appConfigEntityId

bodyType

customContents

entityId

entityUuid

isVideo

articlePageQuery

article

articleSchemaContent

content

hooksArticlePageAdvertisementConfigArticle

hooksArticlePageFloatingButtonConfigArticle

appConfig

filter

json

liveArticle

James David Spellman

Artificial intelligence (AI) has entered its “nasty” phase – the fast-moving, volatile period when economic reality constrains ambitions and markets separate winners from losers.
The capital expenditure boom, accelerated US-China competition and strategic retrenchment threaten to erase as much wealth as the industry’s unprecedented expansion has created. Returns on capital face mounting pressure to deliver outsize performance. Investors are wary and impatient, rotating portfolios towards companies with clearer paths to profitability, as the recent sell-off indicates.
Nowhere is the scale of the transition more apparent than among the hyper-scalers. Amazon, Alphabet, Microsoft and Meta are collectively expected to spend nearly US$700 billion on capital expenditure in 2026 alone.
Recent earnings releases illustrate the divergent consequences. Meta issued a disappointing revenue forecast as it struggled to prove it can monetise AI, while Microsoft reported record and accelerating cloud revenue, which reinforced the “AI trade”.
Analysts estimate that between US$4 trillion and US$8 trillion will be invested globally in AI infrastructure over the next five years – arguably the largest investment cycle ever.
The frenzy is equally under way in China. Chipmaker ChangXin Memory Technologies soared nearly 470 per cent in its recent market debut in Shanghai, the mainland’s biggest initial public offering (IPO) since 2010. At least six Chinese start-up AI companies are reportedly preparing IPO launches by 2027.


Moonshot AI just completed funding that valued the company at about US$35 billion, while ByteDance is reportedly exploring a US$20 billion bond offering. Combined, they signal a profound reallocation in Chinese capital from property to advanced manufacturing, computing infrastructure and AI.
The financial stakes are vast. The world’s biggest publicly traded companies – predominantly AI platforms and the infrastructure providers powering them – account for increasingly large portions of equity market capitalisation. In the US, AI-related companies represent a record 45 per cent of the S&P 500’s value, making markets increasingly dependent on a few firms.
China offers a different picture. Its AI companies account for only about 10 per cent of global AI market capitalisation while generating 16 per cent of industry revenue, according to Goldman Sachs, suggesting they are undervalued relative to their US peers. Technology companies also comprise more than 30 per cent of China’s A-share market, underscoring AI’s deep roots in the economy.
The consequences are racing through global capital markets, reshaping competition and intensifying the US-China technology rivalry. A severe correction could prove to be more damaging than previous downturns because households worldwide are more heavily exposed to equities than ever before, the Bank for International Settlements recently warned.




Credit markets are signalling concerns, causing tech credit premiums to soar. Alphabet’s first negative free-cash flow in 22 years shows the difficulties ahead. The complex matrix of interdependent corporate balance sheets as companies trade data centre leases for hardware purchase commitments, including Chinese companies, creates systemic contagion risks.
In this nasty phase, investors are obsessed with free cash flow, return on invested capital, operating margins, customer retention and pricing power. New entrants are upending long-held assumptions and forcing management to reallocate capital. One estimate suggests that AI revenue must hit US$2.5 trillion per year to recover capital spending, more than the sector’s entire combined revenue today. Cybersecurity threats and AI failure rates add to the risks.
China’s models are accelerating the nasty phase. Moonshot AI’s Kimi K3, DeepSeek-V4 and Alibaba’s Qwen 3.8 Max are approaching the performance of leading US models while requiring less computing power, energy and cost. Their rapid progress is undermining a core assumption underpinning hyperscalers’ investment thesis: ever-larger capital expenditures are the only path to leadership, which provides pricing power and superior returns, since falling behind compounds exponentially.
That assumption drives US companies to believe they can charge premium prices to recover investments. Meanwhile, Chinese models are reportedly available at prices as much as 90 per cent lower than comparable US offerings, raising a fundamental question: will customers continue paying premiums for incremental gains in performance?


The disruption extends beyond pricing. For years, some US developers held that models should remain closed-source or closed-weight to protect intellectual property and lock customers into proprietary ecosystems. Chinese companies, along with Nvidia and Microsoft, have pursued the opposite strategy. Releasing open-weight or open-source models lets users download, adapt and run them on their own infrastructure, providing greater data control, lower costs and more flexibility.
The fibre-optic boom in the 1990s offers clues to what lies ahead. Telecommunications companies laid more fibre than demand required. Prices collapsed. Differentiation among providers mattered little; they were interchangeable. Many went bankrupt. Even so, the infrastructure outlived the bubble and eventually provided the capacity for AI to take off. Customers captured more value than the fibre suppliers.
That is the paradox of the nasty phase: profits become harder for creators to sustain, even as the technology becomes more valuable to those who use it to reinvent industries.


Hong Kong’s ‘AI city brain’ will not tap personal data, breach privacy: John Lee

Huawei unveils latest tech to boost AI power in push to break China’s Nvidia reliance

Hong Kong’s 5-year plan, policy continuity offer investors more stability, experts say

Huawei accused of stealing robot tech to get hands on T-Mobile’s trade secrets

Artificial intelligence (AI)

AI boom enters its ‘nasty’ phase as economic realities set in

A man walks past an electronic screen showing the SK Hynix stock price outside a foreign exchange dealing room at the Hana Bank headquarters in Seoul on July 31. Photo: AFP

America and China can be allies against AI catastrophe

AI safety activists take part in a "Stop the AI Race" protest march in San Francisco on July 11. Photo: AFP

Workers walk inside an AI data centre under construction in the US. Photo: Reuters

US forces launch what the military said was a precision strike against Iran on July 22. Photo: CENTCOM/AFP

Illinois Governor J.B. Pritzker signs the US state’s artificial intelligence safety measures act into law on July 6. Photo: TNS

US President Donald Trump shakes hands with Chinese President Xi Jinping after a visit to the Zhongnanhai Garden in Beijing on May 15. Photo: Reuters

How sell-off at flashy AI-focused US hedge fund is a wake-up call for Chinese investors

Digital screens display trading numbers on the floor of the New York Stock Exchange (NYSE) on July 23, 2026. Photo: AFP

Situational Awareness founder Leopold Aschenbrenner was a former OpenAI researcher. Photo: X.com/leopoldasch

Employees of SK Hynix attend the company’s opening bell ceremony at the Nasdaq market on the day of its IPO in New York City, July 10, 2026. Photo: Reuters

As AI spending soars, can China’s tech giants deliver long-term profits?

Alibaba Cloud signage is pictured during a tour at the Alibaba office in Beijing on April 1, 2026. Photo: AFP

The Tencent booth at the World Artificial Intelligence Conference in Shanghai on July 17, 2026. Photo: AFP

Opinion

World Opinion

Economy

Global Economy

The era of unprecedented expansion is giving way to a volatile period where tech firms’ soaring ambitions feel the squeeze


The era of unprecedented expansion is giving way to a volatile period where tech firms’ soaring ambitions feel the squeeze.


Stocks

Technology

China

United States

Meta Platforms

Microsoft

Amazon

ByteDance

Alibaba

Expresses the views, ideas and conclusions of the author. (Eg. Enough of Crazy Rich Asians, think of crazy poor Hongkongers.)


AI boom enters its ‘nasty’ phase as economic realities set in

.css-1c6uqr6{color:inherit;font-weight:inherit;font-size:inherit;font-family:inherit;line-height:inherit;overflow-wrap:break-word;}The era of unprecedented expansion is giving way to a volatile period where tech firms’ soaring ambitions feel the squeeze

The era of unprecedented expansion is giving way to a volatile period where tech firms’ soaring ambitions feel the squeeze