Investors Question Anthropic's Post-IPO Revenue Sustainability
nashnova research
Anthropic's annualized revenue hit $65 billion in July, yet investors' post-IPO valuation estimates span $1.5 trillion to $4 trillion — a gap roughly the size of Amazon's entire market cap, and itself a verdict on whether AI's business model can last.
$65 billion in revenue — so why are investors more nervous, not less?
Anthropic's annualized revenue reached $65 billion in July. Backers predict it will top $120 billion by year-end, potentially making it the fastest-growing company in history.
Yet investors' post-IPO valuation estimates range from $1.5 trillion to $4 trillion — a spread roughly equal to Amazon's entire market cap. This means → the most bullish and most bearish views on the same company differ by nearly 3×.
In plain terms = the bigger and faster the numbers get, the harder it is to judge whether they can last. The disagreement isn't confusion — it's informed doubt.
OpenAI strikes back — is Anthropic's moat deep enough?
After launching GPT 5.6 in July, OpenAI's weekly user spend on routing platform OpenRouter overtook Anthropic's for the first time — OpenAI's first lead on that metric in over two and a half years.
OpenAI claims its newly released Astra is "the best model on the market" and is in early talks for a new funding round at a $1.2 trillion valuation, while delaying its own IPO.
This means → Anthropic isn't facing a retreating rival. It's facing one that has re-concentrated its firepower and shelved its own listing to fight.
A price war arrives — who is driving costs down?
Chinese firms such as DeepSeek and Moonshot, plus open-source models from Meta, are approaching frontier closed-source performance at much lower prices.
Ramp co-founder Eric Glyman calls it "the first real price competition between labs." Ramp now switches between models by task type, cutting its AI spend by 40%.
His analogy is blunt: "You don't need to rent a Ferrari to get groceries." This reflects a shift — enterprise buyers are learning to shop around, and AI labs' ability to lock in customers is weakening.
Retention leads the pack — but can it translate into pricing power?
Yale economist Aleh Tsyvinski analyzed OpenRouter data and found Anthropic's 12-month user retention at 22.5%, versus 13.2% for OpenAI.
The bull case: token usage has expanded 250× since early last year. The market is growing fast enough for both leading labs to benefit.
But former Andreessen Horowitz partner Mike Paulus offers a cooler read: "Every market will be a fight — OpenAI will be there, incumbents will be there, open-weight models will be there. This is not a comfortable monopoly."
What risks lurk beyond the income statement?
Customer concentration: revenue depends heavily on a handful of large clients and cloud partners. If a major customer switches providers, revenue could swing sharply.
Policy risk: potential intervention from the Trump administration remains an open question.
CEO Dario Amodei recently called for slowing frontier AI development. This means → while it could save tens of billions in training costs, it may also give rivals — especially Chinese competitors — a window to close the gap. Anthropic declined to comment.
市场有风险，内容仅供研究参考，不构成投资建议。