After a year of relentless enthusiasm for AI stocks, cracks may be starting to appear beneath the surface. While investors broadly treated “AI” as a single winning theme in 2025, market experts say the coming year could mark a turning point—one where clear winners and losers begin to emerge.
The sharp rallies and sudden sell-offs seen in technology stocks during the final quarter of 2025 may not have been random volatility. Instead, they could represent the market’s first attempt to separate companies that are profiting from AI from those that are simply pouring money into it.
As AI adoption accelerates and costs rise, investors are being forced to ask a tougher question: Who is actually making money from this boom?
A Market That Treated AI as One Trade
Throughout much of 2025, investors piled into anything associated with artificial intelligence. Exchange-traded funds, retail portfolios, and institutional strategies often grouped AI-related firms together—regardless of whether they had a proven revenue model.
Stephen Yiu, chief investment officer at Blue Whale Growth Fund, believes this lack of differentiation cannot last.
“For a long time, everything linked to AI was treated as a winner,” Yiu explained. “But AI is still in its early stages, and the market will eventually need to distinguish between who is spending capital and who is generating returns.”
That distinction, he argues, may define the AI investment landscape in 2026.
Three Distinct Camps Emerging
According to Yiu, the AI ecosystem is already splitting into three broad categories:
- Private AI developers and startups
- Publicly listed AI spenders
- AI infrastructure and hardware providers
OpenAI and Anthropic are just two examples of private companies receiving enormous sums of venture capital. According to PitchBook data, AI-oriented startups have raised an impressive amount of venture capital in the first three quarters of 2025, approximately 176.5 billion on its own- an unprecedented number that demonstrates the enthusiasm of investors but also casts doubt on the sustainability of the trend in the long term.
Big technology companies such as Amazon, Microsoft, Meta, and Google are pouring money toward AI development on the open markets. The companies are spending big on data centers, custom chips, cloud architecture, and power capacity to accommodate more complex AI models.
Then there are the infrastructure providers—companies like Nvidia and Broadcom—that sit on the receiving end of this spending. Rather than funding AI expansion, they supply the tools that make it possible.
Valuations Under the Microscope
Blue Whale Growth Fund evaluates companies by examining free cash flow yield, measuring how much cash a business generates after capital expenditures relative to its stock price. Using this lens, Yiu believes many high-profile AI spenders are becoming increasingly difficult to justify from a valuation standpoint.
Several members of the so-called “Magnificent Seven” are now trading at substantial premiums compared with historical norms, largely due to optimism around AI-driven growth.
“I believe AI will change the world,” Yiu said. “But that doesn’t mean I want to invest in the companies absorbing the bulk of the costs.”
Instead, his firm prefers exposure to businesses benefiting directly from AI investment rather than those funding it—at least until the revenue picture becomes clearer.
Froth Concentrated, Not Universal
Market strategists say the risk of an AI bubble is real, but unevenly distributed.
Julien Lafargue, chief market strategist at Barclays Private Bank and Wealth Management, notes that speculative excess is largely confined to specific corners of the market.
“The biggest concern isn’t the entire AI space,” Lafargue said. “It’s companies attracting investment purely on future potential, without current earnings to support valuations.”
He pointed to emerging sectors like quantum computing as examples where enthusiasm may be running ahead of fundamentals.
“In many cases, positioning is driven more by hope than results,” Lafargue added. “That’s where differentiation becomes essential.”
Big Tech’s Business Model Shift
One of the most significant changes underpinning this debate is the evolution of Big Tech itself. Once celebrated for being asset-light and highly scalable, many technology giants are becoming increasingly capital-intensive.
Companies like Meta and Google are now operating more like industrial-scale hyperscalers. They are investing billions in GPUs, land, energy infrastructure, and massive data centers—fundamentally altering their cost structures and risk profiles.
Dorian Carrell, head of multi-asset income at Schroders, believes this shift complicates how investors should value these firms.
“If these companies are no longer capex-light software businesses, then applying traditional software multiples may not make sense,” Carrell said in a recent CNBC interview.
While few doubt AI’s long-term potential, Carrell cautions against paying premium valuations before the revenue payoff becomes visible.
Debt-Fueled Growth Raises New Questions
To finance their AI ambitions, several major technology firms turned to debt markets in 2025. While companies such as Meta and Amazon remain in relatively strong financial positions, investors are watching closely.
Ben Barringer, global head of technology research at Quilter Cheviot, emphasized the importance of balance sheet strength.
“Raising debt isn’t the issue,” Barringer said. “The concern is whether companies can absorb these costs without weakening their financial flexibility.”
This distinction separates cash-rich giants from firms with tighter margins that may struggle if AI investments fail to generate incremental revenue quickly.
Margins, Depreciation, and the Road Ahead
Beyond funding, another challenge looms: depreciation. AI hardware and infrastructure lose value over time, and those costs will eventually hit profit-and-loss statements.
“If AI revenues don’t grow faster than expenses, margins will compress,” Yiu warned. “That’s when investors start questioning returns.”
He added that many of these costs have yet to fully appear in company financials. As depreciation and operational expenses surface over the next year, earnings disparities between companies could widen significantly.
A More Selective AI Market in 2026
Taken together, these forces suggest that the broad, rising-tide approach to AI investing may be nearing its end. Instead, 2026 could usher in a more selective phase—one where investors reward cash flow, discipline, and tangible results over ambition alone.
The AI story is far from over. But as spending accelerates and scrutiny intensifies, the market may finally begin separating the monetizers from the manufacturers—and pricing them very differently.
As Yiu put it simply: “There’s going to be more and more differentiation.”