Two months ago, Elon Musk made a bold claim. (I know. Shocking!) Explaining in its initial public offering (IPO) prospectus why Space Exploration Technologies (SPCX -1.20%) was justified in asking investors for a valuation more than $1.5 trillion, Musk & Co. asserted that, in the not-too-distant future, its products and services would serve a $28.5 trillion market for space, connectivity, and artificial intelligence (AI) services.
And the biggest of these was AI.
According to Musk, AI is a market opportunity of $26.5 trillion.
Image source: Getty Images.
SpaceX argues, Anthropic echoes
Such a gargantuan number obviously stuck with me. And when another AI company -- Anthropic -- announced last week that, in its opinion, the total addressable market (TAM) for AI services could reach $30 trillion, well, that rang a big bell.
Anthropic reported $11.6 billion in revenue in the second quarter (Q2) of 2026, more than doubling year over year. According to The Wall Street Journal, the company earned a "small operating profit" as well. But Anthropic sees even bigger things ahead for it as its TAM swells to $30 trillion and beyond.
So, $26.5 trillion? $30 trillion? These are big numbers, and they're suspiciously close to each other. But that's not the only thing they have in common. Anthropic says it's targeting a TAM comprising "the full scope of work that could be completed with AI models," according to the Journal. And it can reach this TAM if it can "theoretically capture ... 100% market-share."
Emphasis on "theoretically."
But here's the problem: Estimating the size of a company's TAM requires "a bit of guesswork," says the Journal. Rarely does the company approaching an IPO tell you exactly what it includes in its TAM. Even more rarely does it tell you when it expects to achieve the TAM it cites.
Unlike actual market-share reports, says the Journal, TAM estimates are "especially squishy."
Which is another way of saying it's impossible to verify them before the IPO has happened -- by which time it may be too late.
NASDAQ: SPCX
Key Data Points
Examples from history
Need examples? In 2019, ride-share company Uber (UBER -0.26%) told investors that its TAM was $6 trillion. But how much revenue did Uber actually pull in last year? $52 billion. Or about nine-tenths of one percent of what it cited as its TAM.
In 2021, Rivian Automotive (RIVN -1.07%) claimed its TAM was $9 trillion. Last year, Rivian booked less than $5.4 billion in revenue -- less than one-tenth of one percent of the claimed TAM.
In 2023, Instacart (CART -1.26%) targeted a $1.1 trillion TAM in groceries. Amazingly, Instacart was the most accurate of these three companies. Its revenue in 2025 was $3.7 billion, or about 3.4% of its TAM estimate.
What's the upshot for investors?
What lesson should investors draw from the above?
Not that TAM is necessarily complete bunk. I suppose if you look out far enough, a TAM in just about anything can grow into any number you like; it just may take a few decades, or centuries, for that to happen. If a company wants to hypothesize a bright future far in the distance and not pin a precise date to it, then, technically, I suppose there's no reason they shouldn't be allowed to do that.
Your job, however, is to realize that when a company like Uber, Rivian, Instacart -- or SpaceX and Anthropic -- gives you a TAM, it's more of a marketing ploy than a realistic prediction of the near future.
You should also be aware that two things can be true at once. SpaceX and Anthropic may both estimate the AI economy will grow to $30 trillion eventually. They may even be right! But if SpaceX and Anthropic are both targeting this same $30 trillion AI market, neither one of them is going to own all 100% of it. It's much more likely that SpaceX will claim one sliver of the market, Anthropic another -- with rivals such as OpenAI, Microsoft, Meta (META +1.00%), and DeepSeek all claiming their pieces of the pie as well.
Ultimately, no matter how big the TAM, every one of these companies is going to have to share with the others, and no one of them is going to get anywhere near the full $30 trillion. Caveat investor.
