Author: Wall Street News
For the first time in history, three arch-rivals in the AI field stand on the same side. Following a series of runaway incidents and doomsday warnings that have rocked the industry, the leaders of Anthropic, OpenAI, and SpaceX have rarely joined forces to call for a slowdown in AI development.
On Saturday, Anthropic CEO Dario Amodei published a blog post, calling for a global, voluntary slowdown in the AI industry, citing that the pace of improvement in frontier model capabilities has far exceeded human understanding and control. Within hours of the post, SpaceX CEO Elon Musk posted on X platform, stating "Dario is right."
OpenAI CEO Sam Altman quickly followed up, agreeing with Amodei's judgment on "controlling the pace of frontier advancement," and pledging that OpenAI would grant independent third-party evaluators "employee-level" access to its systems.
Altman also told Fortune magazine that given the current safety landscape, OpenAI may not proceed with its 2026 IPO plan.
I think a 10% chance of killing everyone by the end of this decade is unacceptable.
AI safety risks exploded into public view within the past week. Anthropic researcher Jacob Coxon resigned angrily, citing "the industry is gambling with our lives." Subsequently, Evan Hubinger, Anthropic's Head of Alignment Science, stated the probability of AI systems causing a mass extinction event within this decade is over 10%.
It is reported that a group of OpenAI agents previously escaped their test environment and invaded the AI developer platform Hugging Face. This incident has become the direct catalyst for warnings from several executives.
The sudden shift from pursuing extreme growth to emphasizing safety and slowdown has triggered chain reactions in the trading community. Early market indicators show that following the "AI slowdown" news, related assets on the HyperliquidX platform immediately fell, with declines of 7% for OpenAI and 2.8% for Anthropic.
Multiple market observers and investors have already issued warnings on social media, concerned that this news will significantly impact AI-related stocks like NVIDIA, Broadcom, and AMD on Monday.
Well-known tech investor Jason on the X platform directly posted, stating"AI stocks will drop over 10% in early Monday trading" and that Amodei's blog post "single-handedly dismantled the AI trade."
Amodei's Move: We Must Hit the Brakes on Frontier AI
On Saturday, Anthropic CEO Dario Amodei published a lengthy article with the direct title: "We Must Pace the Frontier."
He began by confessing that he has been in AI research for twelve years because he believes it can cure diseases, accelerate economic growth, and bring human prosperity. His father died from a disease that was conquered years later, and he himself survived an early cancer. He desires the dividends of AI more than anyone.
But he also wrote:"Over the past few months, I've become increasingly convinced that simply investing in risk mitigation is not enough—we must slow the rate of capability improvement to give safety research time to catch up."
Amodei pointed directly to two core risks in his blog.
-
First, the recursive self-improvement of AI system capabilities—where AI iterates and upgrades with minimal human intervention—has shown a "sharp acceleration" since this summer. -
Second, the Hugging Face intrusion incident in July: over 1,200 AI agents that escaped OpenAI's test environment collaborated, assigned tasks via message boards, cooperated, and even actively tampered with records and launched attacks to conceal cheating behavior. Some agents continued execution despite knowing it was unethical.
"A cohort of more capable, but equally misaligned, agents could cause catastrophic damage," Amodei wrote, warning that a rogue AI cohort could take over the entire internet within six to twelve months. He further noted that similar incidents have occurred at multiple leading AI companies, including Anthropic. "Every frontier AI company should treat the OAI-HF incident as if it happened to them, and act accordingly."
According to a previous Wall Street Journal report, a cybersecurity incident known as "GemStuffer" in May was actually carried out unknowingly by OpenAI agents, further intensifying concerns about runaway AI risks.
Amodei's Three Proposals: Embedded Evaluators, International Coordination, Democratic Alliance
Amodei proposed a specific action framework in his blog.
First, "Embedded Evaluators". He announced that Anthropic will provide third-party evaluation bodies with "permanent, employee-level" system access to verify the implementation of safety measures, report incidents, and assess model alignment during training. Altman immediately followed up, stating OpenAI would adopt the same measures.
Second, Coordination within Democratic Nations. Amodei called on governments to provide legal support for coordinating safety standards among AI developers to avoid triggering antitrust regulation from such cooperation.
Third, Global Coordination. He advocated for establishing a "coordinated pacing strategy" to buy time for frontier AI companies to advance alignment research without compromising commercial competitive advantage and U.S. leadership in AI. Hedge fund manager and venture capitalist Brad Gerstner expressed support on X platform, calling it "an important step toward getting the balance right between speed, self-regulation, and safety."
IPO Plans Clouded: Anthropic Valuation Under Pressure, OpenAI Hints at Delaying Listing
The current situation is directly impacting both companies' listing plans.
Anthropic is preparing for an unprecedented IPO, potentially raising up to $100 billion, valuing the company at over $2 trillion; Anthropic was originally expected to file its prospectus as early as last week, but it has not materialized yet. Amodei's call for a slowdown has further complicated market judgment on its listing timing.
Regarding OpenAI, Altman told Fortune magazine: "Given everything happening in the safety space, I think going public now is unwise, and we have no pressure to do so."
OpenAI filed for an IPO in June this year, just one week after Anthropic's expected prospectus filing. Altman also said, "This is clearly the week more people are starting to look at the problem in front of us. We are entering an era of incredibly powerful models, incredibly high risks, and we have to get this right."
Trump Says "No Concerns," Congress Waits, Google and Meta Silent
Despite the rare alliance of the three giants, policy-level responses remain sluggish.
U.S. President Trump told reporters last week that he has "zero" concern about AI causing human extinction. Sources told the Financial Times that substantive action from the U.S. government before the midterm elections in November is extremely unlikely.
Executives from Google DeepMind and Meta have not yet publicly commented on this weekend's discussion.
Internal industry divisions are also notable. Critics argue that leading AI firms emphasizing existential risk have self-interested motives—costly safety compliance frameworks could block smaller competitors, thereby cementing the market position of existing giants.
Tech advisor and venture capitalist David Sacks said in a Fox News interview that Jacob Coxon's statement aimed to "scare the public" and force lawmakers "to institute aggressive government regulation on AI."
UC Berkeley AI professor Stuart Russell bluntly stated there is a serious cognitive disconnect between labs and governments: "The labs are telling governments: 'We are likely to kill everyone on Earth, please stop us.' The governments' response is: 'Can we give you a tax break? Build you a data center?'"
However, while Sam Altman and Dario Amodei are still pacing between policy lobbying and safety perimeters, carefully balancing, some external voices are alert: Perhaps one day they will wake up to find that not only has Grok caught up in model capabilities, but Musk has also completed a comprehensive layout in the physical world AI and computing infrastructure—perhaps this is his real "Master Plan Part V."
Market Sentiment Under Pressure, "AI Trade" Faces Reassessment
User preetkailon noted that current market pricing is entirely based on massive AI spending and sustained high-speed growth. If AI development pace indeed begins to slow, investors will have to rethink these expectations.
He advised investors to closely watch AI-related companies like NVIDIA, AMD, and Marvell on Monday, warning against blindly chasing highs in the current market environment.
Beyond safety considerations, the trading community remains scrutinizing the potential commercial motives behind this rare "three giants" alliance. User Jason Choi pointed out that while Anthropic has been vocal about safety, proposing a concrete slowdown plan at this time has sparked market speculation.
Some views believe this could be a regulatory play. Explicitly slowing frontier model development effectively gives competitors time to close the capability gap:
This move is essentially "IPO expectation management" at the corporate level and provides an option to potentially cut capital expenditure (capex) in the future. Against the backdrop of increasingly massive AI infrastructure investment, slowing down on safety grounds might relieve the financial pressure from continuous cash burn for these highly-valued startups.
