AI bubble fears intensify as Allianz chief economic adviser issues stern warning

AI bubble concerns are intensifying across markets and boardrooms.

Key Points

Summary — AI bubble, valuation risk, and cross-border capital flows

Mohamed El‑Erian (Mùhǎnmòdé · Āi’ěr‑Āilì’ān 穆罕默德·埃尔-埃利安), chief economic adviser at Allianz (Anlian 安联), warned in a recent speech that the current AI (artificial intelligence) mania could produce significant personal losses for investors and a wave of credit incidents.

Major investment banks including Goldman Sachs (Gāoshèng 高盛) and Morgan Stanley (Mógēn Shìdānlì 摩根士丹利) have likewise flagged elevated valuations at leading AI‑related tech firms as a potential trigger for market pullbacks.

At the same time, foreign capital flows out of Asian markets and growing corporate concern about “AI bubble” dynamics are heightening investor unease.

El‑Erian: expect heavy losses and credit accidents

In his speech, El‑Erian said investors should prepare for substantial personal losses in the AI sector and that the market may experience numerous “credit accidents.”

Working with Nobel laureate Michael Spence (Màikè’ěr Sībīnsè 迈克尔·斯宾塞), El‑Erian concluded that the AI boom resembles a “rational bubble.”

That term means the technology’s enormous real potential can justify venture‑style bets for outsized returns, while simultaneously allowing exuberant investment to produce painful failures.

He pointed to parallels with past speculative periods—such as the internet bubble—where companies effectively labeled themselves to attract capital.

Today, that label is “artificial intelligence.”

Even foundational or legacy companies have drawn big investments simply by attaching an AI label, yet “not every company will succeed.”

El‑Erian also called attention to a policy gap in the U.S. compared with countries such as China (Zhōngguó 中国) and the United Arab Emirates (Āliánqíyìlián Émirates 阿联酋).

He warned that the lack of a comprehensive, orderly national program to promote broad AI adoption could prevent productivity gains from being fully realized if rollout is mishandled.

Market reaction: foreign selling in Asia and softer risk appetite

Data compiled by media outlets show notable foreign selling in regional markets this month.

Foreign investors have withdrawn roughly ¥33.6 billion RMB ($4.6 billion USD) from the South Korean stock market.

In Japan, the Japan Exchange Group reported net foreign sales of about ¥16.8 billion RMB ($2.3 billion USD) through November 7.

These flows indicate that foreign appetite for Asian AI‑related stocks is cooling.

Concerns over lofty valuations are spreading, and the market’s reduced odds of a Federal Reserve rate cut in December have further dampened global risk appetite.

Corporate leaders voice similar worries — valuations, revenue gaps, and capex

Executives at technology companies have increasingly warned about overvaluation in AI.

At the Web Summit technology conference in Lisbon, Jarek Kutylowski, CEO of German AI firm DeepL, said the valuations looked exaggerated in places and that there were signs a bubble could be forming.

Picsart CEO Hovhannes Avoyan bluntly noted that many AI companies carry high valuations despite little or no revenue.

That concern is amplified by expectations for massive capital expenditures to support the AI ecosystem.

A report from venture firm Accel estimates that by 2030 newly built AI data‑center capacity could reach 117 gigawatts, implying roughly ¥29.2 trillion RMB ($4.0 trillion USD) in capital expenditure over the coming years.

According to Accel, roughly ¥22.6 trillion RMB ($3.1 trillion USD) of revenue would be needed to offset that investment.

Major corporations and AI leaders have already announced multi‑billion‑dollar deals to expand global data‑center capacity.

Still, some investors and fund managers caution that publicly quoted megabudget numbers may be exaggerated and that the front‑end, product‑facing implications may not justify the hype.

That suggests a brewing “infrastructure” or “capacity” bubble as well.

Even OpenAI’s co‑founder and CEO Sam Altman has acknowledged concerns that data‑center investment may be becoming overheated.

Where opinions diverge: long‑term AI potential vs. near‑term excess

Although worries about a speculative bubble are rising, many tech executives remain bullish on AI’s long‑term potential while acknowledging short‑term risk.

Lyft CEO David Risher summarized the tension: we are in a speculative financial bubble because AI is an extraordinary, revolutionary technology—but that is also a reason not to fall behind.

Executives expect strong enterprise demand for AI through 2026 and beyond.

Kutylowski said that while there’s broad interest and demand, many companies remain in the “struggle to adopt” stage—underlining a gap between headline promises and real, scalable implementation.

What investors and policymakers should watch — pragmatic signals and early warnings

• Valuation concentration. Monitor whether a handful of firms (for example, NVIDIA (Yīngwéidá 英伟达, NVIDIA)) are disproportionately driving market gains.
• Capital expenditure vs. demand. Track whether announced data‑center and infrastructure spending is backed by sustainable revenue growth—or whether it becomes a speculative capacity build‑up.
• Policy and rollout. Watch for national or industry‑level plans to manage AI deployment in the workforce; poor rollout could cap productivity gains.
• Risk‑management signals. Larger signs of credit stress or increasing defaults in related lending could confirm El‑Erian’s warning about “credit accidents.”

Note on currency conversions: dollar‑to‑RMB conversions in this article use an approximate rate of 1 USD ≈ ¥7.3 RMB (exchange rates fluctuate; readers should check live rates for precise calculations).

Examples above are rounded for clarity.

Quick takeaways for investors, founders, and operators

• If you’re an investor: watch concentration risk and capex-to-revenue math before piling into high‑multiple AI names.
• If you’re a founder: be honest about revenue traction and unit economics when you attach the AI label to your product.
• If you’re a policymaker: consider coordinated national plans to manage deployment and workforce integration so productivity gains are actually realized.

This piece preserves the key facts, figures, and warnings that are shaping the current debate over the AI bubble.

References

• “AI泡沫”担忧愈发密集 安联首席经济顾问发出严厉警告 – 东方财富
• Allianz official site – Allianz
• Accel (venture firm) – Accel

AI bubble