As Anthropic races toward becoming the fastest-growing company in history, investors are beginning to question its IPO prospects.

Investors are now questioning whether Anthropic can maintain its astonishing growth momentum, as fierce competition, a price-sensitive customer base, and the potential risk that AI could destroy humanity all cast a shadow over Anthropic's highly anticipated initial public offering (IPO). The company's annualized revenue soared to $65 billion in July, and its backers expect that figure to exceed $120 billion by year-end, which would make Anthropic the fastest-growing enterprise in history. However, considerable uncertainty remains over whether it can sustain this growth rate. This is mainly reflected in the huge divergence in investors' expectations for its IPO valuation. Some investors familiar with Anthropic's situation said the company is currently valued at $965 billion, while its post-listing trading valuation could range between $1.5 trillion and $4 trillion a spread roughly equivalent to Amazon's market capitalization. Joey Brookhartz, an analyst at SemiAnalysis focused on AI lab research, noted, "This is astonishing, but it's still too early to say what the final landscape will look like." Potential Investor Concerns One of the underlying issues worrying investors is the strong resurgence of Anthropic's main competitor, OpenAI. According to data from model platform OpenRouter, since OpenAI released GPT 5.6 in July, the company has attracted increasing customer spending and, in terms of spending by dimension, surpassed Anthropic for the first time in more than two and a half years. OpenAI claims that its Astra model, released earlier this month, is the best model currently on the market. In addition, the company has postponed its own IPO plans and is currently in preliminary talks with investors, seeking a new funding round at a valuation of $1.2 trillion. Moreover, both companies face an imminent threat from "open" models AI software whose parameters are public and customizable. Some lower-cost models are gradually approaching the performance levels of OpenAI's and Anthropic's more complex "closed" frontier models, and are continuously gaining market share. Eric Glyman, co-founder and co-CEO of corporate payments company Ramp, pointed out, "This is the first time we've seen real price competition among these labs." He added that Ramp and many of its clients switch between different competing models depending on the specific task, and use routing tools to simplify the process of changing providers, thereby avoiding being "locked in" by model developers. As a result, the company has cut its AI spending by 40%. Glyman vividly analogized, "You don't need to rent a Ferrari to go grocery shopping." Industry executives including Microsoft CEO Satya Nadella have long argued that, as competition drives down prices and users gravitate toward the most cost-effective models, AI models will eventually become a "commodity." In terms of customer retention, however, Anthropic outperforms its competitors. Aleh Tsyvinski, a professor of economics at Yale University, analyzed OpenRouter data and found that 12 months after first using an Anthropic model, 22.5% of users were still using it; by comparison, OpenAI's figure was about 13.2%. Another concern is that advances in AI could trigger a social crisis. This month, departing Anthropic researcher Jacob Coxon said, "The people developing AI genuinely believe it could lead to the extinction of all humans by the end of this century." Afterward, Anthropic CEO Dario Amodei called on AI labs to slow the development of cutting-edge systems to ensure their safety. While this could save Anthropic tens of billions of dollars in training costs, it could also give competitors more time to close the technological gap. Mike Paulus, a former Andreessen Horowitz partner who now invests in Anthropic through his family office, said, "These (lab) CEOs are saying 'we should slow down and be careful,' while the market is saying 'yeah... but the profit motive is just too strong.' Maybe in hindsight, we'll wonder why we didn't take their words seriously." This article is reprinted from "Cailian Press"; GMTEight editor: Yan Wencai.