Anthropic's Strong Quarter Builds an IPO Case the AI Trade Can't Borrow
Anthropic's annualized revenue climbed past roughly $1 billion by the end of 2024, and the more telling detail sits beside that number: the pace of growth has started to cool. The triple-digit surges of 2023 are giving way to something steadier. For a company preparing to go public, steady and large is a workable story. For the wider AI trade, it lands differently.
The AI trade and an AI company are now two separate bets. One can do well while the other stalls, and anyone holding AI exposure through a fund, a handful of tech names or a hoped-for IPO allocation should know which bet they are actually making. In October 2026, that distinction decides more outcomes than headlines about how big the technology is.
A company is priced on its own cash, a theme on its growth rate
An IPO buyer asks a narrow question: will this business turn its revenue into profit over time? Anthropic has real answers. Its Claude models compete directly with OpenAI's GPT and Google's Gemini for business customers, and Amazon and Google have invested billions in it. That backing works like a large co-signer on a loan. It doesn't guarantee anything, but it makes the offering easier to place even when the sector feels shaky.
The broad AI trade asks something else. It's the habit of buying almost anything tied to AI on the assumption that total spending keeps accelerating. Chipmakers, cloud providers and data centre builders are priced on that acceleration. When growth slows, even while revenue keeps rising, the trade loses the assumption holding it up. Nvidia's share price still tends to set the mood for private AI valuations, which tells you how much of the sector moves on one expectation.
The lesson: when a sector headline comes out, ask whether it speaks to a single company's earnings or to the growth rate the whole group is priced on. They often point in opposite directions.
Follow where the money goes next
Anthropic spends a large share of its revenue on cloud computing to run its models, and much of that computing comes from the same companies that invested in it. One firm's revenue becomes another firm's sales. That loop can make the whole system look busier than the end demand underneath it.
End demand is still forming. Much of current business AI spending remains concentrated in pilot projects rather than fully integrated operational systems. Statistics Canada's figures show the hardware side holding up: business investment in computers and peripherals rose 16.7% in the second quarter of 2026, and machinery and equipment investment reached its highest level since the second quarter of 2024. Companies are buying the equipment. Whether pilot projects turn into permanent budgets is what separates durable AI revenue from a temporary spending wave.
There is a dependency risk too. A model company renting computing power from investors who also build competing models can find its costs and priorities shaped by someone else.
The lesson: trace each dollar of AI revenue back to a customer who is paying because it saves money or raises sales. Revenue that only circulates between suppliers deserves a lower multiple.
An IPO prices the risk the early investors are handing off
Anthropic's valuation climbed from $18 billion in early 2024 to $183 billion by September 2025. Some analysts argued that a valuation near the top of that range, set against revenue around $1 billion, was a historically risky price-to-sales ratio. The tech IPO market has grown selective since its quiet stretch in 2023 and 2024, and only firms with large revenue and a believable route to profit are getting through. Anthropic fits that description. New AI rules expected in Canada and the U.S. could still add compliance costs that older revenue models didn't include.
An IPO is often the point where private investors take their gains and public buyers take on the next stage of risk. A solid company can still be a poor purchase at the wrong price.
So before adding AI exposure, ask one question of each holding: would I own this if the AI trade went quiet for two years? If the answer depends on the theme, you own momentum. If it rests on the company's own customers and margins, you own a business.
Anthropic's annualized revenue climbed past roughly $1 billion by the end of 2024, and the more telling detail sits beside that number: the pace of growth has started to cool. The triple-digit surges of 2023 are giving way to something steadier. For a company preparing to go public, steady and large is a workable story. For the wider AI trade, it lands differently.
The AI trade and an AI company are now two separate bets. One can do well while the other stalls, and anyone holding AI exposure through a fund, a handful of tech names or a hoped-for IPO allocation should know which bet they are actually making. In October 2026, that distinction decides more outcomes than headlines about how big the technology is.
A company is priced on its own cash, a theme on its growth rate
An IPO buyer asks a narrow question: will this business turn its revenue into profit over time? Anthropic has real answers. Its Claude models compete directly with OpenAI's GPT and Google's Gemini for business customers, and Amazon and Google have invested billions in it. That backing works like a large co-signer on a loan. It doesn't guarantee anything, but it makes the offering easier to place even when the sector feels shaky.
The broad AI trade asks something else. It's the habit of buying almost anything tied to AI on the assumption that total spending keeps accelerating. Chipmakers, cloud providers and data centre builders are priced on that acceleration. When growth slows, even while revenue keeps rising, the trade loses the assumption holding it up. Nvidia's share price still tends to set the mood for private AI valuations, which tells you how much of the sector moves on one expectation.
The lesson: when a sector headline comes out, ask whether it speaks to a single company's earnings or to the growth rate the whole group is priced on. They often point in opposite directions.
Follow where the money goes next
Anthropic spends a large share of its revenue on cloud computing to run its models, and much of that computing comes from the same companies that invested in it. One firm's revenue becomes another firm's sales. That loop can make the whole system look busier than the end demand underneath it.
End demand is still forming. Much of current business AI spending remains concentrated in pilot projects rather than fully integrated operational systems. Statistics Canada's figures show the hardware side holding up: business investment in computers and peripherals rose 16.7% in the second quarter of 2026, and machinery and equipment investment reached its highest level since the second quarter of 2024. Companies are buying the equipment. Whether pilot projects turn into permanent budgets is what separates durable AI revenue from a temporary spending wave.
There is a dependency risk too. A model company renting computing power from investors who also build competing models can find its costs and priorities shaped by someone else.
The lesson: trace each dollar of AI revenue back to a customer who is paying because it saves money or raises sales. Revenue that only circulates between suppliers deserves a lower multiple.
An IPO prices the risk the early investors are handing off
Anthropic's valuation climbed from $18 billion in early 2024 to $183 billion by September 2025. Some analysts argued that a valuation near the top of that range, set against revenue around $1 billion, was a historically risky price-to-sales ratio. The tech IPO market has grown selective since its quiet stretch in 2023 and 2024, and only firms with large revenue and a believable route to profit are getting through. Anthropic fits that description. New AI rules expected in Canada and the U.S. could still add compliance costs that older revenue models didn't include.
An IPO is often the point where private investors take their gains and public buyers take on the next stage of risk. A solid company can still be a poor purchase at the wrong price.
So before adding AI exposure, ask one question of each holding: would I own this if the AI trade went quiet for two years? If the answer depends on the theme, you own momentum. If it rests on the company's own customers and margins, you own a business.
Sources
Read Next
Cancel or Keep the Card? Closing It Can Spike Your Credit Utilization
The Easiest Tax Return in Canada Could Cost Low-Income Filers the Most
India's Biggest Banking Deal Had No Major Investment Bank Behind It
Affordability Improved on Paper in September. Buyer Confidence Has Yet to Follow