Are we in an AI bubble?

Daniel Widenak and Chris Benson discuss the possibility of an AI bubble, defining a bubble as a market situation where company valuations exceed actual value. They draw parallels to the dot-com bubble, expressing concerns about potential economic bubbles in AI and the harm previous bubbles have caused to investments and trust in financial institutions. They reference Jerome Powell's assertion that AI spending is not bubble-like and is supported by actual revenue, contrasting it with the dot-com boom. They speculate on companies like OpenAI and Anthropic, while noting NVIDIA's soaring valuation as indicative of the AI boom.

The hosts highlight that AI-related stocks are driving substantial returns in the S&P 500, although these gains are concentrated in a few companies. They discuss the return on investment for AI, noting that while cloud providers may profit, many companies investing in AI may not see similar returns. Over 50% of venture capital funding in Q2 2025 went to AI companies, indicating a risky investment landscape. AI companies are receiving much higher valuations compared to traditional SaaS companies, with revenue multiples ranging from 15x to 30x.

While current valuation trends echo the dot-com era, they introduce the concept of "beta" as a measure of valuation against earnings, noting that current betas are high but not as extreme as during the dot-com bubble. Chris offers a counterpoint, suggesting that current earnings and business fundamentals may be stronger, indicating a more stable environment for AI investments. Jerome Powell contrasts real earnings in AI with the dot-com era, noting that while some companies have solid earnings, many lack profitability and rely on hype.

The current level of AI investment relative to GDP is low compared to historical transformative revolutions, raising questions about whether we are in an AI bubble. The hosts highlight varying levels of investor engagement and understanding of AI, with many companies labeling products as AI-powered without substantial backing. They suggest that the most valuable AI applications will be specialized solutions addressing real business problems rather than general chat interfaces.

Chris argues that the rationale for AI deployment is deeper than in previous tech bubbles, emphasizing AI's integration into various industries. He expresses confidence that we are not in a classic bubble, citing the long history of algorithmic development, particularly neural networks. Despite some excessive valuations, there is a belief that AI will transform work across industries.

The conversation also addresses AI's impact on employment, highlighting significant layoffs at companies like Amazon as technology replaces human roles. This raises concerns about the balance between AI and human labor, with ongoing debates about synergy versus competition. They reflect on the complexities of the current AI landscape, suggesting that while speculative elements exist, the foundational understanding of AI's transformative potential is strong.

They explore AI's impact on romantic relationships and therapeutic applications, highlighting emotional connections that differentiate the current AI landscape from the speculative nature of the dot-com era. One speaker expresses skepticism about "vibe coding" as a viable strategy for AI use, advocating for treating AI as a "pair programming partner" to enhance software development. There is concern that heavy reliance on AI could diminish cognitive abilities over time, raising alarms about the potential erosion of essential human qualities.

Chris shares his view on the existence of an AI bubble, arguing that while it may not fit the classic definition, there are multiple smaller bubbles present. Dan concurs, acknowledging AI's current utility and integration into various enterprises. Ultimately, both speakers conclude that they do not believe we are in a significant AI bubble, committing to addressing world problems incrementally through their discussions.