The question of whether artificial intelligence (AI) represents a bubble in equity markets resists straightforward answers, according to Henry Fisher, market analyst at CMC Markets. Fisher cautions against reducing the debate to a simple yes-or-no conclusion, emphasizing that investment decisions hinge on a spectrum of probabilities, timeframes, and position sizes rather than clear binaries.
Concerns about an AI bubble have circulated since at least 2022, but some key AI-related stocks have seen dramatic gains since early 2023. Notably, Nvidia’s share price has risen approximately 15-fold, Palantir Technologies has increased by about 30 times, and Micron Technology, a supplier of memory chips for AI applications, has risen roughly 20-fold during this period.
Fisher suggests that being prematurely convinced of a bubble can be akin to being wrong, referencing investor Howard Marks’s observation that timing is critical in markets. He notes that investors do not necessarily need to commit to a definitive stance on the bubble question but can instead adjust their exposure as new information emerges. Traders, he adds, may find opportunities amid the volatility generated by the uncertainties surrounding AI.
He describes technology trends as waves riding on a rising tide: markets may experience sharp surges and corrections while the underlying long-term trend continues upward. Historical examples like major technology stocks and bitcoin illustrate assets recovering from steep declines to reach new highs. Fisher contends that episodes of excess valuation and sustained growth can coexist, but warns that significant corrections do not automatically signal an end to the broader trend.
One of the central challenges for investors is assessing the intrinsic value of AI-related assets amid evolving business models and vast uncertainty about the future. Fisher points to the case of Anthropic, whose leaked IPO documents projected $4.6 billion in revenue but a $42 billion loss in 2025, with infrastructure commitments reaching $15 billion. The company was targeting a valuation around $2 trillion, highlighting the difficulty of making sense of current financial metrics in a rapidly developing sector.
Drawing a historical parallel, Fisher compares the current AI landscape to the internet in 1995, when early assessments underestimated the technology’s eventual economic impact. The internet’s expansion created new markets and business types well beyond initial advertising revenue estimates. He suggests the AI trend may similarly be in its early stages.
Beyond questions of bubble dynamics, AI is already influencing market valuations in other sectors. Fisher notes that AI-driven disruption is exerting pressure on companies traditionally considered to have strong competitive moats, including software firms such as Adobe and Xero. Significant stock drawdowns can expose weakening fundamentals but also generate entry points when markets overreact; Atlassian’s stock more than tripled from its April low, exemplifying this dynamic.
Fisher highlights that while AI can boost productivity—enabling companies to produce more with fewer resources—this does not guarantee improved profitability. Increased productivity may saturate markets and intensify competition, reshaping business economics in unpredictable ways. The productivity gains AI affords could simultaneously undermine existing profit models across industries.
The broader implications of AI extend beyond financial markets to issues of employment, safety, energy consumption, and regulatory responses. Fisher underscores the importance for investors of observing how governments, firms, and markets adapt to these challenges, acknowledging that predicting specific outcomes remains difficult.
Given the rapid pace of AI development and its uncertain future, Fisher advises investors to gather diverse perspectives and closely monitor evolving assumptions and risks embedded in valuations. Rather than trying to definitively predict an AI bubble, investors should adopt flexible strategies that enable them to respond as new evidence emerges.
“If you’re lying awake at night worrying about an AI bubble, you probably own too much of it,” Fisher concludes, underscoring the need for balanced exposure and adaptability in navigating this transformative but volatile sector.