Is the AI Stock Market Bubble About to Pop?

Summary

The current US stock market, including retirement funds, is built on a story that might be ending: the belief that American companies will generate endless profits from AI technology. This video argues that this AI buildout is significantly larger than the dot-com bubble and explores reasons why this narrative might unravel. One key issue is the lack of trust and verifiable value in current AI models, highlighted by figures like Alex Karp, CEO of Palantir. He points out that AI companies charge per token (word), a pricing model that suggests they aren't confident in AI's ability to deliver guaranteed results, unlike traditional service providers. This is further compounded by AI's tendency to "hallucinate" or make up information, and the risk of companies inadvertently giving their proprietary data and trade secrets to AI providers, potentially creating future competitors. The video also critiques the fundamental business model of AI, comparing it unfavorably to traditional software. Unlike software where each additional customer is largely free, AI services incur significant costs per use due to electricity and chip wear, meaning more customers lead to more expenses, not necessarily more profit. OpenAI, for example, reportedly burned through $20.9 billion in 2025, and their costs increase linearly with revenue, preventing the margin expansion typical of tech companies. Adding to the concern, China is emerging as a formidable competitor. While the US is spending trillions on AI, China is investing a fraction, yet producing AI models that are significantly cheaper – up to 7 to 12 times less expensive per task. This is achieved through "distillation," where Chinese companies compress and open-source advanced AI models developed elsewhere, effectively learning from US innovation at a fraction of the cost. This raises questions about the sustainability of US companies' high valuations if a cheaper, comparable alternative exists. Potential triggers for a market correction are discussed, including a major tech company moderating AI infrastructure spending, financing for AI companies falling through, or a significant pullback in data center debt issuance. The video emphasizes that the stock market doesn't always wait for companies to admit problems, as seen in the dot-com crash where spending continued even after the market fell. The full video is highly recommended for its in-depth analysis and compelling data-driven arguments that challenge conventional wisdom about the AI boom.