Wall Street's AI Reckoning: From Honeypot to Financial Contagion?

The hottest stock in the world is suddenly in hot soup.

Nvidia's Jensen Huang, Wall Street's golden boy and the face of the artificial intelligence revolution, has reportedly found himself entangled in an extramarital affair, allegations of a possible Chinese intelligence honeypot and speculation about an impending divorce.

The alleged intelligence connection remains unproven, while divorce speculation is just that — speculation. Nvidia has said it has no reason to believe corporate secrets were compromised.

Ordinarily, a billionaire's marital difficulties would be little more than tabloid entertainment. But when the man concerned heads the world's most closely watched AI chipmaker, questions about corporate security, leadership and market confidence inevitably follow.

Unfortunately for Nvidia, the gossip coincides with something considerably more serious.

Wall Street's Trillion-Dollar Headache

From where we left off on October 24, on Thursday, November 5, the Nasdaq plunged 5.5%, semiconductor shares collapsed by more than 12%, and approximately US$3 trillion disappeared from the S&P 500's market value.

Friday's rebound offered little comfort. Even Monday's recovery left investors questioning the sustainability of the AI investment frenzy.

The Magnificent Seven — Nvidia, Microsoft, Alphabet, Amazon, Meta, Apple and Tesla — have been the principal beneficiaries of Wall Street's technology enthusiasm, although their exposure to AI differs considerably.

The question is no longer whether AI will transform the world. It is whether the astronomical investments being made in anticipation of that transformation can ever generate satisfactory financial returns.

J.P. Morgan estimates global AI capital expenditure at US$900 billion in 2026 and a cumulative US$5.5 trillion through 2030. Bank of America projects annual global data-centre spending could reach US$2.93 trillion by 2030. 

The financing requirements are staggering. Five major hyperscalers — Amazon, Alphabet, Microsoft, Meta and Oracle — could spend close to US$800 billion in 2026. Oracle alone is reportedly planning some US$40 billion in debt and equity financing.

Meanwhile, Anthropic is contemplating an IPO that could raise as much as US$100 billion, while OpenAI is reportedly seeking another US$30 billion in private financing ahead of a possible listing.

These are not small technology ventures seeking start-up capital. They represent an unprecedented appetite for global investment funds. The concern is whether financial markets can continue supplying such enormous sums while investors become increasingly impatient for actual profits.

The danger is shifting from inflated share prices to corporate borrowing, private credit and infrastructure financing. 

Nvidia may continue making extraordinary profits selling chips. But that does not mean its customers will make money operating the data centres housing those chips. A technology can succeed spectacularly while its investors lose fortunes.

When the AI Gold Rush Meets Reality

The global race to build data centres has become almost indiscriminate. Governments compete to attract investment. Technology companies compete to secure computing capacity. Financial institutions compete to finance projects.

Yet few appear willing to contemplate what happens if supply eventually exceeds commercially viable demand. Data centres require enormous investments in land, electricity, cooling systems, networking equipment and advanced chips. Their commercial viability ultimately depends on paying customers, not the number of buildings completed or investment announcements celebrated.

A slowdown in AI spending could leave developers with expensive infrastructure, heavy debts and insufficient revenue. And unlike share prices, debts cannot simply be written down by disappointed investors without consequences for lenders and the wider economy.

Malaysia's RM385 Billion Gamble

Malaysia has enthusiastically joined the race. According to figures disclosed in Parliament on October 28, the country has approved 117 data centres involving RM385.7 billion in investments. Of these, 55 were operational, 19 under construction, 16 had yet to begin construction, and another 27 were at various implementation stages.

The approved investment figure is impressive. But approvals are not expenditure, and completed facilities are not necessarily profitable businesses.

Even more revealing is the electricity requirement. As of June 2026, data centres were drawing approximately 1,256 megawatts from the national grid against maximum contracted demand of 5,650 megawatts.

That gap partly reflects future expansion. Nevertheless, it illustrates the enormous infrastructure capacity being reserved for an industry whose eventual demand remains uncertain.

Meanwhile, Malaysia continues aggressively promoting AI, semiconductors and advanced technology as the country's economic future. There is nothing wrong with embracing new technology. The concern is the apparent imbalance in national economic priorities.

Malaysia already possesses substantial economic strengths in manufacturing, commodities, agriculture, tourism, trade, logistics and established services. These industries generate employment, export earnings, domestic business opportunities and economic activity across the country.

Yet policymakers appear increasingly captivated by the glamour of AI, semiconductors and billion-ringgit technology announcements. The danger is not merely an AI investment correction.

It is that excessive policy attention, infrastructure commitments and financial resources directed towards fashionable technology sectors could come at the expense of improving the competitiveness and productivity of industries that already sustain Malaysia's economy.

Traditional industries are not necessarily sunset industries. Many offer considerable opportunities for modernisation, productivity improvements, downstream development and higher value creation.

An economy does not become advanced simply by attracting advanced technology investments. Nor does economic diversification mean abandoning existing strengths to chase the latest global investment fashion.

The objective should be to strengthen the entire economy, not merely showcase its newest sectors.

The GPU Question Nobody Wants to Ask

There is an intriguing account circulating within Malaysia's data-centre industry. A Malaysian leader reportedly approached Jensen Huang seeking to accelerate the supply of Nvidia GPUs for local data-centre operations. The response, according to the account, was that availability might have to wait until 2035.

The exchange has not been independently verified, but the story raises an uncomfortable question. Malaysia can approve data centres, allocate land, arrange electricity connections and announce billions in investments. But what happens if the chips needed to operate them cannot be obtained when required?

And suppose the global AI investment boom eventually produces excess computing capacity, making GPUs more readily available. Would that necessarily be good news?

If the chips become available because the worldwide market is already oversupplied, would there still be sufficient paying customers to make Malaysia's enormous data-centre investments commercially viable?

Malaysia could find itself building infrastructure for a global market that has already moved on.

The Bigger Economic Risk

Wall Street is beginning to question whether the AI investment boom has run too far ahead of commercial reality. Malaysia should be asking a different but equally important question. Are we allowing enthusiasm for AI, semiconductors and data centres to distort our national economic priorities?

Economic transformation requires more than chasing the industries currently fashionable among global investors. It requires strengthening existing productive capabilities while developing new ones, ensuring that investment translates into sustainable businesses, employment, productivity and national income.

Malaysia cannot afford to neglect the larger economy while betting excessively on a handful of technology industries whose future profitability remains uncertain.

There is a final irony. Today, Malaysia worries about obtaining enough Nvidia GPUs. By 2035, the world might have more GPUs and data-centre capacity than it needs. When the chips finally arrive, will the customers still be there?

And more importantly, what will have happened to the rest of Malaysia's economy while we were waiting?