The most efficient way to stress-test a grand TAM claim is a piece of arithmetic so simple it should embarrass anyone who skips it: take the company’s current revenue, divide it by the market it says it is chasing, and see what percentage falls out.
Anthropic’s annualised revenue run rate hit roughly $65 billion by the end of July 2026. Set that against the $30 trillion total addressable market the company is preparing to present to IPO investors, according to The Wall Street Journal, and the implied penetration rate is 0.22%. Even the company’s own bullish projection for 2028, somewhere between $190 billion and $200 billion, would only nudge that figure to 0.65%.
For perspective, rewind to June and the blockbuster SpaceX IPO. SpaceX went public at a $1.77 trillion valuation carrying its own enormous TAM of $28.5 trillion, a number it described in its S-1 as “the largest actionable total addressable market in human history.” SpaceX’s Q2 revenue annualised to a penetration rate of roughly 0.11%. And buried inside that $28.5 trillion headline was a remarkable detail: $26.5 trillion of it was attributed to AI, with $22.7 trillion assigned specifically to enterprise applications, a category where SpaceX’s xAI division was generating $2.6 billion per quarter while running an operating loss of $1.3 billion.
The message both companies are delivering to investors is identical and unmistakable: we have captured less than one percent of our market, therefore the upside is enormous. They are not wrong that their penetration rates are low. But there is a question that no prospectus will answer: does the market they are describing actually exist as a capturable revenue pool? Or is “$30 trillion of work AI could do” a category error that quietly conflates the total economic value of human labour with the fraction of that value an AI company could realistically charge to partially automate?
The Benchmark That Created a Trap
The timing and sequencing of these claims is not accidental. It reveals a dynamic that will define every AI IPO filing for the next several years.
When SpaceX filed its S-1 in May 2026, the $28.5 trillion TAM drew criticism from valuation experts. NYU finance professor Aswath Damodaran, widely known as “the Dean of Valuation,” said even before the IPO that SpaceX’s AI TAM was “reaching the end of what’s plausible and pushing beyond.” The criticism did not matter. SpaceX raised $86 billion. Institutional investors bought at scale. The IPO was a historic success.
That success quietly constructed a trap for every company filing after it. The new rule is brutally simple: if you file after SpaceX and your TAM is smaller than $28.5 trillion, investors will ask why your opportunity is less attractive. And if your target valuation is higher than SpaceX’s $1.77 trillion, as Anthropic’s roughly $2 trillion target is, you need a TAM that is proportionally larger to make the premium look earned rather than arbitrary.
Anthropic’s $30 trillion is the rational output of that logic. It clears the SpaceX bar by exactly enough to claim the title (“our market is even larger”) without being so inflated that it provokes immediate rejection. The Wall Street Journal reports that Anthropic derived the figure by estimating “the full scope of work that could be completed using AI models,” a definition broad enough to encompass virtually all knowledge work, customer service, software engineering, legal analysis, medical consultation, and administrative coordination on earth.
@shiri_shh on X captured the absurdity with precision: “Anthropic market pitch: $30 TRILLION. Total US GDP: $32 TRILLION. The pre-IPO hype is now officially competing with national GDP numbers.” The comparison is technically imperfect (TAM is global, GDP is national), but the emotional point lands cleanly: these numbers have crossed into a territory where they function as positioning statements rather than market measurements.
And the arms race will not stop here. If SpaceX opened the bidding and Anthropic raised, OpenAI’s filing will almost certainly top both. OpenAI is negotiating to raise as much as $100 billion at a valuation approaching $830 billion, and it would be structurally impossible to present a TAM smaller than Anthropic’s while seeking a competitive valuation. The logical terminus is that every major AI company going public in 2026 and 2027 will claim a TAM somewhere between $25 trillion and $40 trillion, because SpaceX proved the market will absorb those numbers and because claiming anything smaller now reads as an admission of inferiority.
But when everyone owns a market the size of a national economy, the number instantly loses its differentiating power. It becomes a mandatory line item, a figure that bankers insert and sophisticated investors discount. The real valuation, as it always has, falls back on the things that actually move: revenue trajectory, margins, competitive moats, and customer retention.
What the Numbers Look Like Without the Theatre
Strip away the TAM packaging and Anthropic’s case is, on its own merits, genuinely formidable.
Q2 2026 revenue hit $11.6 billion, more than doubling the prior quarter. The annualised run rate reached $65 billion by late July. Enterprise customers spending over $1 million annually crossed 1,000 by April, having doubled from 500 just two months earlier. Claude Code alone is generating a multi-billion dollar revenue stream. Eight of the Fortune 10 are Claude customers. Revenue per monthly active user sits at an estimated $192, compared to roughly $23 for ChatGPT, an eightfold premium that reflects enterprise pricing rather than consumer popularity.
Those metrics, combined with a 2028 revenue projection of $190 to $200 billion, compose a growth story that is extraordinary at any scale. A $2 trillion valuation on $200 billion in projected 2028 revenue implies a forward price-to-sales multiple of 10x, which is aggressive but well within the range that hypergrowth technology companies have sustained during their expansion phases. The fundamentals carry the valuation without assistance.
The $30 trillion TAM adds precisely nothing to that analysis. It is a narrative device engineered to make $2 trillion look modest (“we are capturing 0.065% of the market”), and in that narrow framing it succeeds. But the investors who will determine whether Anthropic’s IPO matches or exceeds SpaceX’s record will base their decisions on the revenue curve, not on whether the theoretical ceiling of “all work AI could do” rounds to $30 trillion or $50 trillion.
For context that the TAM figure quietly obscures: the 191 technology companies in the S&P 1500 generated a combined $2.4 trillion in revenue last year. Anthropic is claiming a market opportunity that is 12.5 times larger than the actual annual output of the entire tech sector. That framing is not dishonest, because TAM by definition measures a theoretical maximum, but it is also not informative in the way that a revenue projection or a competitive analysis is informative. It tells you the ceiling of a ceiling. It does not tell you the shape of the room.
My Take: The Number Is a Vibes Check, Not a Market Analysis
I will say this directly: $30 trillion is not a market size. It is a mood.
SpaceX proved that institutional investors either do not interrogate the plausibility of TAM figures or have collectively agreed to treat them as a necessary fiction that greases the IPO process. Anthropic’s $30 trillion is the rational response to that signal. OpenAI will file at $35 trillion or $40 trillion. The company after them will claim $50 trillion. The number has no natural ceiling because it is measuring a hypothetical, and hypotheticals scale without friction.
The honest version of Anthropic’s TAM slide would read something like this: “Our market is every dollar of knowledge work that can be partially automated. That is a large and growing number that we cannot precisely quantify. Our actual revenue opportunity over the next five years depends on enterprise adoption rates, pricing power, competition from OpenAI and open-source alternatives, and whether the current pace of AI capability improvement continues. We believe that opportunity is substantial.”
That version will never appear in a prospectus, because it does not give investors a number to type into a spreadsheet cell. So instead they get $30 trillion, the figure that fits neatly in the model, produces a valuation that makes everyone in the room comfortable writing a very large cheque, and lets the story continue exactly as planned.
@kimmonismus put the landing question well: “I’m curious to see if that’s possible after the current backlash. In any case, they’ll need a good release before the IPO.” The backlash refers to the export-control disputes and Fable 5 controversies that have complicated Anthropic’s narrative in recent months. Whether investors look past those issues and fixate on the revenue curve, or whether the $30 trillion number becomes a punchline rather than a pitch, will depend on what Anthropic ships between now and October, not on how many zeroes the bankers put on the TAM slide.
FAQs
What is a total addressable market?
TAM is the theoretical maximum annual revenue a company could generate if it captured every possible customer with zero competition. It is not a forecast or a revenue target. Companies include TAM estimates in IPO filings to frame their growth potential for investors.
How does Anthropic’s $30 trillion compare to SpaceX’s $28.5 trillion?
SpaceX claimed $28.5 trillion in its May 2026 S-1, with $26.5 trillion attributed to AI. Anthropic’s $30 trillion tops it by $1.5 trillion. SpaceX IPO’d at $1.77 trillion. Anthropic is targeting roughly $2 trillion.
What is Anthropic’s actual revenue?
Q2 2026 revenue was $11.6 billion, more than double the prior quarter. Annualised run rate hit $65 billion by late July. The company projects $190 to $200 billion in 2028 revenue.
When is Anthropic’s IPO?
The S-1 was filed confidentially in June 2026. The prospectus is expected imminently, with a potential debut as early as September or October 2026. Morgan Stanley, Goldman Sachs, and JPMorgan are managing the deal.