Original post · Original
In its IPO filing, Anthropic tells investors that people inside the company share the worry that advanced AI could pose catastrophic or existential risks to humanity.
In the same filing, it warns that its 1,300% growth rate won’t last and that it can’t promise revenue will keep growing at all.
> “In addition, with the scale of harm that could result from a catastrophic AI-related incident, no single insurer, and potentially the insurance industry in the aggregate, may have sufficient capacity or expertise to underwrite insurance commensurate with such exposure.”
> “Concerns have been raised both within our organization and by other researchers, regulators, policymakers, and industry participants that sufficiently advanced AI systems could, if not developed, deployed, and operated with appropriate safeguards, oversight, and security measures, pose catastrophic or existential risks to humanity, particularly if such systems were misused, replicated, weaponized, or allowed to operate with a high degree of autonomy.”
> “We may be exposed to liability, litigation, and regulatory action arising from harms associated with our AI models, and our contractual protections and insurance may be insufficient, any of which could adversely affect our business, financial condition, and results of operations.”
Anthropic’s S-1 shows how fast its value rose in 2026. Google and Amazon agreed in March and April to buy shares at $259, and weeks later new investors paid $589.
Broadcom, which rents AI chips to Anthropic, has also agreed to finance part of that rent.
> Google and Amazon had the right to buy shares at a fixed $259.
> Amazon’s older loans also turn into shares at a fixed $2.09 to $5.05.
> As Anthropic’s value soared, those fixed-price rights became worth far more, and accounting rules count that rise as a loss for Anthropic: $72.4B plus $9.4B, nearly all of its $80.7B net loss for the first half of 2026. No cash left the company, and the business itself lost only $0.7B.
View this post on X ↗
Anthropic just filed S1 with SEC.
shows $16.3B in 6-month revenue while its operating loss shrank to $693M.
That revenue is 14 times the year-earlier $1.16B, and it’s growing much faster than Anthropic’s biggest operating expense.
Compute and infrastructure cost $8.9B over the 6 months, yet June-quarter revenue rose 1,368% year over year against 300% for compute.
Existing customers supply much of that growth: by September, those outside Claude’s Pro and Max plans spent about 9 times their year-earlier level.
The $80.7B net loss, by contrast, comes mostly from a $72.4B paper charge, because convertible notes issued to Amazon gained value as Anthropic’s valuation rose.
Amazon funded Anthropic partly through loans that later turn into shares, and accounting rules make Anthropic book a loss whenever those future shares become more valuable, which accounts for $72.4B of the $80.7B loss. No cash left the company; the real cost is that Amazon will receive shares worth far more than it paid, which shrinks other shareholders’ slice.
Public buyers get 1 vote per Class A share, while the co-founders’ single Class F share carries 50.1% of total voting power.
The Long-Term Benefit Trust also elects most directors, so Class A holders will have little say over leadership or any takeover.
We are still blank are the price range and total compute commitments running to 2045, the figure that will test whether revenue can keep outpacing compute.
View this post on X ↗