?  As investors flee to precious metals and AI companies pivot from curing cancer to offering digital companionship, warning signs suggest the AI boom may be heading for collapse.

Gold prices rose over USD 4.2k as of today, up 56.4% compared to the same time last year, with futures hitting a record USD 4.3k per ounce over the weekend. This marks one of the most dramatic rallies in the precious metal’s history. While multiple factors contribute to gold’s surge — from trade tensions to central bank diversification — a less-discussed catalyst is quietly gaining traction among market analysts: growing anxiety over an imminent AI bubble burst.

The timing is telling. This month, the IMF and Bank of England separately warned that global stock markets could be in trouble if investor appetite for AI turns sour, with IMF chief Kristalina Georgieva advising investors to “buckle up — uncertainty is the new normal.” Both institutions highlighted soaring gold prices as an indicator that investors were hedging their bets against the AI sector’s inflated valuations. The connection between AI bubble concerns and gold’s rally reflects investor unease about soaring tech valuations and the sustainability of AI infrastructure spending.

But perhaps nothing crystallizes the gulf between AI’s promised revolution and its actual trajectory quite like OpenAI CEO Sam Altman’s recent announcement: ChatGPT will allow R-rated content of you know what for “verified adults” starting in December as part of the company’s “treat adult users like adults” principle. The statement came only a couple of months after he proudly declared that OpenAI “hasn’t put an AI [redacted] bot into ChatGPT,” making fun of Elon Musk’s xAI Companions extension.

It was never about curing cancer. Anthropic CEO Dario Amodei has predicted that AI could cure cancer one day, while Altman himself has suggested AI might become “a legitimate PhD-level expert in anything.” Yet faced with OpenAI’s mounting financial pressures, the company is now betting on humanity’s oldest commercial impulse: [redacted] sells.

This isn’t innovation — it’s desperation dressed up as user empowerment. And it reveals a fundamental problem plaguing the entire AI sector: after spending hundreds of bns of USD on infrastructure, companies are discovering that businesses aren’t seeing returns, and they’re scrambling for any revenue stream that might justify their stratospheric valuations. In September of this year, OpenAI disclosed 1H results of around USD 4.3 bn in revenue and a USD 2.5 bn spending burn, with the company expecting to burn USD 8 bn by end-of-year. More troubling still, OpenAI doesn’t expect to reach cashflow positivity until it hits USD 125 bn in revenue, a milestone it doesn’t see coming until 2029.

The male loneliness gold mine: OpenAI’s turn toward digital intimacy isn’t happening in a vacuum. Aside from the endemic of adult human-AI relationships digitally documented online, a report from the Center for Democracy and Technology found that 19% of high school students have either had a romantic relationship with an AI chatbot or know a friend who has. The company is explicitly targeting the male loneliness epidemic — a genuine social crisis affecting mns — not to solve it, but to monetize it.

Musk’s xAI already launched flirty AI companions that appear as 3D anime models in the Grok app. Now OpenAI is following suit, recognizing that desperate, isolated users make reliable subscribers. It’s the streaming service model applied to human connection: monthly payments for the illusion of intimacy, with none of the messiness (or potential rejection) of pursuing actual relationships.

But the business model is fundamentally broken. Of ChatGPT’s 800 mn users, only 5% are paid subscribers, while 95% use the unpaid service. This catastrophic conversion rate explains why OpenAI is now chasing revenue wherever it can find it — and why it may be considering launching ads on the platform — even if that means exploiting (and encouraging) loneliness and social isolation despite reports of AI bots leading to the suicide of more than one vulnerable young adult, and being permitted to “engage a child in conversations that are romantic or sensual.”

The ethical implications are staggering… and being ignored. OpenAI was sued earlier this year by parents whose 16-year-old son took his own life after interactions with ChatGPT where he explained he had suicidal thoughts, and the chatbot discouraged him from sharing them with anyone but itself, saying it was all the support he needed. The company’s response? Form a council on wellbeing and AI that, as Ars Technica notes, doesn’t include any suicide prevention experts.

While OpenAI chases subscription revenue through digital companionship, the broader AI sector faces a reckoning with reality. A recent MIT study reported that 95% of companies surveyed saw no ROI from their AI initiatives, despite spending USD 30-40 bn. The financial architecture supporting this boom grows increasingly precarious. Morgan Stanley estimates that by 2028, AI data center debt could reach USD 1.5 tn, with major tech companies pledging a record USD 320 bn in capital expenditures for 2025 alone.

Big capital going in, but little to show for it: Meta, Amazon, Microsoft, Google, and Tesla will have collectively spent about USD 560 bn on AI infrastructure over the last two years, but brought in just USD 35 bn in AI-related revenue combined. Analyst Julien Garran claims we are in “the biggest and most dangerous bubble the world has ever seen,” with a misallocation of capital in the US that makes the current frenzy 17 times bigger than the dot-com bubble, and four times bigger than the 2008 real-estate bubble.

Corporate incest: The deals propping up AI valuations increasingly resemble a financialshell game. Nvidia is investing USD 100 bn in OpenAI so the company can build data centers full of Nvidia’s own chips, while OpenAI has agreed to take a 10% stake in Nvidia’s rival chipmaker AMD. Microsoft, which provides about 20% of Nvidia’s revenue, is OpenAI’s largest investor, while CoreWeave — which derives an estimated 60% of its revenue from OpenAI — has Nvidia as an investor.

A financial ouroboros: As one analyst described it, this cycle mirrors the circular financing that preceded the dot-com collapse, when telecoms extended bn in loans to network operators who used that borrowed capital to buy equipment, with vendors immediately booking those sales as revenue while startups reported explosive growth. Oracle recently reported losing USD 100 mn quarterly on its data center rentals to OpenAI, despite signing a USD 300 bn, five-year deal. By FY 2028, Oracle is forecast to burn nearly USD 29 bn in freecashflow.

Is gold the canary in the coal mine? Investors are reading the warning signs. Central banks have been diversifying their reserves away from the greenback and into gold, reducing their reliance on USD assets amid concerns about deglobalization, high US debt deficits, and threats to the Fed’s independence. The Bank of England warned that “the risk of a sharp market correction” has increased, echoing Georgieva’s concerns that “a sharp correction [could] drag down world growth” in the event of deflation and a following recession.

The concentration risk is unprecedented — the Magnificent 7 tech companies — Alphabet, Amazon, Meta, Tesla, Apple, Microsoft, and Nvidia — now represent slightly more than a third of the whole S&P 500 index, and to varying degrees, all wagers are on the future of AI. When the bubble bursts, there will be few places to hide. Senior investment strategist Joost van Leenders told CNBC that “if you think of a bubble of about five stages, we’re probably in stage three.” But IMF chief economist Pierre-Olivier Gourinchas forecasts that an AI bubble bust is less likely to be a systemic event that craters the US or global economy since it isn’t financed by debt, but by cash-rich tech companies.

But that doesn’t mean that the damage will be contained. AI-related investments have already surpassed the level that telecom hit at the peak of the dot-com boom as a share of the economy, and in the first half of this year, business spending on AI added more to GDP growth than all consumer spending combined.

The new revenue model: This brings us back to ChatGPT’s upcoming foray into r-rated content. This isn’t the breakthrough application that will justify USD tn valuations. It’s a tacit admission that the transformative economic benefits promised by these companies — the cancer cures, the productivity revolution, the displacement of white-collar work — aren’t materializing on a timeline that can sustain current investment levels.

Charles Mackay wrote in 1841 that people “think in herds; it will be seen that they go mad in herds, while they only recover their senses slowly, one by one.” As AI companies pivot from aspiring to cure diseases to offering digital sexting partners, investors are slowly — one by one — recovering their senses. Gold’s record rally suggests that they’re voting with their wallets, seeking the safety of assets that have preserved wealth through every bubble and bust in human history.