If you’ve had an artificial intelligence (AI) heavy role in the past two years, it’s probably had many of the same names: Nvidia, Advanced Micro Devices, Microsofta few hyperscalers, and maybe some software-as-a-service (SaaS) plays with “AI” somewhere in the investor deck. Back then, if a CEO or someone whispered “AI implementation” during an earnings call, it felt like the stock would skyrocket 15% overnight.
If you’ve been paying attention, the list of trending AI stocks looks different today. As a result, some AI positions have dropped significantly. You’ve run out of some things you didn’t own.
The rotation away from AI began quietly. In early 2026, investors began asking a question the market had avoided for two years: If AI is going to reshape every sector (i.e., will AI steal my job?), why are companies being reformed trading at the same multiples as those doing the reforming? In other words, why are some of these huge private and public AI companies fundamentally unprofitable, burning huge amounts of money on computers when real customer demand and revenues don’t justify the costs?
Image source: Getty Images.
The market is repricing its AI shares
Morgan StanleyThe Global Investment Committee of the Global Investment Committee has put together a useful framework: The market is shifting from AI “builders,” the infrastructure providers and chip companies, to AI “adopters,” which are companies that use AI to actually increase productivity and margins, as reflected in their profit and loss statements.
The downside of this is the repricing of companies most at risk of disruption. That’s what happened with software. The software sell-off wasn’t irrational, even if it was exaggerated. It was the market that tried to separate companies whose pricing power AI survives from those that lose it.
When Anthropic released agentic tools that could automate business workflows, the market asked a reasonable question: why pay per-seat SaaS costs when AI can do the work? The resulting panic to sell everything punished both good and bad companies, but the underlying demand is legitimate
Meanwhile, semiconductors held their ground. For example the Russell 1000 semiconductor The index deviated sharply from the software sector of the Russell 1000. Physical AI infrastructure continued to grow. Data center cooling companies reported record backlogs. Fiber optic connectivity companies launched new product lines with optimized density for hyperscale environments. The parts of the AI stack that were paid for in real dollars, based on real contracts, continued to grow
What a healthy AI portfolio looks like now
A portfolio built for the next phase of AI trading looks less like a concentrated technology bet and more like a layered infrastructure position.
Think about it in terms of who gets paid, regardless of which AI platform wins. The cooling infrastructure doesn’t care whether OpenAI, Anthropic or Alphabet wins the model race. Data centers need chillers anyway.
A good example of this is Vertiv (VRT +6.97%). This company directly benefits from AI’s power and cooling needs, providing the thermal infrastructure every data center needs, regardless of which models win. Another does Equinix (EQIX +1.68%)that manages the physical backbone of the internet and leases data center capacity and interconnection services that scale with AI workloads.
Today’s change
(6.97%)$16.33
Current price
$250.55
Key data points
Market capitalization
$96 billion
Day range
$233.40 -$251.00
Range of 52 weeks
$53.60 -$282.05
Volume
307K
Avg. full
7.9 million
Gross margin
34.26%
Dividend yield
0.08%
Fiber connectivity doesn’t care whether the winning AI runs on Nvidia or AMD GPUs; it needs fiber optic anyway. Enterprise AI tools deployed at scale for long-term contracts provide revenue visibility that isn’t revised with every quarterly sentiment shift. Amphenol (APH +6.04%) powers AI clusters with high-speed connectors and interconnect systems that are becoming increasingly important as computing density increases.
Today’s change
(6.04%)$7.20
Current price
$126.35
Key data points
Market capitalization
$155 billion
Day range
$121:00 -$126.51
Range of 52 weeks
$56.45 -$167.04
Volume
10M
Avg. full
10M
Gross margin
36.88%
Dividend yield
0.66%
Here’s a subjective view of an AI portfolio that looks different than it did six months ago: The change itself is a signal that trading is maturing and not dying. Bull markets early in any major technology shift tend to cancel everything out, as optimism is high and the future feels limitless, full of venture capital funding.
But the next phase is much less forgiving, as the speculation disappears and the difference between real companies and hype becomes clear. The companies that make it through this kind of rotation, the companies that are worth owning a year from now, will be companies with sustainable demand and a clearly defined role in the AI ecosystem that endures even as sentiment cools.
