John Higgins, chief markets economist at Capital Economics, states that the AI stock bubble has already burst, based on the fall in the price-to-earnings ratio. However, he warns of a new bubble on the horizon, this time focused on tech companies' earnings. This situation is different from the dot-com bubble of the early century, where the price-to-earnings ratio was much higher, exceeding 150% for the IT sector in the early 2000s, compared to a peak of nearly 75% in late 2024.

Higgins notes that AI valuations have soared, with 498 AI unicorns valued at $2.7 trillion in the fall of 2025, 100 of which were founded in 2023 or later. More than 1,300 AI startups have valuations over $100 million. OpenAI's valuation reached $730 billion last month, according to CFO Sarah Friar, up from $500 billion in October, less than six months prior.

The 'SaaSpocalypse', a rapid selloff of software-as-a-service (SaaS) stocks, has contributed to the correction of the tech sector, as investors fear that agentic AI will easily replace traditional business models. Companies like Salesforce and ServiceNow have lost about 30% of their value since the beginning of the year.

Higgins explains that investors have focused on the software services sector as one of the most vulnerable to the implementation of AI, which has led to a significant reduction in the valuation of this sector. In addition, the semiconductor industry has also experienced a recent slowdown, with high demand fueling a chip shortage and geopolitical tensions triggering supply chain challenges.

Higgins suggests that the next bubble could be on the fundamental side of things, i.e., in the earnings themselves. Tech companies' earnings have increased in recent years, raising the question of how sustainable this growth is. Bloomberg Intelligence estimates that the earnings growth of the Magnificent Seven is around 18%, compared to 11% growth from the remaining 493 companies in the S&P 500.

Nvidia reported revenue of $68.1 billion for its fourth quarter last month, a 73% year-over-year increase. Higgins wonders what would happen if these earnings declined.

Higgins identifies several factors that could lead to a market correction in AI earnings. First, demand for AI could be lower than initially anticipated, forcing tech companies to reckon with the estimated $539 billion in AI capex for 2026, according to Goldman Sachs. Although 88% of companies report regular AI use, according to McKinsey, adoption may be stalling due to employees' anxiety about being replaced by the technology.

In addition, the overall economic situation could affect AI earnings. The war in Iran has halted helium production in Qatar, responsible for about one-third of the world's supply of this gas used to manufacture computer chips. Data centers have also become targets of attacks during the conflict, and energy prices could increase the costs of these facilities.

Higgins warns that if the economy weakens, this could affect the stock market and the earnings of companies that profit from the implementation of AI, even if demand for AI itself does not weaken much. This underscores the interconnectedness of the economy and the tech sector, and how macroeconomic factors can influence the success of AI companies.

In summary, Higgins' warning highlights the importance of a cautious assessment of the AI industry, and the need to consider both the opportunities and the risks associated with its rapid growth and its impact on the global economy.