Could Anthropic Stock Fall 25%?

AMZNYTD+16.0%SPYYTD+7.9%XLYYTD+0.4%
Analyze AMZN →

Anthropic priced its IPO at $83 a share on Wednesday evening. Its shares start trading today under the ticker ANT. At that price, could the stock fall 25% in its first year?

Image from Pixabay

What Does $83 A Share Buy?

The $83 price is above the $70 to $78 range in this week’s amended filing. Anthropic also added 155 million new shares to the sale. At $83, it is worth about $1.1 trillion on its 13.2 billion shares outstanding (about $1.5 trillion counting options and other dilution).

Anthropic booked $35.3 billion of sales over the last four quarters. That puts the stock at about 31 times sales. Buyers pay about $31 of stock value for each $1 of yearly sales. At the top of the range, $78, it would have been about 29 times. Each 10% rise in the first day’s price would add about three to that multiple.

Which sales figure you use changes the answer a lot. At its October pace, Anthropic would take in $81 billion a year. On that pace, the price is about 14 times sales. On its 2025 full-year sales, the basis of the table below, it is about 240 times. The gap between those numbers is a bet on how long the growth lasts. That is the question in Anthropic IPO: Priced For 2031, Or 2036?

At $83, IPO buyers pay about 31 times sales. If you would rather wait for a lower price, your cash does not have to sit idle. The right price can take months. Meanwhile, the Trefis Income Solution pays a dividend yield about 3x the S&P 500’s. It holds 50 stocks with at least 10 straight years of dividends. See what it pays.

What Did Day-One Buyers Pay Before?

We looked at ten giant US listings, from Google in 2004 to SpaceX in June. For each, we divided its market value at the first day’s close by its sales in the last fiscal year before it listed.

Snowflake closed its first day at 272 times its sales. SpaceX closed at 114 times, Alibaba at 27, Meta at 22 and Google at 19. The table also shows each one’s return over the next 12 months and its lowest close in that year, both from the first close.

IPO (first close) Price ÷ last year’s sales 12 months later Lowest close in year one
Snowflake (Sep 2020) 272x +27% −26%
SpaceX (Jun 2026) 114x −11% (5 months) −33%
Alibaba (Sep 2014) 27x −32% −35%
Meta (May 2012) 22x −34% −54%
Google (Aug 2004) 19x +179% 0%
Airbnb (Dec 2020) 18x +25% −14%
Reddit (Mar 2024) 10x +149% −22%
CoreWeave (Mar 2025) 10x +94% −11%
Uber (May 2019) 6x −24% −64%
Rivian (Nov 2021) no sales −67% −80%

Source: Trefis, prices through Nov 18. Day-one market value is the first close times the shares outstanding at the first quarter-end after listing. Sales are from the last fiscal year reported before listing. Returns and lows cover the 252 trading days, about one year, after the first close (SpaceX: 110 days, to Nov 18, 2026).

Did A Higher Price Mean A Deeper Fall?

Not in a straight line. Across the nine with sales, a higher multiple did not line up well with a deeper fall. The cheapest, Uber at 6 times, fell the furthest: 64%, in the March 2020 crash. Snowflake, at 272 times, still ended its first year up 27%.

But there is a dividing line, at about 20 times sales. Every listing above it closed at least 25% below its first close within a year. The median low, the middle of the group, was 34% below. For the listings at 20 times or less, the median low was 14%.

Day-one multiple Median low in year one Median 12 months later Higher after a year
Above 20x (4 IPOs) −34% −32% 1 of 3
20x or less (5 IPOs) −14% +94% 4 of 5

Source: Trefis, prices through Nov 18. Same measure and window as the table above, from the first close. SpaceX counts in the low column but has not had 12 months yet. Rivian (no sales) is left out.

Nine listings is a small record, and we drew the 20x line after looking at it. So read it as a pattern, not a law.

Among the four above 20x, Meta’s buyers saw the worst of it. Meta was down 54% by Sep 4, 2012, less than four months after it listed. Three years after its first close, Meta was up 111%.

What Separated The Winners?

Growth, more than price. Google and Snowflake both listed near or above the line. Both more than doubled their sales in the fiscal year they listed, up 118% and 124%. Meta’s sales grew 37% and Alibaba’s 45%. Both stocks ended their first year about a third lower.

In this record, a multiple above 20x came with a deep fall. The next year’s sales decided whether the stock came back.

Anthropic is growing far faster. Its sales in the first nine months of 2026 were $33.2 billion. A year earlier, they were $2.5 billion. That is a 1,206% jump. If that pace holds, the multiple shrinks as sales catch up with the price, the way Snowflake’s did.

What Should Not Surprise You?

The data points one way. The four giants that listed above 20 times sales each fell 26% to 54% below their first close within a year. The median fall was 34%. So it should not surprise anyone if ANT closes at least a quarter below today’s first close. That could happen at any point before Nov 19, 2027.

From the $83 offer price, a 34% fall would take the stock to about $55. That is still more than 20 times the last four quarters’ sales.

Two things would change this picture. One is a first public report with sales running well above the September quarter’s pace of $68 billion a year. The other is a first-day close below the $83 offer price.

What Can A Long-Run Investor Do?

Plan your position. Don’t try to predict the price. Whatever today brings, these steps hold up.

  • Pick your price now. Decide what multiple of sales you would pay, then turn it into a share price. At 20 times the last four quarters’ sales, that is about $53 a share. At 25 times, it is about $67. A limit order buys only at your price or lower. It plans for a fall instead of reacting to it.
  • Size it for the fall. Hold only as much as you could watch drop by half without selling. Meta’s day-one buyers had to do just that.
  • Look for cash. Sales are the headline. Cash is the proof. Anthropic burned cash for three years. Then its business brought in $6.6 billion of cash in the first nine months of 2026. In the same nine months, it put $36.5 billion into investments, mostly short-term securities such as Treasuries. It also has $70.7 billion of computing commitments due in the 12 months after September. At each quarterly report, check whether the cash its business makes is catching up with what it has promised to spend.
  • Ask who pays. That does not have to be clear in 2027. But $451.6 billion of those computing commitments fall due before the end of 2030. At some point, it has to become clear who pays: customers, through the cash the business makes, or new investors.
  • Let the record arrive. Watch two dates: the first quarterly report and the end of the lockup. The lockup is the period when early holders cannot sell. It usually ends about six months after listing.

What Should You Ask Next?

  • When does ANT join the big stock indexes, and what will index funds pay for it then?
  • How fast must Anthropic grow to be worth $1 trillion at 20 times sales by 2028?

So What Matters Most In Year One?

Today’s headlines will be about the company. In our read, the record says the price shapes the first year, and growth decides the years after. If you believe in the growth, buy it only at a price you have chosen. And look for cash.

Nobody knows where ANT will close today, or a year from now. A plan that does not ride on one stock’s first year is easier to live with.

Stop Guessing. Start Positioning.

No one can predict surprises like this one. The Trefis High Quality (HQ) Portfolio doesn’t try: it holds 30 quality stocks in equal weight, so no single surprise decides your result. Come see how it works.

See How HQ Works

For education, not a recommendation to buy or sell.

For questions about this article or underlying research, email help@trefis.com.