Anthropic President Defends AI Margins : president Daniela Amodei defends them before IPO. Just weeks before Anthropic’s public listing, co-founder Daniela Amodei is stepping into the spotlight to shrug off doubts about the profitability of artificial intelligence. With $47 billion in annualized revenue and an unprecedented compute appetite to fund, the case she is already making to investors tells you exactly which story Anthropic plans to sell on Wall Street.
Key Takeaways
- Anthropic posted $47 billion in annualized revenue in May 2026, up from $9 billion at the end of 2025.
- Daniela Amodei owns the scale of training and inference costs and prefers excess demand to the opposite.
- She argues that enterprises are still only beginning to embed AI across coding, finance, legal and healthcare.
Have an AI Sum Up This Article
ChatGPTAn IPO Built on a $965 Billion Valuation
Anthropic filed its confidential S-1 with the SEC on June 1, putting a real date on a public listing that the market had long suspected. Co-founder and president Daniela Amodei is now stepping forward as the lead voice selling Anthropic’s story to public investors.
The filing confirmed last week (see our coverage of the SEC filing details) came just days after a massive private fundraise. That round, described as greatly oversubscribed, brought in $65 billion in fresh capital at a $965 billion post-money valuation.
The commercial trajectory behind that valuation is almost insolent. Anthropic ended 2025 at roughly $9 billion in annualized revenue. Five months later, that number sits at $47 billion, a fivefold jump within a single half year.
This acceleration comes paired with spending commitments that are just as staggering. The compute deal signed with xAI alone represents $1.25 billion per month, a magnitude very few cloud buyers can absorb without dilution.
It is precisely this gap between large headline revenue and the capital intensity of the business model that becomes the center of gravity of Amodei’s pitch as the roadshow approaches.
The Open Bet on Demand Outrunning Supply
Daniela Amodei does not deny the financial weight of the model. Per her statements, the upfront cost of training and the cost of serving inference remain very high, and she says the right move is to state it plainly to investors rather than dress it up.
On the recurring question of corporate AI budgets, her strategic choice is unambiguous. According to her, Anthropic would much rather sit in a position where demand slightly exceeds capacity to deliver, instead of facing the inverse, which she considers far more dangerous.
This stance directly addresses the signals coming from buyers like Uber, which has publicly acknowledged that not every dollar spent on AI has yet generated the expected productivity gains. The risk of corporate clients trimming AI budgets remains the sector’s main blind spot.
Amodei’s argument inverts that reading. If the returns look uneven, she says it is because enterprises are still at the very beginning of embedding AI into their day-to-day operations, not because the technology is underdelivering.
She explicitly cites four verticals where adoption is only starting: software development, finance, legal and healthcare. Each one carries margins and headcount where every percentage gain captured by AI agents becomes financially meaningful. The fundraising history, including the Series H at $65 billion, shows that private investors are buying that thesis without blinking.
Also on Horizon:
- Anthropic Maps AI Cyber Threats Using MITRE ATT&CK
- MiniMax M3: Chinese Open-Weight Model Matches Top Proprietary AI
- How Hackers Used Meta AI to Take Over Instagram Accounts
What Anthropic Will Actually Sell to Wall Street
In the short term, Amodei’s message targets a very specific audience. IPO analysts and institutional fund managers want answers on both future gross margins and Anthropic’s ability to convert its order book into recurring five-year revenue. For context, see our earlier piece on Horizon: Cerebras IPO Gets $10B in Orders Before the Book Even Opens.
The direction she is laying out clearly prioritizes market capture over optimization. Anthropic accepts a high capital intensity as long as demand growth keeps outpacing the expected decline in inference costs on the hardware side.
In the medium term, the bet rides on growing corporate dependency. Daniela Amodei expects AI to become incorporated into the everyday work of employees, which effectively translates into a recurring per-seat revenue model rather than punctual usage.
If that scenario plays out, the late-2026 IPO window could secure Anthropic enough war chest to absorb several years of operating losses without depending on private financing cycles. That option will not be available to non-listed competitors.
Daniela Amodei’s public push ahead of the official roadshow is therefore less about explaining and more about defusing an anxious narrative around AI profitability. And, upstream, about locking in the frame within which markets will price Anthropic.
Follow the story on Horizon.
