Anthropic's IPO investors talk $1.5T, $500B below plan

Three sources told Reuters that investors at Anthropic's October 14 investor day in San Francisco discussed valuing the company at $1.5 trillion in its IPO, the Telegraph reports. The $500 billion "collapse" is measured against a number nobody has paid yet: the $2 trillion target in Anthropic's leaked IPO prospectus.
That gap comes to 25%. According to Reuters, some current and prospective IPO investors are weighing the lower figure after Anthropic's latest financial disclosures. The prospectus Reuters saw shows Anthropic lost $42 billion in 2025.
Set against what investors have actually paid, the picture flips. Anthropic's last priced round was a $65 billion Series H in May 2026 at a $965 billion post-money valuation, according to Revenue Memo and other trackers. Three months before that, in February, the Series G raised $30 billion at $380 billion. A $1.5 trillion IPO would still sit $535 billion above the Series H mark ($1,500 billion minus $965 billion), roughly 55% higher. Falling to 55% above the last priced round is an odd kind of collapse.
Nothing here is a priced offering either. Revenue Memo says Anthropic submitted a confidential draft IPO filing on June 1, 2026. Both the $2 trillion and the $1.5 trillion are targets and talk. Neither is a price at which shares have changed hands.
The disclosures investors are reacting to are mixed. On October 15 we reported that, per the Financial Times, Anthropic brought in $16.8 billion in the third quarter and says it was profitable. The prospectus put revenue for all of 2025 at nearly $4.6 billion. In August, Reuters reported that the IPO valuation rested on a forecast of $190 billion to $200 billion in revenue by 2028.
On the other side of the ledger, the same prospectus lists at least $518 billion in planned infrastructure spending over the next decade with Google, Amazon, Microsoft and Broadcom, and 80% of it can't be canceled, as we reported on September 30.
Up next
Claude cut a quantum circuit's gate count 17x in 53 hours
"No human input between rounds": Claude Mythos 5 redesigned eight quantum circuits in a loop with an automated evaluator, and some beat the best published designs by an order of magnitude or more. When an agent iterates on your code, is it a closed loop or human review each round?
Related stories
- Anthropic's June credit rating: the top tier of junk
- Anthropic's IPO price talk slips toward $1.5 trillion
- One quarter brought Anthropic $16.8B and, it says, a profit
- Per FT, OpenAI's run rate sits at $50B, below Anthropic's
- OpenAI closes in on Anthropic in OpenRouter coding data
- Anthropic trains AI engineers like medical residents
Comments
No comments yet. Be the first.
Join the conversation
Sign in to leave a comment. It appears after moderation.

