TOKYO – After two years of near-vertical capital spending, euphoric earnings calls, and GPU shortages that became their own economic indicator, the mood toward AI has dramatically shifted. Hyperscalers, model makers and investors are now asking the same uncomfortable question: did the future arrive too fast?
The recent wave of slowdown warnings isn’t just caution — it’s the first real test of whether the artificial intelligence trade is a revolution, a slow-motion bubble or something in between. Every boom eventually reaches a moment when optimism stops compounding and starts questioning itself. Many now believe AI has hit that inflection point, the place where booms either harden into durable growth or crack under the weight of their own expectations.
“There are plenty of signs that we are now in the late stages of a bubble in AI,” says John Higgins, economist at Capital Economics, who argues. He argues the epicenter of any bust would likely be the US stock market, where most of key indicators sit at or near levels that have historically preceded market peaks.
Adding to the unease, Higgins doubts markets will get the kind of shock-absorbing rally in US Treasuries that followed the dot-com crash of the early 2000s. With the US national debt recently topping US$40 trillion and President Donald Trump weighing on the Federal Reserve’s independence, confidence in Treasuries and the dollar is harder to come by.
Meanwhile, “AI fatigue,” as Wolfe Research puts it, appears to be spreading. “If this does lead to a slowdown and a rethink of AI spending, that will have ramifications for the economy and some important sectors of the stock market, because essentially we’ve been running hot based on AI spending,” says Steve Sosnick, analyst at Interactive Brokers.
Old-economy variables could compound the problem. Ruchir Sharma, chairman at Rockefeller International, warns that if the 10-year Treasury yield “decisively breaches” 5% — the top of its range since the dot-com era — it would signal the start of a tighter-money environment in which AI megaprojects become harder to fund.
Debt-servicing costs are also far higher today, he notes, meaning rising public borrowing costs will squeeze other borrowers sooner and hit the “bubbly” AI market harder.
Investor Michael Burry, of “The Big Short” fame, has been warning since May that AI stocks are overpriced, calling the boom “just an asset bubble, plain and simple” and drawing direct comparisons to the dot-com era. More recently, he’s accused corporate leaders of “hype and puffery” designed to mask “real, uncontrollable slowing growth.”
Nobel laureate Paul Krugman takes a different view, describing the AI craze not as a conventional asset bubble but as “a kind of fad, almost a social delusion.”
Not everyone, however, is convinced a pullback is coming. “The key question is whether this is the first sign that the extraordinary AI investment cycle might eventually moderate,” analysts at Deutsche Bank write. “For now, that seems unlikely. The competitive race between companies and countries remains intense, and it’s difficult to imagine firms voluntarily stepping back while rivals continue to push ahead.”
There’s also reason to think the broader economy is more insulated than the stock market suggests. Morgan Stanley chief US economist Michael Gapen estimates AI-related spending has added just 0.4 percentage points to annual GDP since 2025 — modest compared to how reliant economies like South Korea or Taiwan are on the sector.
Bank of America disagrees that a real slowdown is likely, arguing the economic stakes are “too large for any sustained meaningful deceleration,” and that current AI network utilization rates still point to healthy demand.
Some see an upside to a slower pace. Salesforce President Patrick Stokes argues a slowdown could actually help builders “who have a lot of catching up to do” to perfect their AI models. Mohamed El-Erian, chief economic adviser at Allianz, similarly believes a slowdown in AI investment “need not trigger an equivalent slowdown in broader tech-related economic activity.”
Goldman Sachs economist Joseph Briggs adds that the conversation is too narrowly focused on hardware capex. “What I think is less appreciated is how much is being invested in softer forms of investment at the company level” and the data infrastructure and strategy needed to actually deploy AI effectively.
One of the more consequential fault lines is geopolitical: which nation — the US or China — ultimately gets AI right. That question gained urgency after Anthropic CEO Dario Amodei published a lengthy essay over the weekend urging the industry to slow development of the most advanced AI models.
“The people building AI earnestly believe that it could kill us all by the end of the decade. This is not a marketing stunt,” Amodei wrote. “Many executives and senior researchers will couch their phrasing in the press to sound sensible — but I hear the same people express fear privately. No other human activity poses this level of danger.”
Trump sees it differently. “We’re leading China in AI. 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 Sunday, adding that concerns about guardrails are being pushed by “very negative forces” raising issues “that won’t happen.” Vice President JD Vance has gone further, calling government efforts to regulate AI a “Trojan horse” for stifling innovation.
Democrats are pushing back. Former President Barack Obama and former Vice President Kamala Harris this week became the latest prominent Democrats to call for stronger government action.
“We need government — and specifically our leaders in Washington — to get proactive in coming up with concrete proposals, laws and regulations that deal with serious safety concerns, anticipate AI’s impact on jobs and our kids, and make sure that AI’s benefits are widely spread,” Obama said.
Obama added: “We can’t stuff AI back in a box. But we can collectively determine how it’s developed and used, rather than letting AI and its fallout happen to us.”
China, meanwhile, has dismissed American calls to slow down as “fearmongering” lifted from a “Cold War playbook” — and the US political infighting may be giving Beijing room to expand its own ambitions. The US still leads in model capability, research output, and citations, but China is closing the gap faster than expected, despite American export restrictions on advanced chips.
Chinese models from DeepSeek to Moonshot have narrowed the distance considerably, catching Silicon Valley off guard. Stanford’s Institute for Human-Centered Artificial Intelligence notes China is rapidly eroding whatever lead the US once had. Samm Sacks, a senior fellow at Johns Hopkins’ School of Advanced International Studies (SAIS), says “the gap between US and Chinese models is narrow and fragile.”
Within the GOP itself, the divide is notable. House Speaker Mike Johnson is aligned with Trump’s position, but Senate Majority Leader John Thune — long an advocate for reining in AI developers as former chair of the Commerce Committee — is more cautious, putting him at odds with Trump’s dismissal of AI risk as a “hoax.”
For now, Thune is playing it carefully: “There are going to continue to be a lot of conversations about potential paths forward, but getting anything done in the near term is going to be challenging given the other stuff we’re dealing with.”
If the industry does coordinate a real slowdown — slower model releases paired with new guardrails — the market impact could be immediate. As Neil Wilson, a strategist at Saxo Bank, puts it, analysts would be left “scrabbling around to assess likely impact on earnings and valuations.”
Xi Jinping’s China, meanwhile, is pressing ahead with its open‑source approach to AI. And Trump’s laissez‑faire stance may be overlooking the strategic advantages China gains from its deployment‑first strategy.
“China is competing on a different axis, with focus on AI deployment,” argues Reva Goujon of Rhodium Group. “In recognition of its own constraints, due in large part to US export controls, China has focused on deploying and diffusing cheap, near‑frontier open‑weight models to deliver both cost‑efficient AI capabilities and tech sovereignty.”
Goujon adds that when “paired with its footholds in wireless networking and device manufacturing, and a playbook to globalize standards that play to those strengths, China is positioned as the default infrastructure supplier for physical AI and connected industries — especially across aspiring manufacturing hubs in the Global South.”
Trump’s America, meanwhile, is actively alienating the Global South, if not repelling it in real time.
If the AI music really is stopping, East Asian economies are in for a shock. Frenzied demand for AI chips, advanced memory modules and other tech goods is propelling growth in China, Taiwan and South Korea while giving Japan a meaningful tailwind. Semiconductor giants — Samsung and SK Hynix in Korea, TSMC in Taiwan — are minting record profits that underpin their broader markets and economies.
But the AI boom is masking deeper weaknesses: real‑estate crises, record‑low birth rates, slowing domestic demand, US tariff headwinds and fallout from the Iran conflict. As Moody’s economist Stefan Angrick puts it, Asia‑Pacific economies are running at two speeds: “the AI boom is boosting exports and production, while higher inflation and tighter policy drag on growth.”
If today’s AI turbulence is the start of the reckoning many fear, the economic reality would likely hit Asia fastest and hardest.
Follow William Pesek on X at @WilliamPesek
