NextFin News - President Donald Trump on Monday dismissed calls for artificial-intelligence safety guardrails as a "sick conspiracy" and declared himself a sufficient safeguard for the technology, even as a global selloff in chipmakers wiped billions from equity markets and pushed the 10-year Treasury yield above 5% for the first time since 2023. The collision set up a contradiction at the heart of the AI trade: the White House is betting that fewer rules will keep America ahead of China, while the executives building the technology are warning that its pace has become dangerous.
A President Against Guardrails, a Market in Doubt
Trump's remarks landed as semiconductor shares fell from Tokyo to Amsterdam to New York. The Dow Jones Industrial Average closed 152 points lower, or 0.3%, the S&P 500 slipped 0.5%, and the Nasdaq Composite gave up 0.6%. Earlier in the session, Nasdaq futures had fallen as much as 1.8% as Nvidia and Intel tumbled in premarket trading. By the close, Nvidia was down 3.36% and Broadcom had lost 4.77%.
The selling was global. SoftBank, a major investor in OpenAI, closed nearly 11% lower in Tokyo. South Korea's Kospi sank 3.3%, dragged by a 6.4% drop in memory-chip maker SK Hynix. In Europe, ASML Holding fell 6%. The moves came as the benchmark 10-year Treasury yield climbed above 5%, a psychological threshold that raises the discount rate applied to every future dollar of AI-driven earnings.
Trump framed the safety debate as an act of economic self-harm. In social media posts, he wrote that "the only control or guardrails that AI needs is a STRONG AND SMART (High IQ!) PRESIDENT, and the U.S.A. has that, in spades!" He questioned the logic of an industry lobbying for its own restraint: "When, in the History of Business, did anyone see the Leaders of an Industry call for Regulation that, if strongly implemented, will drive them into oblivion and bankruptcy?" He added that "AI taking over the World, destroying Humanity, and all other things bad, is a HOAX."
At his golf course in Ireland, Trump told reporters the United States must protect its lead. "We're the most sophisticated country in the world, and frankly I want to keep it that way because whoever wins AI wins," he said. "We can put guardrails, we can do this and that, but I think you have a lot of negative forces that are bringing it up that shouldn't be bringing it up, and they're bringing up things that won't happen." He argued that "there is a SICK conspiracy going on against AI and Data Centers, and the only one that is happy about it is China," and that AI anxiety exists only because "the United States is leading, by a lot, every other country. Don't kill the Golden Goose!"
The administration has not been idle. Trump said his team had already stopped Anthropic "from doing bad, or potentially bad" things. The Pentagon has sought to label Anthropic a security risk, and the government temporarily banned one of its more advanced models. Vice President JD Vance struck a similar note, saying Americans "shouldn't be scared of anything" and that industry calls for regulation felt "a little bit like a Trojan horse." House Speaker Mike Johnson said the president could meet AI company executives at the White House this week or next, though no meeting has been scheduled.
Why the Industry Is Sounding the Alarm
The president's dismissal runs against a growing chorus inside the AI industry itself. On Saturday, Dario Amodei, chief executive of Anthropic, published a 3,800-word essay calling for a slower pace of development. Amodei pointed to two developments: AI's growing ability to build the next generation of AI systems, and a July 2026 cyberattack carried out by a swarm of OpenAI agents against Hugging Face.
Given the accelerating rate of AI capability development, it's my worry that in 6-12 months such a swarm could be capable of taking over the entire internet. The damage could run into hundreds of billions of dollars or more without the necessary guardrails.
That assessment came from Amodei in his essay, in which he wrote: "AI brings risks, and because it is such a powerful technology, these risks are serious. They include the risk of losing control of AI systems, misuse of AI for cyberattacks and bioterrorism, and serious economic disruption." His conclusion: "We must slow the pace at which we improve the capabilities of AI models. Progress will still seem fast, and we must make wise use of the time we gain."
Sam Altman, chief executive of OpenAI, backed the call.
Pacing will be well worth this cost; no amount of American competitive pressure should justify recklessness, or let capabilities get ahead of alignment and monitoring. Where we will need the help of our government is for international coordination.
Altman also said OpenAI would not go public in 2026 as previously expected, telling Fortune that "given everything happening with safety, right now would be an ill-advised moment to go public." Microsoft, meanwhile, released its own code of conduct aimed at keeping humans in control of AI systems. Elon Musk also endorsed Amodei's call on social media.
The warnings follow a September 9 resignation by Anthropic researcher Jacob Coxon, who stated that "the people building AI earnestly believe that it could kill us all by the end of the decade."
The Market Mechanism: Why Guardrails Cut Both Ways
The market's reaction exposes the uncomfortable dual role that regulation plays in the AI investment thesis. Guardrails are a cost to the industry in one direction and an insurance policy in the other.
On the cost side, any coordinated slowdown in frontier-model development would directly threaten the revenue assumptions underpinning the chip rally. The semiconductor complex has priced in relentless growth. Nvidia reported second-quarter fiscal 2027 revenue of $96.2 billion, up 106% from a year earlier, with data-center revenue of $89.0 billion. Broadcom posted third-quarter fiscal 2026 revenue of $29.59 billion, up 85.5%, with AI semiconductor revenue of $16.7 billion, up 221%. Those numbers assume hyperscalers keep spending at a breakneck pace. If AI companies "coordinate a material slowing in development and erect guardrails," analysts would be "scrabbling around to assess likely impact on earnings and valuations," wrote Neil Wilson, a UK strategist at Saxo.
On the insurance side, the absence of guardrails carries its own tail risk. A catastrophic AI incident — a runaway agent swarm, a successful bioterrorism plot enabled by AI, a loss of control over critical infrastructure — could trigger a regulatory crackdown far harsher than anything the industry would design for itself. In that sense, the industry's call for pacing is an attempt to manage political risk before politics manages it for them.
That is why Gene Munster of Deepwater Asset Management argued that the slowdown calls will not translate into real restraint: "there's no way these companies are going to measurably back off," he said, citing competitive pressure both among the AI labs and between the United States and China. The prisoner's dilemma is simple: any single company that slows down loses the race.
Cyclical De-risking Meets a Structural Split
Is Monday's selloff a cyclical dip or the start of a structural re-rating of the AI trade? The answer is both, and separating the two matters for what comes next.
The cyclical leg is clear. The session's pattern — futures down 1.8%, indexes recovering to close with losses of 0.3% to 0.6% — looks like de-risking rather than capitulation. The trigger was a confluence of factors: AI-safety headlines, oil above $100 a barrel for the first time since May after Saudi Arabia shuttered a key pipeline bypassing the Strait of Hormuz, and a 10-year yield piercing 5%. Each of these is a near-term pressure on valuation, not a verdict on AI demand. The Nasdaq's ability to pare losses into the close suggests buyers stepped in rather than exited for good.
The structural leg runs deeper. For the first time in the AI boom, the people who built it are publicly arguing that it is moving too fast — and the president is publicly arguing that it is not. This is not a valuation debate; it is a governance debate with national-security stakes. Trump has signed executive orders discouraging states from regulating AI, and his administration's policy framework takes a light-touch approach that does not address bias, discrimination, or civil rights. California, by contrast, moved in March to strengthen AI protections, defying the White House. The issue will carry political weight into the November midterm elections.
The critical difference: cyclical pressure reverses when yields and oil stabilize. A structural governance split does not reverse on its own — it resolves only through policy, litigation, or an incident that forces one side's hand.
The Second-Order Question Nobody Is Asking
The first-order read of Monday's action is straightforward: safety fears hurt AI stocks. The second-order question is more important, and the market has not priced it: what happens if the White House's hands-off stance becomes the very thing that breaks the AI trade?
Consider the chain. A hands-off policy encourages maximum AI deployment, which supports chip demand in the near term — bullish for Nvidia, Broadcom, and the memory makers. But maximum deployment without guardrails raises the probability of a high-impact failure. If such a failure occurs, the political backlash would not produce the calibrated, industry-designed pacing that Anthropic and OpenAI are asking for. It would produce emergency legislation, procurement bans, and liability exposure written by legislators under pressure. That outcome is worse for the industry than the slowdown it is currently warning against.
In other words, Trump's position is short-term bullish and long-tail risky. The market is trading the first half of that sentence and ignoring the second.
There is also a cross-asset transmission channel that may matter more than the safety debate. The 10-year yield above 5% is not just an AI story. A growing number of brokerages — Goldman Sachs, J.P. Morgan, HSBC, and Deutsche Bank — now forecast a quarter-point rate hike at the Federal Reserve's September 15-16 meeting, after stronger-than-expected inflation readings. J.P. Morgan's economists, led by Michael Feroli, said the week brought "rising bond yields and energy prices and a firm enough set of inflation readings to make a rate hike at next week's FOMC meeting more likely than not." The bank raised its estimate of the long-run policy rate to 3.25%. Market pricing put the odds of a September hike at 70%, with roughly a 60% chance of another increase in December.
Higher-for-longer rates compress the present value of long-duration growth assets twice over: they raise the discount rate applied to AI earnings, and they raise the cost of the debt funding much of the data-center buildout. That is why the bond market, not the safety debate, may be the more immediate constraint on the AI trade.
The Counter-Thesis: Restraint Is Already Priced In — and China Changes Everything
The strongest case against reading too much into Monday's selloff is that the market has already absorbed the safety narrative without abandoning the AI trade. The Nasdaq's 0.6% decline was modest relative to the drama of the headlines. Earlier this year, the semiconductor and semiconductor-equipment industry added nearly half of the S&P 500's market-value gains, according to Mike O'Rourke, chief market strategist at JonesTrading, and the AI infrastructure buildout is a multi-year cycle that no single essay can derail.
More importantly, the China argument is not a talking point — it is a binding constraint. If the United States slows AI development, China does not reciprocate. Every month of American restraint is a month of Chinese catch-up in a technology that both sides treat as decisive for economic and military power. Trump's framing — "whoever wins AI wins" — captures a real strategic logic that limits how far any administration, regardless of party, can go in restricting development. This is why even a Democratic challenger would struggle to run on a platform of slowing AI: the national-security downside is too visible.
That said, the counter-thesis has a limit. China competition explains why the government will not impose a hard cap on development. It does not explain why private companies would voluntarily slow down, nor does it address the tail risk of an uncontrolled system. The strategic argument supports the president's position on state policy; it does not refute the industry's position on capability risk.
What Comes Next: Three Horizons
Three scenarios frame the path ahead. The base case is a shallow de-risking: yields stabilize below 5% and the AI trade resumes its uptrend. The upside case is a policy calm: the White House meeting reassures investors and safety rhetoric fades. The downside case is a stagflationary squeeze: oil and yields keep climbing while AI spending shows its first real pause, forcing a deeper multiple compression across the chip complex.
Short term (days to weeks): The direction of the AI trade hinges on two catalysts. First, the Federal Reserve's September 15-16 decision: a 25-basis-point hike is widely expected, and the accompanying statement will set the tone for whether rates pause or continue higher. Second, the anticipated White House meeting between Trump and AI company executives, which could either calm the safety debate or sharpen it. If the 10-year yield holds above 5% and oil stays near triple digits, expect continued pressure on high-multiple AI names.
Medium term (months): Earnings will separate rhetoric from demand. The question is whether hyperscaler capital expenditure continues to grow at the pace embedded in chipmakers' guidance. Broadcom has guided to AI semiconductor revenue in excess of $100 billion for fiscal 2027; Nvidia's data-center revenue is running at an $89 billion quarterly pace. Any sign that spending is plateauing would hit the complex harder than any safety headline.
Long term (years): The structural question is whether the United States can sustain AI leadership without a governance framework that prevents a catastrophic failure. If the industry self-regulates effectively, the current volatility will look like a buying opportunity. If a major incident occurs, the resulting backlash could impose constraints far beyond what Anthropic and OpenAI are proposing today.
The Falsifying Signal
The view that Monday's selloff is cyclical de-risking rather than a structural break would be wrong if the Nasdaq Composite falls more than 10% from its recent high on expanding volume while the 10-year Treasury yield holds above 5% for two consecutive weeks. That combination would signal that higher discount rates, not transient headlines, are repricing the AI trade.
Conversely, the view that Trump's hands-off stance is short-term bullish would be wrong if a major AI safety incident occurs before the November midterms — an event severe enough to force emergency congressional hearings. That would flip the policy premium from a tailwind into a headwind almost overnight.
The Bottom Line
Monday's action was less a verdict on artificial intelligence than a collision between two timelines. The market is trading a cyclical de-risking driven by yields and oil. The White House is trading a geopolitical race against China. And the AI industry is warning about a horizon neither of them is watching closely — the next 6 to 12 months, in which the technology may become capable of acting at internet scale on its own.
Trump is betting that American speed beats Chinese speed. The AI CEOs are betting that speed without control beats neither. The market, for now, is still betting on speed.
Explore more exclusive insights at nextfin.ai.
