When the Brakes Are Applied: How the Debate on AI Slowing Down Has Perturbed the Trillions in Value of Chip Giants
Hello everyone, I'm your financial observer. Today, we're not talking about some esoteric academic paper, but about an "earthquake" that's happening in the global capital markets.
In simple terms, someone in the AI industry is saying, "Slow down a bit," and as a result, chip stocks have taken a massive hit.
On September 14th, the semiconductor sectors in both the U.S. and European markets experienced a significant drop. NVIDIA's market value evaporated by $176.6 billion in one night, with giants like Broadcom, AMD, and ASML also seeing substantial declines, especially European chip stocks.
Why? Because several key players in the AI industry (including the CEOs of Anthropic and OpenAI, as well as Elon Musk) have suddenly reached a consensus: AI is developing too fast, and we need to put the brakes on, conduct safety assessments, and establish global regulations.
For chip manufacturers that sell the "shovels" (chips) to the gold miners (AI companies), if the miners suddenly say, "Let's first figure out how to mine safely," can sales of those shovels continue to soar as before? That's the core logic behind this sharp drop in stock prices.
Next, I'll break down the reasons behind this turmoil for you in plain language, from five different perspectives.
---
1. The Trigger: The Big Players Suddenly Unite to Call for a Stop, and the Market Panics
The direct cause of this drop in stock prices wasn't a technical bug or poor financial reports; it was a change in public opinion.
Previously, the mainstream sentiment in the AI community was "Faster! Faster still!" The company that developed the super-intelligent system first would win. But this time, the tide has turned.
- The Lead Player Speaks Out: Amodei, the CEO of Anthropic (the company behind the Claude model), publicly called for controlling the pace of AI development. His proposal was specific: conduct third-party evaluations, collaboratively establish safety standards, and coordinate global regulations.
- Rivals Join Forces: Most surprisingly, OpenAI's CEO, Sam Altman, and Elon Musk, who are usually rivals, aligned on this issue. Altman stated, "It's not yet the time to blindly push forward with model capabilities without considering safety."
- The Whistleblower Effect Within: Before the big players spoke up, Jacob Coxon, a researcher at Anthropic, resigned and warned that AI could "kill us all by the end of this century." Although this sounds like something out of a science fiction movie, several top scientists from OpenAI and Google DeepMind also issued safety warnings shortly after.
In Plain Language: It's like a group of racers on the track, and suddenly the lead racer says, "Wait, there seems to be a problem with the brakes; let's stop to check and discuss the traffic rules." The manufacturers of tires, oil, and cars (chip companies) start to worry: If the racers decide to drive more slowly or even take a break, will their sales decrease?
What the market fears most is not slowing down, but uncertainty. When even the most aggressive AI players start talking about safety and slowing down, investors wonder: Has the era of rapid AI infrastructure construction come to an end?
---
2. The Chain of Events: From "AI Safety" to the Breakpoint in Chip Orders
Many people might ask, what does AI safety have to do with chips? If AI slows down, do we still need chips?
The connection is significant. Here's the logic:
- The Driving Force of AI Development is Computing Power: Training large models requires massive amounts of GPUs (graphics processing units), which are the lifeline for chip manufacturers like NVIDIA.
- Slowing Down Means Fewer Orders: If AI companies slow down on model iteration due to safety assessments and regulatory requirements, their demand for new chips will decrease.
- Potential Reduction in Capital Expenditure: Barclays Bank analyzed that the market is concerned about whether capital expenditure will decrease. Previously, AI companies spent heavily on chips to stay ahead. If safety concerns or increased compliance costs slow down research and development, they might delay or cut their chip purchases.
In Plain Language: Imagine you're renovating your house (training AI models). Before, you bought as many bricks and cement (chips) as possible. Now, the construction authority (safety regulations) says, "Wait, you need to do a safety assessment first, and we need to wait for global standards to be set before you can start." Even though you still need to renovate, the speed of your purchases will definitely slow down.
---
3. NVIDIA's Dilemma: Selling Safety Solutions While Discussing Slowing Down
NVIDIA CEO Jensen Huang's response in this situation was quite nuanced. He didn't completely deny the need to slow down, nor did he fully support it; instead, he presented a mixed approach:
- Acknowledging the Complexity: Huang acknowledged the researchers' concerns and agreed that there are complex reasons behind the call for a slowdown. He didn't directly refute it but pointed out that the current panic created a "false choice" between rapid innovation and AI safety.
- *Subtext*: Don't choose one or the other; safety and speed can be achieved, but it will be more complex.
- Luring New Business: Last week, Huang proposed that cybersecurity will become the next major application for AI. Since AI makes code more vulnerable to attacks and vulnerabilities, companies will need more AI security tools.
- *Subtext*: Even though you're worried about safety, solving these issues will also require more chips. So, don't worry; I can still sell my products.
This approach aims to reassure the market, but investors are more concerned about the overall reduction in demand due to slower development, not just the additional sales from security solutions. After all, the market for security tools is still much smaller than that for general-purpose computing chips.
---
4. Market Sentiment: A Sharp Shift from FOMO to FUD
To understand this drop, we need to look at the context:
- Previous Overexuberance (FOMO - Fear of Missing Out):
- On August 27th, NVIDIA projected a 70% revenue growth for the 2028 fiscal year, and its stock price soared 8.74%, reaching a market value of $5.55 trillion.
- The market sentiment was that AI was the future, and chips were essential; buying was a sure bet.
- Current Panic (FUD - Fear, Uncertainty, Doubt):
- In the following five trading days, NVIDIA's stock price fell by 8.32%.
- On September 14th, as the debate about slowing down intensified, panic reached its peak.
In Plain Language: It's like a roller coaster in the stock market. Before, everyone thought AI was an endless source of profit. Now, when someone says there might be safety issues, the first reaction is to sell.
Such emotional selling often leads to excessive price drops, which may exceed the actual impact of the slowdown debate on fundamentals.
---
5. The Future Outlook: A Long-Term Negative Trend or Just Short-Term Noise?
Finally, let's discuss the implications for individuals, investors, and the AI industry:
- For Chip Manufacturers: It's a short-term setback, but in the long run, safety could become a new driver of growth. If AI development really slows down due to safety concerns, chip manufacturers' growth will be affected. However, if Huang's vision of a new AI security market becomes a reality, they might find new opportunities. But this will take time, and it's difficult to make up for the loss of demand in the short term.
- For the AI Industry: Standardization is inevitable, but slowing down doesn't mean stopping. Establishing safety standards is good for making AI more reliable and socially acceptable. However, coordinating global regulations is a long and difficult process. During this period, AI development may experience a plateau or adjustment, but it won't come to a complete halt.
- For Investors: Be cautious of changing narratives and focus on fundamentals. This incident reminds us that the investment logic for tech stocks is shifting from pure growth to growth plus safety and compliance. Don't just look at stock prices; check whether AI companies' capital expenditures are actually decreasing. If major players like Microsoft, Google, and Meta are still purchasing chips in large quantities, this drop might just be a temporary correction, not a trend reversal.
In Summary:
AI isn't dead; it's just shifting from a sprint to a marathon. For chip manufacturers, runners in the marathon may not need as many high-speed shoes, but they will need more comfortable, durable ones. The market is reevaluating this change, and the process is bound to be turbulent.
Advice for the Public:
If you hold related stocks, there's no need to panic, but be aware that the AI industry's growth myth may be being challenged by safety and regulatory factors. In the future, "safety" will become a key word for the AI industry. Those who can address these security issues will remain leaders.