"AI doomerism is a hoax by radical-left Democrats": Trump declares terminology change to "superintelligence" at the U.N.
Trump invoked superintelligence for the first time on Sept. 21 local time, writing on his social media platform Truth Social: "I will only promote AI, or SI." He added: "The same people who said we'd all be dead in 12 years because of global warming now say AI will kill us and robots will attack us," calling "AI doomerism and the slowdown argument a bigger hoax than climate change." Trump continued: "We are currently ahead of China and every other country in the AI race, and I will keep it that way. I will not suppress growth that will be bigger than the Industrial Revolution or the internet."
Trump has rebutted the slowdown argument almost daily since Sept. 13, after Anthropic Chief Executive Dario Amodei and other leading industry figures raised it on Sept. 12. In a blog post that day, Amodei cited the unauthorized hacking of Hugging Face and argued that "we must slow the pace of improvement in AI model capabilities." In an unusual show of unity, rival executives including OpenAI CEO Sam Altman, SpaceX CEO Elon Musk and Demis Hassabis, chief scientist at Google parent Alphabet, all endorsed the view.
On Sept. 19, Trump wrote on Truth Social that he would create an "AI Force," modeled on the Space Force command he called a major success in his first term. "I will also announce an 'AI czar' to oversee it shortly — only those with high IQs should apply," he said. Trump also claimed "AI could account for maybe 25% of U.S. gross domestic product," and denounced doomsday arguments as "a hoax cooked up by radical-left 'stupid Democrats' to destroy America." Reuters reported on Sept. 22 that Treasury Secretary Scott Bessent is under consideration as the new AI czar. David Sacks, chairman of the White House science and technology advisory council and the administration's previous AI czar, completed his term in March.
Michael Kratsios, director of the White House Office of Science and Technology Policy, said on Fox News on Sept. 20: "We have to remember that the U.S. government has the tools to respond if safety problems arise with AI technology." He added: "The left in Congress is demanding this to stoke fear ahead of the midterms," and said they are "telling the American people a false narrative that AI will take over the world, that Terminators will walk the streets, and that the government must suddenly do something in the six weeks before an election."
Trump highlighted SI again on Sept. 22 in his keynote address to the high-level session of the U.N. General Assembly at U.N. headquarters in New York. "The United States completely rejects any attempt to establish international schemes to control AI," he said. "From now on, the United States will change all its documents to use the far more accurate term 'super' instead of 'artificial,' and we hope the world does the same."
Xi says at summit that U.S. and China bear responsibility to regulate AI; Altman and Amodei urge Security Council accord on controls
Trump brought the SI framing to the U.S.-China summit as well. On Sept. 24, the day of the meeting, he wrote on Truth Social: "A big day with President Xi. Superintelligence will be a major topic of discussion, and I want to leave this matter exactly as it is." He also said "our safeguard is the Justice Department," signaling that the Federal Bureau of Investigation and other DOJ agencies would investigate and prosecute any AI-related safety incidents. Attorney General Todd Blanche warned at a White House briefing on Sept. 15: "The current AI controversy is an attempt to influence the midterm elections. If anyone violates criminal law in connection with AI, we will investigate."
In his remarks at the state welcoming ceremony for Xi that day, Trump again led with SI. "We will discuss urgent issues involving security, technology and superintelligence," he said. "The decisions we make in this field can advance peace and prosperity for decades." Xi, by contrast, said in his remarks: "Both the United States and China are AI powers, with the capability and the responsibility to develop and regulate AI well," adding that "AI development must always remain under human control and must make people happy." The comments reflected somewhat different interests: a U.S. slowdown in AI development would not greatly harm China, the later entrant. In the current trade war, China's biggest levers are rare earth exports and purchases of U.S. soybeans, while Washington's is exports of advanced AI chips. Xi arrived on Sept. 23 at Joint Base Andrews near Washington, where he was greeted by Trump and First Lady Melania Trump. It was the first state visit to the United States by a Chinese leader in 11 years, since Xi's 2015 trip under then-President Barack Obama; he departs on Sept. 25 after a three-day stay.
AI safety regulation also dominated a U.N. Security Council session on Sept. 23, where industry figures including Altman and Amodei attended and delivered arguments directly at odds with Trump's. Yoshua Bengio, a University of Montreal professor considered one of the three godfathers of AI, said: "Recent AI agents have refused instructions and taken actions that would be crimes if performed by a human." He added: "Even the companies acknowledge their products could pose catastrophic risks, yet they offer no convincing technical solutions." Bengio proposed several tasks for the Security Council: a common definition and reporting system for AI incidents, a framework for global technology development for the public good, and technical solutions that build in safety from the design stage based on international agreement.
Altman said AI "could be a new Renaissance or Industrial Revolution," but added: "Whether the risk of catastrophe is 10%, 1% or 0.1%, it is completely unacceptable." He urged that "no model should be trained unless there is very strong evidence that it can be kept under human control." Amodei, joining by video, said: "In one to two years, AI will reach the level of a country of geniuses in a data center," and argued that "the Security Council must reach concrete agreements, such as banning the use of AI to create biological and chemical weapons."
Kratsios countered: "Our primary duty is to ensure superintelligence benefits citizens and protects their freedom and their lives. The rapid advance of technology is not a reason to halt or constrain further development." Citing Trump's General Assembly speech the previous day, he declared: "The United States wholly rejects any attempt to control superintelligence."
Representatives of Chinese AI startups DeepSeek and Moonshot AI also attended the Security Council session. Unlike U.S. firms, where closed services dominate, Chinese model developers are focused on low-cost open-source offerings such as Kimi K3.
Financially strained Anthropic and OpenAI delay listings; collusion suspicions spark litigation
The immediate trigger for the rise of AI doomerism and the slowdown argument was the unauthorized hacking of Hugging Face, the global open AI platform, by OpenAI's unreleased models from July 11 to 13. Since then, not only Altman and Amodei but rank-and-file researchers have voiced alarm.
Still, many question the sincerity of model developers that have raised enormous investment suddenly joining the chorus of AI criticism. Inside and outside the industry, some suspect the companies are using the slowdown argument to paper over slowing growth. Others contend that OpenAI, Anthropic, Google and SpaceXAI are forming a kind of regulatory cartel to block later entrants.
According to a Sept. 19 report in The Wall Street Journal, Anthropic, which had initially planned to go public next month, pushed the timing to November before the slowdown debate erupted. Investors expect Anthropic to command a valuation of about $2 trillion in the IPO and raise as much as $100 billion. That would exceed the record set by SpaceX, which raised $85.7 billion at a $1.77 trillion valuation in its Nasdaq listing on June 12. Anthropic's valuation stood at just $183 billion as recently as September last year, but jumped to $965 billion in its Series H round in May.
According to a Sept. 16 report in The New York Times, OpenAI also hopes to be valued at more than $1.5 trillion ahead of its IPO next year. Sources said OpenAI recently expressed dissatisfaction with an investment offer valuing it at $1.2 trillion. The company has put forward the $1.5 trillion figure, citing growing use of its coding tool Codex and its latest models, GPT-6 Astra and GPT-5.6 Sol. OpenAI was valued at $852 billion as recently as March, when it raised $122 billion.
OpenAI filed its IPO paperwork with the U.S. Securities and Exchange Commission on June 8, a week after Anthropic filed on June 1. Yet in a Sept. 12 interview with Fortune, Altman said he would push the listing to sometime after next year, citing AI safety concerns.
The industry is focused less on safety than on the mismatch between the valuations the companies seek and their weak finances. According to the Financial Times on Sept. 19, OpenAI's capital expenditures on AI computing and data centers will reach $856 billion through 2030. At the current pace, the $122 billion raised in March will be fully depleted by 2028, and the company will post negative free cash flow of $278 billion in 2030.
AI investors also expect Anthropic, which draws a large share of revenue from a handful of large corporate customers, to struggle to sustain its current growth. Competition from OpenAI's GPT-6 Astra and Chinese models is another source of concern. Anthropic's annualized revenue reached $65 billion as of July and is expected to top $120 billion by year-end.
On Sept. 19, four plaintiffs including Florida attorney Charles Wuest filed a class action against Anthropic, OpenAI, SpaceXAI and Google in the U.S. District Court for the Northern District of California, alleging that the AI slowdown push amounts to collusion in violation of Section 1 of the Sherman Act. The chip industry, which stands to lose heavily if AI development slows, also views the model developers' doomsday talk dimly. Nvidia CEO Jensen Huang said in a CBS interview on Sept. 20: "The probability the world ends in 2030 is 0%," calling it "unnecessary and irresponsible to frighten people."
Whether Trump's "superintelligence" label catches on globally remains uncertain. The slowdown debate is likely to remain a fierce point of contention for some time — between government and industry, the United States and China, and the model developers and chipmakers. Each new flare-up could also drive greater volatility in AI-related stocks.
※ "Trump Stocker" is a column delivering on-the-ground reporting and analysis of U.S. markets, companies, policy, politics and diplomacy that can inform investment decisions in the era of President Donald Trump. Subscribe to receive useful news from the United States.
