The Great AI Split: Why 2026 Will Separate Winners from Losers

The party's over for the 'all-in-on-AI' trade. The rollercoaster ride for tech stocks in the fourth quarter of 2025 wasn't just market noise—it was the opening bell for the great AI market split of 2026. Investors are finally starting to ask a crucial question: who's spending the money, and who's actually making it?

The Three Camps of the AI Economy

According to Stephen Yiu, chief investment officer at Blue Whale Growth Fund, the market is set to splinter into three distinct groups. So far, he told CNBC, investors haven't really differentiated between the types of AI plays, but that's about to change.

The first camp includes private startups like OpenAI and Anthropic, which lured a massive $176.5 billion in venture capital in the first three quarters of 2025, per PitchBook data. The second is the 'AI Spenders'—Big Tech giants like Amazon, Microsoft, and Meta that are cutting huge checks. The third camp consists of the 'AI Infrastructure' firms, such as Nvidia and Broadcom, that are on the receiving end of that spending. Yiu said he'd rather be positioned with the latter.

Warning Signs of a Concentrated Froth

This differentiation reflects a fundamental shift in Big Tech's business model. Once asset-light, companies like Meta and Google are morphing into asset-heavy hyperscalers, pouring billions into GPUs and data centers. According to Dorian Carrell of Schroders, valuing these companies with the high multiples of capex-light software plays may "no longer make sense."

Julien Lafargue, chief market strategist at Barclays Private Bank, told CNBC that the AI "froth" is "concentrated in specific segments." He sees the bigger risk in companies that secured investment but are yet to generate earnings, driven more by optimism than tangible results.

The Balance Sheet Reckoning

This AI arms race is expensive. While tech giants turned to debt markets in 2025 to fund infrastructure, Meta and Amazon are "still net cash positioned," an important distinction. The real test is coming. Yiu said that from 2026 onwards, the depreciation of all that new hardware will start hitting companies' profit and loss (P&L) statements. If AI revenues don't outpace those expenses, margins will compress, and the performance gap between the spenders and the earners will widen significantly.