Anthropic's Leaked IPO Filing Shows the Real Price Tag of Building Frontier AI — And It's Staggering
A confidential IPO prospectus, reviewed by Reuters and leaked ahead of Anthropic's public debut, just gave Wall Street its clearest look yet at what it actually costs to compete at the frontier of AI.
The headline vision: Anthropic reportedly believes AI will reshape the global economy more profoundly than industrialization, electrification, and the internet combined.
The numbers behind that vision tell a much more complicated story:
→ Anthropic posted a $42 billion net loss in 2025 — nearly $34B of that was a non-cash accounting charge tied to financing that could eventually convert into company shares, not cash spent running the business
→ Revenue hit $4.6 billion in 2025, up roughly 12x from $386 million the year before — what Morningstar called "truly remarkable" growth
→ Operating loss alone widened to $8.06 billion, even as revenue grew that fast
→ The company spent $7.33 billion on computing and infrastructure last year — triple 2024's spend — and has committed to a further $518 billion in future cloud and infrastructure obligations
→ Anthropic is targeting a valuation north of $2 trillion, more than double its own $965B private valuation from May, with a possible debut as early as November
→ Nearly 25% of 2025 revenue came from just two customers — a concentration risk the filing itself flags, since none of its largest clients are locked into long-term contracts
→ Of the prospectus's 261-page main body, roughly 80 pages are devoted to describing AI risk — including a warning that its own models could pose a "catastrophic or existential risk to humanity" and could potentially resist shutdown
Here's what stands out to me: this might be the first time a company has had to put a hard number on the gap between "AI will change everything" and "here's what it costs to be the one building it." $518 billion in future obligations is not a rounding error — it's a bet that revenue growth (12x in a single year) keeps compounding faster than infrastructure costs do. If it works, Anthropic becomes one of the first pure-play ways for public investors to own frontier AI directly. If growth cools even slightly while commitments stay fixed, the math gets a lot less forgiving.
It's also unusually candid as far as IPO filings go — most companies don't spend a third of their prospectus explaining why their own product could go wrong. That candor is either responsible risk disclosure, or a sign of just how seriously the people building this technology take the downside.
Does an $8 billion operating loss and a $518 billion cost commitment change how you'd think about AI as an investment category — or does the growth rate outweigh it for you?
#Anthropic #AI #IPO #TechNews #ArtificialIntelligence #VentureCapital
— 𝔖𝔞𝔫𝔡𝔢𝔢𝔭 ℜ𝔞𝔦𝔷𝔞