AI Stocks at Risk: What Rickards Says to Sell

Jim Rickards does not mince words about which AI stocks he thinks are in a bubble. He names specific companies and provides specific data points for each one.

The scale of the AI bubble is staggering. Rickards cites JP Morgan’s Chair of Investment Strategy, who noted that three-quarters of gains in the S&P 500 since the launch of ChatGPT came from AI-related stocks. Without those AI-driven gains, the S&P 500 would be worth roughly half what it is today. AI expenditures accounted for 92% of GDP growth, meaning AI-related spending now contributes more to the nation’s GDP growth than all consumer spending combined.

Nvidia, which designs the advanced chips at the heart of the AI boom, became the first company in history worth $5 trillion. That single stock represents almost 20% of all U.S. GDP. As Rickards points out, Nvidia does not even manufacture its own chips. Taiwan Semiconductor and other manufacturers build them. Nvidia just draws up the designs.

OpenAI is losing more than a billion dollars a month. For every dollar the company makes, it spends at least three. Deutsche Bank estimates OpenAI will need to accumulate $143 billion in negative cash flow before making a single dollar in profit. A Deutsche Bank analyst noted: No startup in history has operated with losses on anything approaching this scale.

Despite this, OpenAI plans to IPO for nearly a trillion dollars. Sam Altman, OpenAI’s CEO, once admitted: I have no idea how we are going to generate revenue. Anthropic, another major AI lab, has warned its business could go bankrupt if AI growth forecasts are off by just one year. Elon Musk’s xAI was burning through cash so fast it had to be merged with SpaceX to keep it from going under.

Rickards singles out Palantir (PLTR) as a particularly egregious example of AI overvaluation. The company has a P/E ratio of 222, which means that if you bought this stock today it would take 222 years at its current earnings to make your money back. He also points to AI startups with no products and no revenue that are supposedly worth a billion dollars, and notes that tech-sector valuations are now well above dotcom era levels.

Oracle (ORCL) is another company Rickards flags as at risk. He notes that Oracle’s default risk hit record highs as the company took on massive debt to build data centers for OpenAI. The company executed major layoffs as it poured capital into infrastructure for a customer that, in Rickards’ view, may not be able to pay. This fits into his broader thesis about data center debt and the illusion of demand in the AI sector.

Beyond individual stocks, Rickards warns about systemic risk. The S&P 500 has reached its highest concentration since just before the Great Depression, with most of the nation’s wealth sitting in just ten stocks. Eight of those ten stocks are soaring based almost solely on AI hype. Torsten Slok, chief economist at Apollo Asset Management, which manages nearly a trillion dollars in assets, has said: The top 10 companies in the S&P 500 today are more overvalued than they were in the 1990s.

Beyond the individual stocks, Rickards points to mounting evidence that AI progress itself may be hitting a wall. Tim Dettmers, a professor at Carnegie Mellon University and AI researcher, has said that rack-level optimization will likely hit a physical wall in 2026 or 2027, and that GPUs can no longer be meaningfully improved. Venture capitalist Marc Andreessen has warned that “we’re increasing GPUs at the same rate, we’re not getting the intelligent improvements at all out of it… AI is hitting a wall.” George Noble, a former Peter Lynch protege who ran the No. 1 Fidelity Fund in the U.S., has said it will cost 5 times the energy and money to make these models 2 times better, and that “the AI hype cycle is peaking. The diminishing returns are becoming impossible to hide.” Yann LeCun, Meta’s longtime chief AI scientist, has said that LLMs are “basically a dead end.” Gary Marcus, a neural scientist and AI researcher, has stated that “when everyone realizes this, the financial bubble will burst quickly.”

Rickards also cites Jeremy Grantham, who once managed over $118 billion in assets, saying: “This is obviously a bubble. The probabilities it doesn’t bust are slim to none.” Former SEC Chairman Gary Gensler has warned that “AI will be the center of the future financial crisis.” BBC reported that AI entrepreneur Jerry Kaplan told a packed audience at Silicon Valley’s Computer History Museum that “when the AI bubble breaks, it’s going to be really bad, and not just for people in AI.” Paul Tudor Jones has said this “is so much more potentially explosive than 1999.” Stacy Rasgon, an analyst for Bernstein Research, which manages $867 billion, has said that “Altman has the power to crash the global economy for a decade.”

The AI Black Paper presentation promotes Strategic Intelligence, Rickards’ monthly newsletter from Paradigm Press. The price is $49 for 6 months, originally $299, an 83% discount that works out to about $8 per month. The guarantee is 3 months: subscribers can request a full refund for any reason within that window and keep all reports.

The package includes six months of Strategic Intelligence plus six special reports: AI Fallout (the biggest AI losers to remove from your portfolio immediately), The AI Black Paper Blueprint (his personal million-dollar roadmap), AI Meltdown Insurance (how to profit from the coming crash), Trump’s AI Arsenal (how investing in AI superweapons could turn $1,000 into $162,000), The Perfect Physical Gold Portfolio, and How to Make Your Home Your Personal Fortress.

Where to Learn More

For the complete analysis, read our AI Black Paper review covering Jim Rickards’ full thesis on the AI Minsky Moment.

For a deeper dive into the bubble thesis, see our analysis of AI bubble warning signs.

Read our guide to portfolio protection strategies for the AI age.

Ready to explore Jim Rickards’ full research? Learn more about Strategic Intelligence here.

This is not financial advice. Always do your own research before investing.