The AI market is about to splinter—and the split won't favor the companies you'd expect. For the first time, investors are starting to ask the question that separates winners from losers: Who's spending money versus who's making it? Meta and Google once thrived as asset-light software companies. No longer. They've morphed into hyperscalers pouring billions into data centers, GPUs, and infrastructure. That transition—from software economics to hardware-heavy capital requirements—marks the 2026 inflection point where the AI market segments into monetizers and manufacturers.

The AI market is tipped to splinter in 2026, and nobody's positioning for it yet. The last three months of 2025 were a rollercoaster of tech sell-offs and rallies, but the real signal isn't market volatility—it's the emergence of three completely different business archetypes that investors are still treating as interchangeable.

Stephen Yiu, chief investment officer at Blue Whale Growth Fund, put it directly to CNBC: "It's very important to differentiate" between companies with a product but no business model, those burning cash on AI infrastructure, and those profiting from AI spending. Right now, he said, "every company seems to be winning." But AI is still in early innings, and that changes fast.

Here's what's happening. Private AI companies like OpenAI and Anthropic captured $176.5 billion in venture capital in the first three quarters of 2025 alone, according to PitchBook data. Meanwhile, Big Tech names—Amazon, Microsoft, and Meta—are cutting massive checks to infrastructure providers like Nvidia and Broadcom. The money is flowing from one type of company to another. And for the first time, investors are asking who actually makes money.

The answer is becoming obvious. Nvidia and the infrastructure specialists are making it. Big Tech is spending it. And the market hasn't priced that in yet.

"When I'm looking at valuations in AI, I would not want to position—even though I believe in how AI is going to change the world—into the AI spenders," Yiu told CNBC. He'd rather be "on the receiving end" as AI spending impacts company finances. Most companies within the Magnificent 7 are "trading a significant premium" since they started heavily investing in AI, he noted—a premium that assumes those infrastructure investments pay off on software-company timelines. They won't.

This isn't hypothetical. The business model shift is already happening. Meta and Google have morphed into what the industry calls hyperscalers—companies that invest heavily in GPUs, data centers, and physical infrastructure for AI. That changes everything about how you value them. They were software companies with 70%+ gross margins. Now they're capital-intensive hardware operators with depreciation schedules stretching a decade.

Dorian Carrell, Schroders' head of multi-asset income, said valuing these companies like software plays "may no longer make sense—especially as companies are still figuring out how to fund their AI plans." The question investors should ask: Should you pay a 40x multiple with 25% growth expectations baked in for a company that looks more like a utility than a software vendor? That's what's happening now with Big Tech. It's a valuation mismatch waiting to resolve.

The funding mechanics are revealing the strain already. Tech turned to the debt markets this year to fund AI infrastructure, and while Meta and Amazon have raised substantial funds—and are still "net cash positioned," per Ben Barringer at Quilter Cheviot—that's a distinction that matters. Other companies won't have that luxury. When hardware depreciation enters the P&L in 2026 and beyond, it will gradually "confound the numbers," Yiu said. Margins will compress if incremental AI revenues don't outpace infrastructure expenses. And gaps between winners and losers will widen.

Julien Lafargue, chief market strategist at Barclays Private Bank and Wealth Management, sees AI "froth" concentrated in specific segments—particularly quantum computing companies and other high-expectation bets with no earnings yet. "Investor positioning seems driven more by optimism than by tangible results," Lafargue told CNBC. Differentiation is key. You can't treat a company spending $50B on data centers the same way you treat a company profiting from the infrastructure those centers run on.

Here's where 2026 becomes the inflection point. The AI market isn't consolidating around winners—it's splitting into winners and losers. Infrastructure manufacturers are winning now. AI spenders are betting on future returns. And the market is just beginning to price the difference.

Retail investors exposed through ETFs haven't differentiated yet. They own AI as a theme, not as distinct business models. When that stops—when investors start asking "Is this company spending or making?"—valuations will reset. Some companies will trade down. Infrastructure providers will re-rate up. The private debt markets in 2026 will be the accelerator. When capital gets expensive and margins matter, the companies that own infrastructure will thrive while spenders struggle to justify their capex.

This isn't about AI being a bust. It's about the market finally recognizing that not all AI exposure is equal. By 2026, you'll know exactly which companies should be valued like manufacturers and which like software vendors. The split is coming. The timing just depends on when investors stop treating all AI as the same opportunity.

The AI market's 2026 bifurcation represents the first major valuation reset in Big Tech's asset transition. For investors, the shift from software to infrastructure economics means differentiating between capital spenders (rerating down) and manufacturers (rerating up). Decision-makers at enterprises should monitor whether their Big Tech vendors remain committed to the products they're building or pivot to chasing infrastructure ROI. Builders should note: the era of cheap, seamless AI integration from asset-light vendors is ending. 2026 determines winners. The positioning happens now.