Summary of Key Points
The American AI company Anthropic is secretly applying for an IPO (with a potential listing as early as October) and its valuation has surpassed that of OpenAI (965 billion US dollars). Meanwhile, it is calling on the global community to slow down the development of cutting-edge AI, citing concerns about the social risks associated with "AI self-iteration" (the ability of AI to improve itself recursively). This contradictory approach, combined with Broadcom's underwhelming AI-related revenue figures (a chip giant), has caused a sharp decline in AI-related stocks worldwide (with markets in the US, Japan, and South Korea all suffering). It has also sparked intense debates about whether an AI bubble has formed: some believe it is already present (such as Bridgewater's Ray Dalio), while others argue that AI is already profitable (like NVIDIA's Jensen Huang), and still others point out that this bubble is "profit-driven" (different from previous valuation bubbles).
Breakdown and Interpretation
1. Anthropic's Contradictory Moves: Going Public to Raise Funds vs. Calling for a Pause in AI Development – What’s the Real Motive?
Anthropic’s actions seem contradictory: on one hand, it is eager to go public and raise funds (partnering with Morgan Stanley and Goldman Sachs as underwriters), and on the other hand, it urges a slowdown in AI research and development. Why?
- Are They Really Concerned About AI Safety? The concept of "recursive self-improvement" they mention is crucial – if AI can upgrade itself without human intervention, it could happen within two years. As a company that focuses on "AI safety" (founded by former OpenAI employees), this is indeed their core philosophy.
- Marketing and Regulatory Arbitrage? A professor at Wharton University suggests there may be marketing tactics at play: by highlighting AI risks, they can enhance their image as responsible entities (which is beneficial for their IPO valuation). Additionally, if development slows down globally, companies like Anthropic, which already have mature models (such as Claude), could consolidate their advantages and squeeze out smaller firms or open-source AI initiatives (White House officials have accused them of trying to "regulate to capture the market" and suppress competitors).
- There are direct counterarguments: Former Meta AI scientist Yann LeCun argues that current AI capabilities are still limited, similar to those of a cat, and far from achieving true self-iteration.
2. Why Is the Market Panicking? The Combined Impact of Broadcom’s Performance and Anthropic’s Calls for a Pause
The sharp drop in AI stocks is not accidental; there are two main triggers:
- Broadcom’s Underperformance: Broadcom supplies chips to AI companies, and its second-quarter AI chip revenue fell short of expectations, signaling that the demand in the industry might not be as strong as anticipated. This caused Broadcom’s stock price to plummet by 12.6%, and other AI chip stocks (such as NVIDIA and AMD) also declined.
- Anthropic’s Risk Warnings: The market was already skeptical about an AI bubble, and now that even a major AI player is calling for a slowdown, investors are likely seeking to lock in their profits and shift to more stable sectors (such as energy and industry).
- Chain Reactions: Stock markets in Japan and South Korea were also affected – the Korean KOSPI index fell by 5.54% (with Samsung and SK Hynix losing more than 6%), and the Nikkei 225 fell by 1.31%, as AI-related stocks account for a significant portion of these markets’ performance (the Japanese stock market’s 70% growth came from AI).
3. The Debate About the AI Bubble: Is It a “Phantom” or Real?
The focus of the debate is not whether an AI bubble exists, but what type of bubble it is:
- Bubble Advocates: Ray Dalio of Bridgewater argues that AI companies are overvalued (with book values far exceeding actual cash flows), and there is too much speculation, resembling the internet bubble. Barclays also believes that semiconductor and AI stocks are overheated.
- Anti-Bubble Critics: Jensen Huang of NVIDIA asserts that AI is already generating substantial profits, creating trillions of dollars in value, and those who doubt it are simply crazy.
- A New Perspective on the Bubble: BCA strategists suggest that this bubble is different from previous ones; it is "profit-driven" (AI companies are indeed making money). However, if future AI demand declines, the investments made (such as in data centers and chips) could become excess capacity, dragging down the economy. Moreover, analysts are less likely to lower profit forecasts in advance, making it harder to predict when the bubble will burst.
4. The Secret IPO: Anthropic’s “Safety Strategy”
Anthropic is using a US stock market procedure called a "secret S-1" application for listing. What does this mean?
In simple terms, they submit their documents to the SEC for review in secret and only release the prospectus closer to the actual listing date. The advantage is that if the IPO fails, they avoid exposing internal information (such as financial data or business details) and prevent embarrassment.
- Why This Approach? The AI industry is highly dynamic, and Anthropic may want flexibility: if market sentiment is negative, they can delay the listing; if demand is high, they can proceed quickly. With 65 billion US dollars in funding (and a valuation of 965 billion), they are confident, but they also want to avoid any potential issues during the process.
5. What If the AI Bubble Bursts? Don’t Panic – The Technology Will Not Disappear
Both Dalio and BCA point out that a bubble burst is inevitable, but the impact of AI technology will persist, just as the internet did after its bubble burst.
- Consequences of a Bubble Burst: Many smaller AI companies may fail, but leading firms (such as Anthropic, OpenAI, and NVIDIA) will survive and continue to drive technological progress.
- Impact on Ordinary People: In the short term, there may be fluctuations in tech stocks, but in the long run, AI will continue to integrate into daily life (in areas like work, healthcare, and education), bringing tangible benefits.
In summary, Anthropic’s actions have made the market more aware that AI is not an unstoppable rising star, nor a fleeting phenomenon. Investors should be cautious of potential bubbles, while ordinary people can look forward to the long-term changes that AI will bring.