หุ้นเทคโนโลยีปัญญาประดิษฐ์ (AI) ที่ทำผลงานได้ดีในปี 2025 ไม่ได้ทำผลงานได้ดีในปี 2026 นี่คือแผนการใหม่

Key Points
After a swell of hype, investors are now looking for adequate profits.
Not every AI-powered solution brings actual marketable value to the table.
Every player in the artificial intelligence business is being forced to think about power efficiency.
- 10 stocks we like better than Palantir Technologies ›
Last year was another fantastic one for artificial intelligence (AI) stocks, extending a rally that began in early 2023 (shortly after OpenAI's ChatGPT launch in late 2022 set off an AI race). Memory chip company Sandisk led the charge with a stunning 559% gain in 2025, while decision-intelligence software powerhouse Palantir Technologies (NASDAQ: PLTR) saw its stock soar an impressive 135%. Of course, Nvidia (NASDAQ: NVDA) had another good year as well, gaining 36%, only held back by its sheer size.
Something's happened in the meantime, though. Most of these stocks have stopped making forward progress. Nvidia shares are still priced where they were as of September. Palantir's stock has fallen back to its mid-2025 level. What gives?
Will AI create the world's first trillionaire? Our team just released a report on the one little-known company, called an "Indispensable Monopoly" providing the critical technology Nvidia and Intel both need. Continue »
In short, investors have come face-to-face with the fact that simply being in the artificial intelligence business isn't enough. The hype needs to be followed by adequate profits. The steep valuations eventually need to make sense. Too many of these names aren't truly meeting either requirement.
That doesn't mean you should simply give up on the entire AI revolution just yet. You'll just want to think about what the market is no longer rewarding -- and what it is rewarding -- within the AI arena.
Here's the AI investing playbook for the new year, and perhaps for the industry's new era.
Profitability matters now
In the AI business's earliest days, hardware outfits like Nvidia and Broadcom were the only companies making real money, but they were making tons of it! It didn't really matter, though. Investors were willing to take a shot on any company with a compelling growth story.
After three years, however, the market is rightly asking where many of these companies' profits are. They're not where many of them were expecting them to materialize.
Take the aforementioned software name Palantir as an example. It would be naïve to believe that last year's net income of $1.6 billion was anywhere near satisfactory, given the organization's $330 billion market cap, even if its per-share profits are expected to improve more than 70% this year and grow another 40% next year. That's at least part of the reason this stock's peeled back more than 30% from its November peak.
At the other end of the spectrum, AI-capable data center stocks are doing great, with their underlying companies turning solid profits by serving customers that can't or don't want to incur the expense of building their own facilities. Data center Digital Realty (NYSE: DLR) was able to improve its 2025 top line to the tune of 10% last year, for instance, and perhaps more importantly, grow its operating bottom line by nearly 40%. It's looking for similar progress this year as well. That's why DLR shares remain in a long-term (albeit choppy) uptrend that's been underway since 2023, performing pretty well of late even when most other AI stocks haven't.
Of course, these are just a couple of examples from the extreme ends of the spectrum. The bigger takeaway for investors is simply that the market's starting to separate the leaders and laggards here, using profitability and subsequent valuations as a dividing line.
AI solutions must serve a clear, marketable purpose
At the risk of drilling too deeply into any particular facet of the AI movement, not every AI-powered solution is demonstrating enduring, marketable value.
Take artificial intelligence "agents" -- AI-powered digital assistants utilized via a text-based chat -- as an example. All of them are novel. However, not all of them do their users enough actual good to make them worth their cost. They also make mistakes that are difficult to pinpoint and then fix (particularly computer coding agents). This is one of the chief reasons a recent survey performed by PwC alarmingly indicates that 56% of CEOs say they have yet to see any fiscal benefit from investments in AI.
This isn't to suggest that AI agents don't have their rightful place. They can, and do. The automated customer service solutions powered by NiCE (NASDAQ: NICE), for instance, are well-received. Indeed, technology consulting and industry research outfit Gartner has now rated NiCE as a leader of the contact-center-as-a-service business for 11 consecutive years, reflecting how well its tech and platform handle certain kinds of customer service interactions. This is also why last year's revenue growth of 9% was led by cloud computing growth of 14%, where its AI-capable automated customer service agents operate.
The bigger point for interested investors is simply that we're seeing more discernment and discrimination from companies exploring AI tools. Enterprises aren't interested in paying for solutions that don't offer demonstrable value.
Power efficiency has become enormously important
Finally, perhaps the most underestimated effect of the rise of AI is the strain it's putting on the global power grid, which is only going to grow as AI data centers proliferate. The International Energy Agency (IEA) expects data centers' consumption of electricity to grow by 15% per year through 2030, in fact, which is more than four times faster than overall energy usage growth for this timeframe.
Of course, soaring utility prices are exacerbating the industry's operating cost headaches.
But the industry is responding. Processing chips designed by Arm Holdings (NASDAQ: ARM), for instance, are quickly becoming AI data center favorites because they can run using less than half the power that competing chips require. The power that's being delivered to data centers' racks is also being rethought. As it turns out, the 415-volt AC (alternating current) power supplies that owners/operators have historically used aren't nearly as efficient as 800-volt DC (direct current) systems. This impending shift bodes well for a company like Vertiv (NYSE: VRT), which will launch its new 800-volt systems for Nvidia hardware later this year.
These are just a couple of examples, of course. But they're also a representation of one of the AI business's newer and most pressing priorities. It's unlikely that any discussion of any investment in AI solutions will not include some consideration of its ongoing electricity costs. Investors can expect more from the AI companies that are more competitive in this regard.
Should you buy stock in Palantir Technologies right now?
Before you buy stock in Palantir Technologies, consider this:
The Motley Fool Stock Advisor analyst team just identified what they believe are the 10 best stocks for investors to buy now… and Palantir Technologies wasn’t one of them. The 10 stocks that made the cut could produce monster returns in the coming years.
Consider when Netflix made this list on December 17, 2004... if you invested $1,000 at the time of our recommendation, you’d have $532,066!* Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you’d have $1,087,496!*
Now, it’s worth noting Stock Advisor’s total average return is 926% — a market-crushing outperformance compared to 185% for the S&P 500. Don't miss the latest top 10 list, available with Stock Advisor, and join an investing community built by individual investors for individual investors.
*Stock Advisor returns as of April 4, 2026.
James Brumley has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Digital Realty Trust, Nice, Nvidia, Palantir Technologies, and Vertiv. The Motley Fool has a disclosure policy.
The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.