Anthropic Filed for an IPO. Here Is What the 25 Biggest Tech IPOs Since 2012 Did on Day One, One Month In, and a Year Later

Anthropic filed to go public today, October 26, 2026, on the Nasdaq under the ticker ANT, with $16.3 billion of revenue in the first half of 2026 in the filing and the share count and price range left blank, as they always are at this stage. The listing is reported for mid-November. If it prices anywhere near the figures being reported, it will be the largest technology IPO ever, so the question people are typing into search is the practical one: what happens to a stock like this on day one, and after?

We looked at the 25 largest US technology IPOs and direct listings since Facebook in 2012, from the offer price through the first day, the first month and the first year. The pattern is consistent enough to state plainly. The median first-day close was +37.4% above the offer price, and 23 of 25 closed above it. But the median first-day open was already +38.2% above the offer, so almost all of that gain went to people who were allocated shares before trading started. 14 of the 25 closed the first day below the price they opened at. A year after the first close, the median return was -3.5%, only 11 of 25 were higher, and 10 of 25 were trading below the price the IPO was sold at.

Day one: the pop is real, and it happens before you can buy

The first trade in a hot IPO prints well above the offer price because the offer was sold the night before to institutions, and the opening auction is where everyone else bids for it. On this list the opening print was a median +38.2% above the offer. From that opening print to the first close, the median move was -1.0%: 14 of 25 finished the day below where they opened. Snowflake opened 104% above its offer and closed 112% above it; Facebook opened 11% above and closed 0.6% above, after the underwriters defended the price all afternoon.

The biggest day-one closes against the offer: Figma (2025) +250%, Circle (2025) +169%, Airbnb (2020) +113%, Snowflake (2020) +112%, DoorDash (2020) +86%, Zoom (2019) +72%, CrowdStrike (2019) +71%. The smallest: Uber (2019) -8%, CoreWeave (2025) flat, Facebook (2012) +0.6%, Klarna (2025) +15%, Lyft (2019) +9%.

One month and one year: the median goes nowhere

Measured from the first close, the price most people could actually have paid, the median return after one month was -6.0% and only 11 of 25 were higher. After one year the median was -3.5%, with 11 of 25 higher. Buying at the opening print and holding a year was worse: a median -9.4%, higher in 12 of 25 cases. The lockup expiry, typically 180 days after the listing, sits inside that window, and so does the first full earnings cycle as a public company.

The spread is enormous, which is the point. Five of these roughly doubled or better in their first year as public companies and five lost two-thirds or more:

• Roku (2017): offer $14, first close +67.9%, one year later +210.8%.
• Palantir (2020): offer $7.25, first close +31.0%, one year later +153.1%.
• Reddit (2024): offer $34, first close +48.4%, one year later +148.7%.
• Zoom (2019): offer $36, first close +72.2%, one year later +140.3%.
• Datadog (2019): offer $27, first close +39.1%, one year later +128.2%.
• Figma (2025): offer $33, first close +250.0%, one year later -78.4%.
• Klarna (2025): offer $40, first close +14.5%, one year later -69.8%.
• Rivian (2021): offer $78, first close +29.1%, one year later -67.3%.
• Lyft (2019): offer $72, first close +8.7%, one year later -65.6%.
• Coinbase (2021): offer $250, first close +31.3%, one year later -54.4%.

A first-day pop did not predict the year. The 8 listings that closed day one 50% or more above the offer had a median one-year return of +26.0%, higher in 5 of 8; the 17 with smaller pops had a median of -26.4%, higher in 6 of 17. Small samples, but nothing in them says the size of the opening cheer tells you anything about the year.

Ten of twenty-five were below their IPO price a year later: Facebook, Alibaba, Lyft, Pinterest, Uber, Coinbase, Rivian, Chime, Figma, Klarna. That list includes two of the three most anticipated listings of their years, Facebook and Uber, and three of the five 2025 listings.

What is different about Anthropic, and what is not

• Size. The reported figures would make this the largest tech IPO on record, larger than Alibaba's $25 billion in 2014. Larger deals tend to be priced more carefully and pop less, because there are fewer buyers who can move the price on day one. Facebook, Uber and Alibaba were the three largest on this list; they closed day one at +0.6%, -8% and +38%.
• The filing numbers. $16.3 billion of revenue in six months is more than any company on this list had at its IPO. The losses reported alongside it are also larger than any. The S-1 is the document to read, not the coverage; Anthropic's own announcement links it, and NPR's summary covers what is and is not in it.
• Timing. The listing lands in the same month as a Fed decision and the heaviest earnings weeks of the quarter, which the Big Tech earnings data covers. New listings trade on the market's mood more than on their own numbers in the first weeks.
• What is not different: the mechanics. The offer price goes to allocated buyers, the opening auction sets the price for everyone else, and the record above is the record of buying at that opening price.

What this means if you trade it

• The first print is not a price you chose. Market orders in an opening auction fill wherever the auction clears. If you are going to buy on day one, a limit order is the only way to decide the price.
• The first week is where day-one buyers find out. 14 of 25 closed below the open on day one, and the median one-month return from the first close was -6.0%. Position size and a stop you set in advance matter more here than anywhere, because there is no trading history to size a stop against.
• Nothing systematic. JorgAI's auto-trader does not buy new listings: its scanners need price history to score a setup, and a stock with three days of data has none. Members can trade an IPO by hand in their own account; the automated rules, stops and daily loss limit apply to the automated trades only. The start page covers how the rules work.

A disclosure. JorgAI's AI analysis runs on Anthropic's models. That is a vendor relationship, not a position; we hold no shares and have no view on the offering. This page is about what large tech IPOs have done, not about this one.

Quick answers

When is the Anthropic IPO? The S-1 was filed October 26, 2026, for a Nasdaq listing under ANT. The share count and price range are not yet set; the listing is reported for the week of November 16, which is not confirmed in the filing.

Do tech IPOs go up on the first day? Usually, measured from the offer price: 23 of the 25 largest since 2012 closed above it, by a median +37.4%. Measured from the first trade, which is the price the public can pay, 14 of 25 closed lower on day one.

What happens to IPO stocks after a year? On this list the median return one year after the first close was -3.5%, 11 of 25 were higher, and 10 of 25 traded below their offer price.

Can you buy an IPO at the offer price? Only with an allocation from an underwriter or a broker that distributes shares, which for most retail accounts means a small or zero allocation. Everyone else buys in the opening auction or later.

Data: offer prices from each company's pricing announcement; first-day open and close, one-month and one-year returns from Yahoo Finance daily data (CrowdStrike and Klarna first-day prices from the pricing-day coverage where Yahoo's history was incomplete), measured from the first close and dividend-adjusted; 25 US technology IPOs and direct listings, May 2012 to September 2025. Anthropic filing facts from the company's October 26, 2026 announcement. Past listings are not predictions, and nothing here is a recommendation to buy or sell any security.