Who Is Actually Locked Up After an IPO?

On October 22, 2026, up to 1.3 billion SpaceX (SPCX $141.31 +3.62%) shares became free to sell, 28% of the shares under its 180-day lock-up. It wasn't the end of the lock-up, which runs to December 8, 180 days after the June 11 prospectus. It was the seventh of 15 release dates the prospectus lays out. The first came two full trading days after SpaceX's first earnings report, in August, and others followed on the 70th, 90th, 91st, 105th and 120th days. Day 135 came two days later, on October 24. Still to come are day 180 and an extended lock-up that covers about 60% of the company, including every share Elon Musk owns, with releases until mid-2027. Musk's shares get no early release. SpaceX reported its third-quarter results after the close on October 20, the stock rose 7.8% the next day, and on October 22, the release day, it fell 9.1% on 2.3 times its average volume of the previous four weeks.

The textbook lock-up is simpler: insiders agree with the underwriters not to sell for 180 days, and on day 181 the shares come loose all at once. We read the lock-up terms in the prospectuses of 440 IPOs on US exchanges from January 2024 to October 2026 to see how much of that still holds. The 180 days mostly does. The rest has changed in two directions: the biggest deals now release shares early, and in most small Nasdaq IPOs the lock-up binds only the insiders and the largest holders.

The short answer

• Six months is still the standard. 392 of the 440 IPOs (89.1%) lock up insiders for 180 days or six months. 21 use one year and 19 use 90 days.
• The big deals unlock early. 26 IPOs (5.9%) let some locked shares sell before the period ends, after the first earnings report, once the stock trades well above the IPO price, or on a fixed schedule. Among the 43 IPOs with the most trading in their first month, 14 did (32.6%). Among the 320 with the least, 3 did.
• In small IPOs, few people sign. On the Nasdaq Capital Market, 154 of 180 lock-ups (85.6%) bind only the directors, officers and largest shareholders, usually those with 5% or more. On Nasdaq's larger tiers and the NYSE, 134 of 210 (63.8%) bind all or substantially all shareholders.
• The day itself is quiet for the typical stock. Over the three trading days around the end of 330 standard lock-ups, the median stock beat the S&P 500 by 0.7%, and 154 of them (46.7%) fell. A 2001 study of 1,948 lock-ups found a 1.5% drop. Volume still jumps on the larger exchanges, by about a third, and doesn't move on the small ones.
• Few insiders rush for the exit. Only 32 of 322 companies had an insider file a Form 144 notice of a planned sale in the month after their lock-up ended, and the filings came at a steady pace for two months, not in a day-one burst.

How a lock-up works

A lock-up isn't a law. It's a contract between the underwriters and the people who owned the company before the IPO, in which they agree not to sell or hedge their shares for a set period without the underwriters' consent. The company signs one too, promising not to issue new shares. The prospectus has to describe it, usually in the Underwriting section and again under "Shares Eligible for Future Sale", and that description is what we read.

The point is supply. In a typical IPO only a slice of the company is sold, and the rest belongs to founders, employees and investors who bought in privately. Without a lock-up they could sell into the first days of trading. With one, the supply arrives later, on a date everyone knows in advance. The underwriters can release shares early at their discretion, and nearly every prospectus says so; we count only the releases a prospectus writes into the terms.

Six months is still the norm

392 of the 440 IPOs lock up their insiders for 180 days or six months, counted from the date of the prospectus or, in some small deals, from the closing of the offering. The longer one-year lock-up is mostly a small-company feature: 16 of the 21 are on the Nasdaq Capital Market.

The "other" lock-ups are mostly near misses of the standard, such as 150 or 160 days. Neutron Holdings, the parent of the scooter company Lime, chose 160 days.

The biggest deals unlock early

26 of the 440 prospectuses write in an early release. They come in three kinds, and many deals combine them:

• Earnings-based: the lock-up ends on the earlier of day 180 and the second or third trading day after a named earnings report, usually the second one as a public company. Reddit (RDDT $96.15 +5.69%)'s ended at the opening of trading on the third trading day after its June 2024 quarter results.
• Price-based: a slice of the shares is released early if the stock closes well above the IPO price, usually 25% or 30% above it on five of ten trading days. CoreWeave (CRWV $53.74 +7.16%), Figma (FIG $21.03 -5.27%), StubHub and Netskope combined the two.
• Staged: the shares come out in tranches on fixed days. SpaceX is the extreme case. Cerebras (CBRS $97.82 +1.52%) released part of its employees' shares on its first and second trading days, the second release only if the stock closed more than 33% above the IPO price on the first.

21 of the 26 are on Nasdaq's Global tiers or the NYSE, four more are large deals whose cover named only "the Nasdaq Stock Market", and one, Brazil Potash, is on NYSE American. None of the 180 Capital Market IPOs has one. Size is the clearer line: ranked by median trading value in the first 20 sessions, the top tenth of IPOs (43 deals trading at least $35 million a day) includes 14 with an early release. The rest of the top quarter has 8 among 63, and the bottom three quarters 3 among 320.

For anyone holding one of these stocks, the 180-day date in most IPO calendars is the wrong date to watch. Under their clauses, and dated from each company's earnings filing, Astera's and Reddit's lock-ups ended 38 days early and Figma's 80 days early, and in each case two or three trading days after an earnings report, when the stock was already moving.

The next very large IPO on the calendar follows the same pattern and stretches it. Anthropic's registration statement, filed on October 26 and not part of our sample, lays out 23 release dates, and the November 2 amendment, which added results through September, left them unchanged. The first, on the first trading day, covers shares held by non-founder directors and officers and some current and former employees and consultants. The last is the earlier of two full trading days after its September 2027 quarter results and day 365, which is also when the restriction on its seven co-founders ends, with timed releases along the way. The share counts are still blank in the filing, and the terms can change before the IPO prices.

In small IPOs, fewer people sign

The second change is quieter. A lock-up binds only the people who sign it, and in small IPOs fewer people sign.

On Nasdaq's larger tiers and the NYSE, the standard wording is that "our directors, executive officers and the holders of substantially all of our outstanding securities" have agreed to the lock-up. 134 of the 210 prospectuses there say that or something close, and many of the rest are companies controlled by one parent or sponsor that signs for most of the shares. On the Nasdaq Capital Market, the usual wording is that "our directors, officers and holders of 5% or more of our ordinary shares" have agreed. 154 of the 180 Capital Market IPOs (85.6%) lock up only the insiders and the largest holders.

The difference decides who can sell on the first day. A shareholder from before the IPO who sits below the threshold, a 3% holder in a deal that locks up holders of 5% or more, never signed, so those shares can trade from the start, subject only to the securities laws. How much stock that frees depends on the company: where the founders and a few investors own nearly everything, the narrow wording locks up nearly everything anyway. We read who signs, not how many shares each group holds. Most of these deals come from Asia: by the business address in their SEC filings, 123 of the 180 Capital Market companies are based in mainland China, Hong Kong, Singapore, Malaysia or Japan, and 37 in the United States. On the larger tiers, 177 of 210 are US companies.

Nasdaq's own rules approach the problem from the other end. Its float rules count only "unrestricted publicly held shares", those not held by officers, directors or 10% holders and not subject to resale restrictions, and for a company listing through an IPO the minimum must now come from the shares sold in the offering. In December 2025 the SEC approved raising the lowest of those minimums on the Capital Market to $15 million, and in May 2026 it approved a $25 million minimum offering for companies that mainly operate in China, Hong Kong or Macau. In our sample, Capital Market IPOs fell from 98 in 2025 to 11 in the first ten months of 2026; the sample can't say how much of that is the rules.

What happens on the day

The best-known study of lock-up expiries, by Laura Field and Gordon Hanka, looked at 1,948 lock-ups and found that when they ended, trading volume rose 40% for good and the stock fell 1.5% against the market over three days. We ran the same test on the 330 standard lock-ups in our sample that have ended and have prices on both sides: the stock's return over the three trading days around the first day shares could be sold, minus the return of the SPY S&P 500 fund over the same days.

• The typical stock didn't fall. The median three-day return against SPY was +0.7%, and 154 of the 330 stocks fell. In the same stocks' three-day windows a month earlier, the median was -1.2%, so the expiry didn't stand out.
• The average can still look like 2001. For the 80 deals on the larger exchanges where substantially all holders were locked up, the mean was -1.6%, close to Field and Hanka's figure. It comes from a few large drops: five health-care stocks, three drug developers and two medical-device makers, fell 22% to 43% in the window. It is also too noisy to tell apart from zero, since resampling the 80 deals puts it anywhere between -4.0% and +0.7%. Their median was +0.6%.
• Volume tells the story the price doesn't. For those 80 deals, volume on the first three days shares could be sold was a median 1.34 times the stock's average in the month before, against 0.90 in the same window a month earlier. On the Capital Market and NYSE American, where most holders never signed, it was 0.59 against 0.57. Nothing happened.

Day 0 is the first trading day after the lock-up period ends. The small-company line sits below 100% throughout because their volume comes in bursts: on most days it is well below its own average.

The two results fit together. The date is printed in the prospectus months in advance, so anyone who expects selling can trade ahead of it, and the price has no reason to move on the day. Volume can't be moved ahead: the released shares are new supply, and they change hands when they're allowed to. Where the lock-up never covered most holders, there is little left to release.

Few insiders rush to sell

An affiliate of a company, an officer, director or large holder, who plans to sell more than 5,000 shares or $50,000 of stock in three months under Rule 144 has to file a Form 144 first, so the filings show when insiders intend to sell. During a lock-up they mostly can't, and across the 322 standard lock-ups that ended before August 2026 we found 3 filings in the last 30 days. After the end, the filings came at a steady pace rather than a rush: 44 in the first ten days, 68 in the next twenty and 140 in the thirty after that, three to five a day across all 322 companies.

Most companies saw none. Only 32 of the 322 had a Form 144 filed in the month after their lock-up ended. The filings cluster in the bigger deals: 24 of the 122 companies on the larger exchanges had at least one, against 6 of the 166 on the Capital Market. A filing is a notice, not a sale, and the sale itself shows up later on a Form 4. The pattern suggests that in most IPOs, the insiders who signed weren't waiting for day 181 to sell.

Lock-up expiration dates, November 2026 to January 2027

43 IPOs in our sample reach the end of their lock-up between October 28, 2026 and January 31, 2027. These are the 12 with the most trading, by average daily value over the last 20 sessions. Day 180 is 180 days after the prospectus date; depending on the wording, shares can be sold that day or from the next trading day.

SpaceX's December release is the larger of two possible amounts. The prospectus would have released an extra 455.8 million shares after the first earnings report if the stock had closed at least 30% above the $135 IPO price on five of the ten trading days ending that day. It closed between $108.37 and $125.33 in that stretch, so those shares wait for December 8.

The same dates as short answers, from the prospectuses:

• When does the SpaceX lock-up end? The 180-day lock-up ends on December 8, 2026, with a last tranche of up to 797.6 million shares. An extended lock-up on about 60% of the company, including every share Elon Musk owns, releases in stages until mid-2027.
• When does the Cerebras lock-up expire? Day 180 is November 9, 2026. Employee tranches were released in May, and an earnings trigger could only have moved the rest earlier; Cerebras has scheduled its September-quarter results for November 12.
• When does the Fervo Energy lock-up expire? Day 180 is Sunday, November 8, 2026, so the first trading day after it is Monday, November 9.
• When does the Quantinuum lock-up expire? On November 30, 2026, 180 days after its June 3 prospectus. It binds holders of 1% or more.
• When does the Jersey Mike's lock-up expire? On January 25, 2027, 180 days after its July 29, 2026 prospectus.

How to check a company

The terms are in the final prospectus, the 424B4 filed a day or two after pricing. On a company's Equibles page, open Documents, for example Cerebras's filings, find the 424B4, and search it for "lock-up". Three things decide what the date means:

• The parties. "Holders of substantially all of our outstanding securities" means the date matters. "Holders of 5% or more" means most small holders could already sell.
• The start. Most periods run from the date of the prospectus, some from the closing of the offering, a few from the effective date of the registration statement.
• The exceptions. Search for "earlier of", "release" and "trading day". An earnings or price trigger can move the real date months ahead of day 180.

After the date, Form 144 filings show which affiliates plan to sell, and the company's Documents page lists them.

How we built this

• Sample. Every final prospectus (Form 424B4) filed from January 1, 2024 to October 26, 2026 that says "there has been no public market" for the shares and mentions a lock-up, found with EDGAR full-text search: 824 filings. We dropped 314 from blank-check companies (SIC code 6770) and kept one prospectus per company, the first one, or its repricing if a second followed within 30 days. That left 479 companies. We dropped 25 direct listings and resale registrations, which register existing shareholders' stock and have no underwriter lock-up, and 14 prospectuses where we couldn't read a period. The study covers 440 IPOs.
• Lock-up period. We read every passage that mentions a lock-up, a market stand-off or a restricted period, and took the period stated for directors, officers or shareholders, not the company's own promise not to issue shares, which is often shorter. A hand check of 30 randomly chosen prospectuses found the right period in 29.
• Early release. A prospectus counts if it writes in a release before the main period ends: an earnings-based end date, a price condition, or a schedule of tranches. Discretionary waivers by the underwriters don't count. We read all 26 by hand.
• Who signs. We classified the sentence naming the parties to the lock-up. "Substantially all", "all of our shareholders" or a stated share of 90% or more counts as all; a list of directors, officers and holders above a threshold, or "certain shareholders", counts as insiders and large holders. A hand check of 20 found 19 right.
• Tier. From the prospectus cover. 16 covers named no tier and are left out of the tier comparisons.
• Event test. For each standard lock-up that has ended, day 0 is the first trading day after the end of the period, counted from the prospectus date stated in the document. The three-day window runs from the day before to the day after, so it contains the first sale day under either reading of the end date. Returns are split-adjusted closes against SPY. The comparison window is the same stock 20 trading days earlier. Volume is the average of days 0 to 2 against days 6 to 25 before.
• Form 144. Filings from our Form 144 data matched to each company, by filing date, relative to the day after the end of the lock-up period. The 322 lock-ups ended before August 1, 2026, so each has at least eight weeks of filings after.

Limitations

• We read the prospectus, not the agreements. The lock-up agreements themselves are exhibits we didn't parse, and the prospectus summary can leave out detail.
• Waivers are invisible. Underwriters can release anyone at any time. Some do it with a press release, many quietly. A deal counted as a plain 180 days may have released shares earlier.
• Text classification makes mistakes. The hand checks suggest a few percent of periods and parties are wrong, mostly in small prospectuses with unusual wording.
• Prices cover listed companies. IPOs that have since been delisted often have no price history in our data, so the return and volume figures lean toward survivors.
• The event test isn't a trading signal. Many expiries fall near earnings dates, the windows overlap other news, and the sample is a few hundred deals. The point is the absence of a typical drop, not the size of any average.

Conclusion

The lock-up date that IPO calendars print is still 180 days for nine in ten deals, and for the deals where everyone signed it still does something: the shares come loose and volume jumps, though the price, on the typical stock, doesn't fall. What has changed is around that date. The largest IPOs now hand out shares in earnings-linked and price-linked tranches, so the 180-day mark can be the last release instead of the first, SpaceX turned it into a 15-step schedule, and Anthropic's draft has 23 steps. In the smallest IPOs the lock-up names only the insiders and the largest holders, and the date barely moves trading.

So the useful question about an IPO's lock-up isn't when it ends. It's who signed it and what can end it early, and both answers are in the prospectus.

Sources

• SpaceX, final prospectus (Form 424B4), filed June 12, 2026, accession 0001628280-26-042639; current report on Form 8-K (Item 2.02), October 20, 2026, accession 0001628280-26-066482; Form 8-K (Item 2.02), August 4, 2026, accession 0001628280-26-052515.
• Cerebras Systems, final prospectus (Form 424B4), filed May 14, 2026, accession 0001628280-26-035214.
• Anthropic, PBC, registration statement on Form S-1, filed October 26, 2026, accession 0001628280-26-067168 (preliminary; share counts blank).
• Anthropic, PBC, amendment No. 1 to Form S-1, filed November 2, 2026, accession 0001628280-26-068676 (same lock-up schedule; share counts still blank).
• SEC, Release No. 34-104450 (December 18, 2025), order approving SR-NASDAQ-2025-068 on Nasdaq's initial listing requirements; Nasdaq Rule 5005(a)(46) on unrestricted publicly held shares.
• Laura Casares Field and Gordon Hanka, "The Expiration of IPO Share Lockups", Journal of Finance 56(2), 2001, pages 471 to 500.
• Reddit, final prospectus (Form 424B4), accession 0001628280-24-012380, and Form 8-K (Item 2.02), August 6, 2024; Astera Labs, final prospectus, accession 0001193125-24-073873, and Form 8-K (Item 2.02), August 6, 2024; Circle Internet Group, final prospectus, accession 0001193125-25-135795, and Form 8-K (Item 2.02), November 12, 2025; Figma, final prospectus, accession 0001628280-25-037014, and Form 8-K (Item 2.02), November 5, 2025.
• Bending Spoons, final prospectus, accession 0001104659-26-079884; Quantinuum, final prospectus, accession 0001628280-26-041003; Neutron Holdings, final prospectus, accession 0001628280-26-046635.
• EDGAR full-text search for the sample, and SEC Form 144 filings.