Anthropic Books $16.3B Revenue, $80.7B Loss in S-1 Filing

Anthropic filed its S-1 with the SEC, opening its books for the first time and setting up an IPO roadshow that could begin as soon as November 9 in pursuit of a targeted $1.5 trillion valuation.

The numbers inside the filing are eye-popping in both directions. Anthropic booked an $80.7 billion net loss on $16.3 billion in revenue, though operating losses are falling. Revenue rose 1,300% year over year, up from $1.2 billion. Every employee contributes $6.6 million to revenue.

The customer base is just as striking. Around 250 customers pay Anthropic more than $10 million a year under contract, and more than 10,000 pay at least $100,000. Net revenue retention stands at roughly 900%.

The takeaway: extraordinary growth, paired with losses that far outstrip revenue.

What does the S-1 reveal about the risks?

Plenty. Between five and seven cents of every dollar Anthropic receives comes from users the company doesn't want on its platform — customers who presumably violate its terms of service or usage policies and whom it intends to remove.

Almost half of Anthropic's revenue in the first six months of the year came through cloud marketplaces run by Amazon, Google, and Microsoft. Those three firms are simultaneously Anthropic's computing suppliers, its business partners, and its competitors.

Anthropic itself acknowledges the risks in the S-1, including the possibility that cheaper Chinese models offering similar performance could undercut its business. And the 900% net revenue retention figure, while remarkable, doesn't necessarily show how broadly that growth is distributed or how sustainable it will be.

In short: the filing confirms both the demand and the fragility behind the valuation.

Is the market ready to pay $1.5 trillion?

The IPO will test the market's appetite for AI's promise — what the filing era makes a "put up or shut up" moment for the technology.

"Anthropic's IPO is shaping up to be a major test of investor appetite for AI," says Susannah Streeter, chief investment strategist at Wealth Club, a non-advisory investment service. "Especially as it's arriving at a particularly turbulent time for markets."

The backdrop is uneasy. Private investment in AI firms has increased 48-fold since 2013, with last year's $290 billion up from $131 billion in 2024, according to Our World in Data. But returns are far from certain. This weekend, Torsten Slok, chief economist at Apollo Global Management, warned that AI has "delivered zero financial payoff for corporate America outside of tech." (Slok declined to speak for this story.)

Slok explained that "the AI story will be decided in the S&P 493, not the Mag 7, and so far, there is no evidence that revenue growth and margins are improving outside the tech sector."

Not everyone buys that framing. Tae Kim, who runs the Key Context Substack, calls the expectation of economy-wide financial gains premature. "A ridiculous statement and thesis because it's a very myopic way of thinking about it," he says.

What does the demand side actually look like?

Kim points to hard supply data. The cost of renting Nvidia's H100 GPUs rose nearly 40% between October 2025 and March 2026, according to research firm SemiAnalysis. Micron's high-bandwidth memory output is largely sold out for 2027, with shortages expected to continue through 2028.

"This is a major computing shift," Kim says, comparing AI's arrival to the transition from horses to cars, or from the abacus to spreadsheets on personal computers.

Wall Street analysts sit between the bulls and bears. "There are some concerning—not unexpected—risk factors in the S-1 including compute," writes Dan Ives, partner and senior managing director at Yorkville Ives & Co. But "overall, the numbers are showing massive eye popping growth from a broadening group of customers," Ives adds. "The near-term losses are massive but investors on the IPO will be laser focused on the growth trajectory into 2027 and 2028 as the next phase of this AI revolution takes hold."

The bears have a case, too. "Anthropic's IPO is likely to be a big moment for generative AI," says Gary Marcus, a self-described AI skeptic. "People had originally pegged it at $2 trillion, but that was before the death of tokenmaxxing, the rise of Chinese models, and the deceleration of their rapid growth. It will be very interesting to see if the market is still up for that valuation at the IPO and the days that follow."

Dot-com bubble or enduring shift?

The big question for investors is whether the AI boom echoes the dot-com bubble or marks a lasting technology shift. Kim argues the comparisons are misplaced and that investors are still "suffering from PTSD" after the dot-com crash. This time, he says, AI infrastructure is already heavily utilized and chip supply is constrained by manufacturing capacity — unlike the overbuilt fiber-optic networks of the internet boom.

The test is coming fast, and not just from Anthropic. Rival OpenAI looks likely to launch its own IPO in 2027 at a valuation rivaling Anthropic's. OpenAI is reportedly looking to raise another $30 billion at a $1.25 trillion valuation. The investor pools are likely to overlap, which leaves skeptics worried that an already frothy bubble is about to burst.

Ultimately, the question isn't only whether investors embrace Anthropic's valuation on day one. It's whether the growth behind that valuation can convert into sustainable profits, and whether AI's economic impact spreads beyond the companies selling the technology. Public filings from Anthropic — and potentially OpenAI — will let investors test those assumptions against real financial results.

Kim, for one, isn't bracing for an end. "I don't think this is going to end anytime soon," he says. "I think we're still very early. It's early innings."