The AI market is likely to split in 2026.
The last three months of 2025 have been a rollercoaster of declines and rebounds in tech stocks, with cyclical trading, bond issuance, and high valuations fueling fears of an AI bubble.
Stephen Yiu, chief investment officer at Blue Whale Growth Fund, said such volatility could be an early sign of how AI investing will evolve as investors pay close attention to who is spending their money and who is making it.
Investors, especially retail investors exposed to AI through ETFs, typically don’t differentiate between companies that have a product but no business model, companies that are burning cash to fund AI infrastructure, or companies that are on the receiving end of AI spending, Yu told CNBC.
So far, “every company seems to be winning,” he said, but AI is in its infancy. “It’s really important to differentiate” between different types of companies, Yui added, and that’s “what the market could start to do now.”
This illustration taken in Paris on April 20, 2018, shows apps from Google, Amazon, Facebook, and Apple, as well as the reflection of binary code on a tablet screen.
Lionel Bonaventure | AFP | Getty Images
He sees three camps: private companies or startups, publicly traded AI investors, and AI infrastructure companies.
The first group, which includes OpenAI and Anthropic, attracted $176.5 billion in venture capital in the first three quarters of 2025, according to PitchBook data. Meanwhile, the names of Big Tech are: Amazon, microsoft and meta Cutting checks on AI infrastructure providers: Nvidia and broadcom.
The Blue Whale Growth Fund measures a company’s free cash flow yield (the amount of money a company generates after capital investment) compared to its stock price to determine whether a valuation is justified.
Most companies in the Magnificent 7 have been “trading at a significant premium” since they started investing heavily in AI, Iu said.
“When we look at valuing AI, we don’t want to position it as a spender, no matter how much we believe it will change the world,” he added, adding that the company would rather be a “receiver” as AI spending would have a further impact on corporate finances.
Julien Lafargue, chief market strategist at Barclays Private Bank and Wealth Management, told CNBC that AI “bubbles” are “concentrated in specific segments rather than across the market.”
The bigger risk lies with companies that have secured investment from the AI bull market but are not yet profitable — “for example, some quantum computing companies,” Lafargue said.
“In these cases, investor positioning appears to be driven more by optimism than tangible results,” he added, adding that “differentiation is key.”
The need for differentiation also reflects the evolution of Big Tech’s business models. Once asset-poor companies have become increasingly wealthy, gulping down the technology, power, and land needed for aggressive AI strategies.
Companies like Meta and Google have transformed into hyperscalers, investing heavily in GPUs, data centers, and AI-driven products, changing their risk profiles and business models.
Dorian Carrel, head of multi-asset and income at Schroders, said valuing these companies on light software and capital expenditures may no longer make sense, especially as companies are still figuring out how to finance their AI plans.
“We’re not saying it won’t work, we’re not saying it won’t happen in the next few years, but we’re saying why should you pay such a high multiple when you have such high growth expectations built in,” Carell told CNBC’s “Squawk Box Europe” on Dec. 1.
Tech companies turned to the bond market this year to fund AI infrastructure, but investors were wary of relying on bonds. Meta and Amazon are raising money this way, but “they still have net cash,” Ben Ballinger, global head of technology research and investment strategists at Quilter Cheviot, told CNBC’s “Early Edition Europe” on Nov. 20. This is an important difference from companies with potentially tighter balance sheets.
Mr. Carell added that the private bond market “is going to be very interesting next year.”
If AI revenue growth does not outweigh these costs, profit margins will be compressed and investors will question their return on investment, Yiu said.
Furthermore, as the value of hardware and infrastructure declines, the performance gap between companies is likely to widen further. Those spending on AI need to consider the investment, Yiu added. “It’s not yet part of the income statement. From next year onwards, the numbers will gradually get messed up.”
“So there will be more and more differentiation.”
