The winners and losers of the changing AI trade
Add Axios as your preferred source to
see more of our stories on Google.
Illustration: Sarah Grillo/Axios
Investors are getting more discerning about AI, looking for companies that will define, and make money from, the next phase of the buildout.
Why it matters: The AI winners of 2026 will be determined by who can prove a return on their investment in the technology.
What they're saying: "Now we're in the name of the game where ROI matters," Futurum chief market strategist Shay Boloor tells Axios.
- When the AI boom began in 2022, Boloor says, the winners were obvious. Think of Nvidia selling its chips to the whole industry.
- But being an AI building block isn't enough anymore. Investors want players who can make AI more efficient, responsive and personalized.
Between the lines: That means the inference and memory companies could get a leg up, while the traditional infrastructure names could lose favor with investors.
🟢 Here are some potential winners
- Memory suppliers: Models need more memory for personalization and chatbots, and the companies that make chips designed to store data and remember information are now catching a bid.
- Micron Technology makes these chips and is growing nearly 100% on the top line. It still trades at just 8 times earnings.
- Data storage firms: These companies help store the data that train AI models, including Sandisk, which rallied 46% over the last five days.
🔴 Here are some potential losers
- Cooling companies: At the Consumer Electronics Show this week, Nvidia CEO Jensen Huang announced a new chip that does not require cooling, which pressured suppliers that provide fans and liquid cooling to keep servers from overheating.
- AI companies that are not focused on responsible budgeting. "OpenAI is having a code red moment, because (Google) Gemini runs so efficiently on inference," Boloor says. "Their economics for compute is unmatched."
Zoom in: The market is making a "mistake" by thinking ChatGPT or Gemini is "what AI is all about," Haim Israel, global strategist at Bank of America, said on a call with reporters.
- "To say AI is about ChatGPT is to really to say, 'we built the internet revolution to have Spotify,'" he added.
- AI is about applications across many industries, and investors could also seek out old-economy stocks, like Walmart, that are using AI tech to cut costs. This is another AI return on investment story to follow in 2026.
What to watch: AI is now "powering record capex, commodity demand and a diversified innovation ecosystem, with the S&P 500 potentially surpassing 8,000 in 2026," JPMorgan noted in its outlook for 2026.
- But the market winners that drive those gains could look different this year as the AI buildout and its bottlenecks change.
The bottom line: "The whole point of AI is going to be agents," Boloor says.
- To get there, you need models that can truly be your colleague, remembering your preferences and storing data along the way.
- That phase of the AI buildout will create different winners and losers.
