What the Anthropic S-1 Tells Us About Its Mega IPO Plans

Anthropic’s S-1 supports the view that its enterprise business is already profitable and scaling efficiently, but its prospective valuation depends on sustaining rapid innovation while absorbing major compute commitments.

• Anthropic reported a 19% adjusted operating margin in Q2 2026 and $3.9 billion of operating cash flow in the first half, including $3.4 billion of free cash flow after approximately $400 million of capital expenditure.
• More than 1,900 customers spend over $1 million annually and roughly 250 spend over $10 million, while the revenue share of the five largest customers is declining, indicating broader enterprise adoption.
• Claude generated 90% of Anthropic’s code contributions in September, and daily code contributed per engineer increased 45-fold from fiscal 2024 to Q3 2026.
• Revenue per dollar of compute rose from $0.40 in Q1 2025 to $2.08 in Q2 2026, a 5.2-fold improvement, while Anthropic held nearly $132 billion in cash and short-term investments against $141 billion raised by June 30.
• The company faces risks from unpredictable model behavior, regulatory and reputational damage, approximately $800 billion of potential compute and infrastructure spending, and the need to keep reinvesting as models released in the prior six months generated over 80% of Q3 revenue.

Takeaway: The filing presents Anthropic as a profitable, cash-rich AI company with improving unit economics and a likely November IPO, but investors must assess whether its valuation and infrastructure commitments can be justified by durable earnings and continued technological leadership.