The AI market is projected to fragment in 2026. The final quarter of 2025 experienced significant volatility in the technology sector, characterized by a series of sell-offs and subsequent rallies. This turbulence was driven by circular transactions, debt issuances, and elevated valuations, which collectively heightened apprehensions regarding a potential AI bubble. This volatility could indicate the initial stages of the evolution of AI investment, as investors increasingly scrutinize the spending and revenue generation of various entities, as noted by Stephen Yiu. Investors, particularly retail investors engaging with AI via ETFs, generally fail to distinguish among firms that possess a product yet lack a viable business model, those incurring losses to support AI infrastructure, and those benefiting from AI expenditures, Yiu remarked. Currently, “every company seems to be winning,” yet AI is still in its nascent stages, he remarked. “It’s very important to differentiate” between different types of companies, which is “what the market might start to do,” Yiu added.

The initial cohort, comprising OpenAI and Anthropic, attracted $176.5 billion in venture capital during the first three quarters of 2025, according to data. Meanwhile, prominent technology firms such as Amazon, Microsoft, and Meta are the entities financing AI infrastructure providers like Nvidia and Broadcom. The Blue Whale Growth Fund assesses a company’s free cash flow yield—defined as the cash generated post-capital expenditure—relative to its stock price to determine the validity of its valuations. According to Yiu, a considerable number of firms in the Magnificent 7 are currently “trading a significant premium” as a result of their substantial investments in AI. “When I’m assessing valuations in AI, I would not want to position — even though I believe in the transformative potential of AI — into the AI spenders,” he noted, emphasizing that his firm would prefer to be “on the receiving end” as AI expenditures are poised to further influence corporate financials.

According to Julien Lafargue, the AI “froth” is “concentrated in specific segments rather than across the broader market,” as reported. The greater risk is associated with firms attracting investment during the AI bull run that have not yet produced earnings — “for example, some quantum computing-related companies,” Lafargue stated. “In these instances, it appears that investor positioning is influenced more by optimism than by concrete outcomes,” he remarked, emphasizing that “differentiation is essential.” The necessity for differentiation also signifies a transformation in the business models of major technology firms. Asset-light firms are progressively becoming asset-heavy as they acquire the technology, power, and land essential for their aggressive AI strategies. Firms such as Meta and Google have evolved into hyperscalers, making substantial investments in GPUs, data centers, and AI-driven products, thereby altering their risk profile and business model.

Dorian Carrell stated that valuing these companies as if they were software and capital expenditure-light plays may no longer be rational — particularly as firms continue to navigate the funding of their AI initiatives. “We’re not asserting that it won’t be successful, nor are we claiming it won’t materialize in the coming years. However, we are questioning whether it is prudent to pay such a high multiple given the elevated growth expectations that are already factored in,” Carrell stated during an interview. This year, the technology sector has sought financing from the debt markets to support AI infrastructure, although investors have expressed caution regarding an overreliance on debt instruments. While Meta and Amazon have raised funds in this manner, “they’re still net cash positioned,” Ben Barringer stated. The private debt markets are expected to be quite intriguing in the coming year, according to Carrell. Yiu stated that if the growth in AI revenues does not exceed the associated expenses, profit margins will shrink, leading investors to scrutinize their return on investment. Moreover, the disparities in performance among companies may expand as hardware and infrastructure continue to depreciate. Investors in AI must consider this aspect in their financial allocations, Yiu noted. “It has not yet been incorporated into the profit and loss statement.” From next year onward, it will gradually confound the numbers. “Thus, an increasing degree of differentiation is anticipated.”