Anthropic is preparing to tell investors that its potential revenue opportunity is more than $30 trillion.
Thirty.
Trillion.
That’s bigger than the entire annual output of the U.S. economy.
It’s roughly a quarter of global GDP.
And according to the Wall Street Journal, Anthropic is getting to that number by looking at the full scope of work that could theoretically be completed using AI models.
Now here’s where this gets interesting.
Anthropic doesn’t currently make anything remotely close to that.
Its Q2 revenue was $11.6 billion.
That’s actually an incredible number. Revenue more than doubled.
But Anthropic is reportedly projecting roughly $190 billion to $200 billion of revenue in 2028.
Do the math.
$30 trillion divided by $200 billion is 150.
So the “potential revenue” number they’re putting in front of investors is about 150 times the company’s projected 2028 revenue.
That’s not a forecast.
It’s a giant theoretical ceiling.
And that’s exactly why I think we should pay attention to it.
Because the AI boom has started producing numbers that sound less like business forecasts and more like economic thought experiments.
SpaceX already did this.
Its IPO materials put its total addressable market at $28.5 trillion, with $26.5 trillion attributed to AI.
Now Anthropic is going even higher.
$30 trillion.
And Reddit immediately noticed something important.
The 191 technology companies in the S&P 1500 collectively generated about $2.4 trillion of revenue last year.
Anthropic’s theoretical TAM is more than 12 times that entire revenue pool.
One commenter pointed out that it’s around 25% of global GDP.
Another joked that Anthropic alone is apparently going to produce the equivalent of the U.S. economy.
Someone else basically said:
If you can make up any number you want for TAM, why stop at $30 trillion?
That’s the part that deserves more attention.
Because technically, Anthropic isn’t saying:
“We will make $30 trillion.”
TAM means something much more extreme.
It asks:
“How much money could theoretically be spent on everything our technology could eventually touch if we captured the entire opportunity?”
That’s a completely different question.
And it can produce almost any number you want if you define the market broadly enough.
If AI eventually touches lawyers, doctors, programmers, accountants, customer service, advertising, logistics, finance, manufacturing, research, entertainment and basically every other occupation, you can start adding up the economic value of all those activities.
Suddenly the TAM becomes enormous.
But here’s the catch.
Economic value isn’t the same thing as AI-company revenue.
If an AI model helps a lawyer produce $1 million worth of legal work, Anthropic doesn’t automatically get $1 million.
If AI helps a company replace 100 employees, Anthropic doesn’t automatically collect those employees’ salaries.
If AI makes a factory more productive, Anthropic doesn’t automatically own the additional output.
Someone has to actually pay Anthropic.
And that payment has to be large enough to support the enormous infrastructure required to deliver the service.
That’s where the $30 trillion story runs into reality.
Anthropic is reportedly targeting a valuation around $2 trillion and could seek to raise as much as $100 billion in an IPO.
At the same time, its projected 2028 revenue is around $200 billion.
So investors are being asked to believe that a company doing $11.6 billion in quarterly revenue today can eventually become a roughly $200 billion annual-revenue company…
while also believing that its theoretical market is $30 trillion.
That’s an enormous gap.
And the Reddit comments picked up on another problem.
Competition.
Claude isn’t operating in a vacuum.
OpenAI has ChatGPT.
Google has Gemini.
Meta has its own models.
xAI has Grok.
Chinese companies are producing increasingly capable models.
And many of these products can perform the same broad categories of work.
One of the better comments in the thread made the argument that if the models become increasingly interchangeable, the competition eventually moves toward price.
That’s a huge distinction.
If ten companies can write your code, answer your customer-service calls, analyze your documents and generate your marketing material, the customer doesn’t necessarily pay all ten.
They shop around.
The price falls.
And AI has already demonstrated how quickly costs can fall when models become more efficient and competition increases.
That’s why a gigantic TAM doesn’t automatically mean gigantic profits.
You can have a $30 trillion market and still have brutal competition over who gets the revenue.
Look at airlines.
Look at restaurants.
Look at smartphones.
Look at cloud computing.
Huge markets don’t guarantee huge margins.
And that’s the part I think gets buried underneath the headline.
Anthropic’s $30 trillion number isn’t evidence that Anthropic will make $30 trillion.
It’s evidence of how broadly the AI industry is now defining what it believes AI could eventually monetize.
That’s much more interesting.
Because we’re reaching the point where the bullish case for AI increasingly depends on AI becoming involved in an enormous percentage of the world’s economic activity.
Not just software.
Not just chatbots.
Not just coding.
Work itself.
And if that’s the bet, then the question becomes much bigger than Anthropic.
Can AI actually capture enough economic value to justify the hundreds of billions being poured into chips, data centers, electricity and models?
Because the infrastructure bill is real.
The capex is real.
The debt is real.
The power demand is real.
The leases are real.
The $30 trillion TAM isn’t.
It’s a theoretical ceiling.
And that’s why I wouldn’t laugh this number off completely.
If AI really does become capable of performing a huge percentage of human economic work, today’s market estimates could eventually look laughably small.
But there’s another possibility.
The industry may be taking the value of all the work AI could theoretically touch and quietly turning it into the amount of revenue AI companies might eventually capture.
Those are not remotely the same thing.
And that’s the question investors should be asking before they throw a $2 trillion valuation at Anthropic.
How much of that $30 trillion actually becomes revenue?
Not theoretically.
Not eventually.
Not if AI replaces half the world’s workforce.
How much money are customers actually willing to hand over?
Because $30 trillion sounds enormous.
But right now, Anthropic is still trying to turn $11.6 billion a quarter into something much, much bigger.
And going from $11.6 billion to $30 trillion isn’t a growth story.
It’s an entirely different economic universe.
Disclaimer: This is not financial advice and is for educational purposes only. Please conduct your own due diligence.