The dazzling allure of artificial intelligence, with its promise of transformative breakthroughs and staggering market valuations, has captivated investors worldwide.
Yet beneath the surface of this trillion-dollar spectacle lies a financial web that is increasingly precarious, with risks that extend far beyond the Silicon Valley boardrooms where the AI revolution was born. Nigel Green, CEO of deVere Group, one of the largest independent financial advisory firms, has sounded a clarion warning: the AI trade is dangerously financing itself in a closed loop, and the real game-changer might be China’s quiet emergence as a formidable competitor.
This week’s violent sell-off in semiconductor stocks on Asian markets has laid bare the fragility underpinning the AI sector’s dazzling growth narrative. South Korea’s Kospi index plunged by as much as 8.1%, dragged down by memory chip giants Samsung Electronics and SK Hynix, both falling over 9%, their lowest levels in months. Japan’s Nikkei 225, heavily weighted with chipmakers, also saw steep declines. These sharp moves are not random market jitters but a reckoning with the circular flow of capital that fuels the AI ecosystem; an ecosystem where the same dollars are counted multiple times as revenue, inflating valuations without corresponding real demand.
At the heart of this circularity is a peculiar financial dance. Nvidia, the world’s leading chipmaker valued at an eye-watering $4.5 trillion, has invested heavily, up to $100 billion, into OpenAI, the AI research lab behind ChatGPT. OpenAI, in turn, pays cloud giants like Oracle and Microsoft for vast computing capacity, which those companies then reinvest into purchasing Nvidia’s chips. The same money cycles through these companies repeatedly, creating an illusion of booming revenue streams. Yet OpenAI is on track to lose roughly $14 billion this year, and Oracle’s backlog of cloud contracts, much of it tied to AI, is north of $600 billion, built heavily on commitments from customers who may not be able to pay.
Would any prudent bank underwrite loans on such terms? Almost certainly not. But when cloaked in the narrative of AI’s boundless potential, Wall Street has embraced this circular financing as growth. The danger, Green argues, is that the market has been blind to the fact that this isn’t organic demand but an accounting trick. When Samsung and SK Hynix lose nearly a tenth of their value in a single session, the market is signaling that confidence in this sector is finally cracking.
More ominously, the challenge for Western investors is not only the internal fragility of this loop but the external threat posed by China. Unlike the tightly knit Silicon Valley cluster, where capital recirculates among a handful of players, Chinese AI labs are deploying frontier open-source models trained on domestic chips. These models deliver comparable performance at a fraction of the cost. Developers and customers are already gravitating toward these alternatives, signaling a shift in the global technology landscape. China is no longer trying to catch up; in some segments, it’s setting the pace and pricing, forcing Silicon Valley to react.
This shift is seismic. Historically, every major technological leap; from electricity to the interne; has reshaped the competitive order, spawning winners, losers, and ruthless new challengers. AI will follow this pattern, and China’s ascent marks a new chapter in the race. Investors who ignore this reality do so at their peril. It’s not a matter of if China will win parts of this race; it’s which parts and how quickly.
Yet Green is clear that this does not mean investors should abandon AI altogether. Rather, it’s a call to refocus on fundamentals. The companies that will endure and thrive are those with real customers paying real money, whose margins are sustainable without relying on supplier loans or circular financing. Their technology must remain viable even if the current flood of financing dries up. This moment of turbulence, marked by loud, leveraged, circular capital flows, is the crucible that separates durable companies from fragile ones.
Investors must now perform a simple but rigorous test on their AI holdings: trace the true origin of revenue, ask whether customers would remain if suppliers stopped funding them, and assess how much of a company’s competitive edge can survive when faced with Chinese competitors delivering 80% of the performance at a fraction of the price. Those who do the hard work to answer these questions will be rewarded. Those who mistake the noise of circularity for genuine growth will pay the price.
The broader financial backdrop underscores the warning. Oracle, a key cloud player in the AI infrastructure chain, reported a staggering $638 billion backlog of AI-related cloud contracts, equivalent to nearly eight years of current revenue. Yet much of this backlog is built on deals with customers, including OpenAI, whose ability to pay remains uncertain. Oracle has raised tens of billions in debt financing to support its cloud expansion, even as it faces intense scrutiny over its rising debt load. The scale and opacity of these commitments spotlight the precariousness of the AI infrastructure trade.
Meanwhile, the semiconductor sector’s recent turmoil in Asia reflects deepening investor doubts. South Korea, home to Samsung and SK Hynix, dominates the global memory chip market, a critical component of AI hardware. The Kospi index’s sharp fall, one of its worst in months, signals growing fears about oversupply and demand sustainability in the semiconductor space. This volatility has ripple effects beyond South Korea, influencing markets and tech valuations worldwide.
Against this backdrop, the Silicon Valley versus China AI competition intensifies. Chinese companies have developed AI models that rival the best U.S. technologies, leveraging open-source frameworks and domestic chip manufacturing to keep costs low. Increasingly, global developers are turning to Chinese AI tools, attracted by their competitive prices and growing capabilities. This dynamic has sparked heated debates within Silicon Valley, with some startups adopting Chinese models and others warning of strategic risks in relying on foreign AI technologies.
China’s strategic investments in AI education, research, and infrastructure have cultivated a robust domestic ecosystem. This has enabled Chinese firms not only to catch up but to innovate independently, challenging U.S. dominance in key AI segments. While the United States remains the global leader in AI research and private investment, the gap is narrowing, and China is shaping the rules in parts of the market.
The AI boom’s current financial underpinnings, as Nigel Green highlights, resemble a house of cards propped up by circular financing loops and optimistic valuations. The real test for investors is to look beyond the hype and assess which companies have solid financial foundations and genuine market demand. Those that pass this test will emerge as the true winners in the AI era.
The AI trade is not just about technology. It’s a financial story with profound implications for global markets and geopolitical competition. The fragile confidence in semiconductor stocks, the ballooning cloud backlogs, and China’s strategic rise all point to a pivotal moment. Investors who recognize these forces and act with prudence will navigate the storm. Those who don’t risk being swept away by the tides of hype and circularity. The AI revolution is real, but its financial fabric is woven with threads that must be carefully unraveled. Only then can the true winners be distinguished from the rest.
For investors, the advice is clear: stop chasing the story and start scrutinizing the balance sheets. The next chapter of AI will be written not by those who cheer the circularity as growth, but by those who see through it and invest in resilience.