as high as $429.54, then fell to around $310 intraday.
• Closed at $328.28, about $85.8 billion fully diluted.
• No capital raised, no lockup restrictions on existing holders.

Notice the shape of both days: a strong open followed by a decline. That is not a coincidence, and it is the single most instructive feature of the structure.

2. The Reference Price Is Not a Price

Nearly every report of these two events described a percentage gain against the reference price. Those percentages are meaningless, and the misunderstanding is worth correcting because it distorts how founders evaluate the structure.

A reference price is not an offer price. Nobody buys at it. No capital is raised at it. It is a figure the exchange publishes — informed by recent private transactions and the financial advisers' input — purely as a starting anchor for the opening auction. It carries no commitment from anyone.

Compare that with an IPO offer price, which is a real transaction: a defined number of shares actually change hands at that number, and the company actually receives the proceeds. When Snowflake priced at $120 and opened at $245, billions of dollars genuinely moved from the company to the buyers who were allocated stock at $120.

When Coinbase's reference was $250 and it opened at $381, nothing moved. No one bought at $250. The company was not underpaid by $131 a share, because the company was not paid at all.

The practical consequence: a direct listing cannot be underpriced, because there is no price to be under. That is the structure's central and genuine achievement, and it is obscured every time a headline reports a reference-price pop as though it were the same phenomenon.

3. What a Direct Listing Genuinely Solves

• Underpricing, entirely. No discounted allocation exists, so there is no transfer from existing shareholders to favoured buyers. This is not an improvement at the margin — it removes the mechanism.
• Allocation politics. Nobody decides who gets to own you cheaply. The register is formed by whoever buys in the open market.
• Cost. Advisory fees are meaningfully lower than a full underwriting spread, which on a large offering is a substantial sum.
• Immediate liquidity for existing holders. Employees and early investors can sell from day one instead of waiting six months. For a company that has been private for a decade, this is a real benefit to real people — and it addresses the pressure that otherwise accumulates behind a lockup, which is why structured tender offers and secondaries exist in the first place.
• No dilution. No new shares are issued in the classic form of the structure.

4. What It Does Not Solve

This is where the enthusiasm outran the evidence.

It does not produce a stable price. Coinbase moved from $429.54 to roughly $310 within a single session — a range of nearly 40% of its own low. Spotify closed well below its opening print. An opening auction with no stabilisation agent and no greenshoe discovers a price by letting it move, which is exactly what it is supposed to do. It is also uncomfortable, and it is one reason companies with a fragile equity story avoid the structure.

It does not raise capital in its classic form. If you need money, this is not the mechanism — though see section 7.

It does not build a shareholder register deliberately. A book-built IPO lets a company place stock with long-only institutions it wants as holders. A direct listing takes whoever turns up. For a company that wants patient, named holders, that is a real loss.

It does not come with a marketing engine. The IPO roadshow is a sales process. Spotify and Coinbase did not need one — both were household names in their categories with businesses a generalist could understand in a sentence. A company whose story requires explanation forfeits the machine that does the explaining.

It does not come with price support. There is no stabilisation, no over-allotment option, no syndicate desk with an incentive to defend the level. When it falls, it falls.

5. The Lockup Difference Cuts Both Ways

Existing holders in a direct listing are generally free to sell immediately. This is usually presented as a straightforward advantage. It is a trade.

The advantage: no calendared supply cliff. The lockup expiry is the mechanism that did the real damage to Facebook in 2012 — a published schedule of hundreds of millions of shares becoming sellable on a known date, into a market that had not priced it. A direct listing has no such date because the supply is available from the start.

The cost: that supply arrives on day one, into a market with no established clearing level and no stabilisation. The strong-open-then-fade pattern in both listings is at least partly this — initial enthusiasm meeting genuine sellers immediately rather than six months later.

Which is preferable is a real question rather than a settled one. Front-loaded supply into a discovering market is not obviously better than deferred supply into an established one. What it certainly is, is honest: the price reflects actual willingness to sell from the first print, rather than a level held up by a temporary restriction on the people most likely to sell.

6. The Liability Question Almost Nobody Mentions

This is the most consequential difference for founders, boards and counsel, and it is missing from most coverage of the structure.

In a traditional IPO, underwriters perform diligence and carry liability. They act as a gatekeeper — not from altruism, but because they are exposed if the registration statement contains a material misstatement. That exposure is a genuine quality control on disclosure.

In a direct listing, no underwriter occupies that role. The company's obligations under securities law remain fully in force, but the private-sector gatekeeper with money at risk is absent.

The second-order effect is subtler and became a live legal issue. In a direct listing, registered shares and previously unregistered shares begin trading simultaneously and become indistinguishable in the market. That matters because a claim under Section 11 of the Securities Act — the provision covering false or misleading registration statements — has historically required a purchaser to trace their shares to the registered offering. When the two classes are commingled from the first trade, tracing may be impossible.

The Supreme Court addressed this in 2023 in litigation arising from Slack's direct listing, holding that a Section 11 plaintiff must plead that they purchased shares issued under the registration statement they are challenging. The practical result is that a direct listing can make certain shareholder claims substantially harder to bring.

Whether that is good or bad depends entirely on where you sit — but it is a material structural difference, and any founder choosing this route should have counsel explain it rather than discovering it later. Our guide to choosing counsel is relevant well before this point.

7. Raising Capital in a Direct Listing

The original objection — that you cannot raise money this way — has been partly addressed. Exchange rule changes approved around 2020 and 2021 permit a company to sell newly issued shares into the opening auction, combining primary capital with direct-listing price discovery.

Adoption has been limited, for reasons worth understanding:

• The amount raised is uncertain until the auction clears, which is difficult to plan a balance sheet around.
• No underwritten commitment means no guarantee the capital arrives at all.
• Banks have little incentive to promote it, since it compresses the economics of the most profitable transaction they do.

So the structure remains, in practice, mostly for companies that do not need the money — which is precisely the profile of both Spotify and Coinbase.

8. Who This Is Actually For

It fits a company that: is already well known to the buyers it needs; has a business a generalist investor can understand quickly; does not need primary capital; has a shareholder base that wants liquidity; and has disclosure clean enough to withstand scrutiny without an underwriter having pressure-tested it first.

It does not fit a company that: needs capital on a schedule; has a complex or unfamiliar story; wants to choose its long-term holders; would struggle with a volatile first week; or has anything in its filings it would rather a diligence process did not find — as the WeWork filing demonstrated, disclosure gets read either way.

That is a narrow band, which is why direct listings did not displace the IPO. It is also a real band, and the companies that fit it have saved their shareholders considerable sums.

Frequently Asked Questions

Did Spotify and Coinbase "work"?

On the structure's own terms, yes — both achieved liquid public trading with no underpricing transfer and materially lower fees. Their subsequent share price performance is a separate question about the businesses and their markets, and conflating the two is the most common error in assessing the structure. A listing mechanism does not determine what a company is worth in three years.

Why do the reported percentage moves not count as a pop?

Because the reference price is not a transaction. In an IPO the offer price is a real trade at which the company receives money, so the gap to the first-day close measures value transferred away from the issuer. In a direct listing the gap measures the distance between an exchange's estimate and where the auction cleared — which costs the company nothing.

Do the banks lose out?

They earn less, not nothing. Advisers still handle the registration statement, investor education, reference price input and the opening process. The fees are well below a full underwriting spread, which is a genuine saving for the issuer and a genuine reason the structure has not been enthusiastically marketed by the people who would otherwise market it — see our guide to how investment banks work with companies.

Is a volatile first day a problem?

Less than it feels. Price discovery is what the opening auction is for, and a wide range on day one is the mechanism functioning rather than failing. The genuine issues are practical: employees making decisions in a fast-moving market, and a public narrative that fixates on the intraday high. Both are manageable with preparation, and neither is a reason to pay for underpricing instead.

What about SPACs — do they solve this?

They solve a different problem, largely speed and negotiated price certainty, at the cost of dilution from sponsor promote and warrants, and far lighter scrutiny than a registered offering. Our comparison of IPOs, direct listings and SPACs works through the three side by side. They are not substitutes for one another; they suit genuinely different situations.

Would this suit the very large companies listing now?

Partly. The largest private companies are exceptionally well known and could clear an auction without a roadshow — but most of them are listing precisely because they want capital, and at a scale where auction uncertainty is unattractive. The more likely direction is hybrid structures that borrow direct-listing price discovery while retaining a committed raise, which is one of the tensions we track in our coverage of the AI listing pipeline and the SpaceX offering.

The Bottom Line

Spotify and Coinbase proved that a well-known company can reach the public market without underwriters, without an offer price and without a lockup — and that doing so eliminates the underpricing transfer entirely.

They also proved the structure gives you nothing else: no capital, no stabilisation, no chosen register, no sales process, and a materially different liability position that cuts in the company's favour and against its future shareholders.

That is not a revolution. It is a well-defined tool for a well-defined situation — and the companies that fit it should stop being talked out of it by people paid on the spread.

Global Capital Network connects founders, investors and the bankers, counsel and advisors who structure these transactions. See upcoming events or get in touch.

Accurate as at 2 August 2026. Figures are as reported contemporaneously. Primary sources: CNBC on the Spotify reference price, CNBC on the Spotify debut, CNBC on Coinbase, CoinDesk, and filings via SEC EDGAR. This article is general information and market history, not legal or investment advice. The Section 11 discussion is a simplified summary — take specific advice.