By Avi Salzman

Has artificial intelligence been excessively hyped, and are AI-related stocks vulnerable? To assess the risks, it makes sense to check in with someone who has invested through prior hype cycles -- and lived, professionally, to tell the tale. Veteran technology investor Paul Wick fits the bill.

Wick, the chief investment officer and lead portfolio manager for Seligman Investments at Columbia Threadneedle Investments, has been investing in tech companies for more than 35 years. His fund, Columbia Seligman Technology & Information Strategy, is up 36.6% this year, well ahead of the S&P 500's 17% gain.

Barron's spoke with Wick on Dec. 15 about AI risks and opportunities, how energy plays into the AI theme, and which one (or two) of the tech mega-giants he would buy now. An edited version of the conversation follows.

Barron's : Will we still be talking about AI in five or 10 years?

Paul Wick: That's my sense. The development and adoption of AI may not go straight from point A to point B. It may very well zigzag, just as the internet economy did. We don't see as many speculative indicators or red flags today as we saw in the late 1990s when many internet companies were born. Areas that look to be overheating today tend to be in late-stage venture [private companies backed by venture capitalists] more than in the public markets.

So far, AI-enthused investors have poured money mostly into chip stocks. Where does the AI trade go next?

Chips have garnered most of the benefits so far. Software has been viewed as an AI loser. People are worried that AI agents are going to look fairly similar at software companies. And there have been fears that people will create their own AI applications instead of using, say, Workday or Salesforce. Those fears are somewhat overstated, simply because the purchasers of enterprise software don't change their work flows quickly.

Software stocks have also struggled this year because there have been a lot of layoffs in Corporate America. The economy outside of AI hasn't been particularly strong. A lot of companies have been shedding workers, and enterprise software is typically sold on a per-seat basis.

When will these dynamics start to change?

Some companies viewed as AI losers right now will probably end up doing reasonably well. They will get revalued higher as the worst-case scenarios don't come to pass. Over time, it is probable that AI use will broaden out, and things like robotics and industrial automation will have more AI technology built in. This is going to be positive for a range of areas in tech.

Which stocks that have been left behind by the AI rally are underappreciated as potential beneficiaries?

We are fans of Wix.com, an Israeli company that is a leader in website design software. There had been concern that AI "vibe coding" would replace the need for professional software for designing websites. [Vibe coding is a programming method that uses conversational prompts to guide AI in creating and revising code.] That fear has been blown out of proportion. Wix's base business has continued to do pretty well, and it bought a vibe coding start-up in Israel back in June named Base44, which has been doing extremely well.

Wix stock is down about 50% this year because people fear competition from companies like Cursor, which is privately held. Cursor, an AI code editor, was just valued at $29 billion in a late-stage venture-financing round. It has annual revenue of $1 billion heading into 2026. Wix is profitable. It has a $5.5 billion market cap and generates more than $500 million a year in free cash flow. We see more positive risk/reward in the public markets than in late-stage private companies.

Turning to chip stocks, Nvidia has had another good year. Broadcom has done well. What other opportunities do you see in chips?

One of the most controversial names is Marvell Technology. People have been too much glass-half-empty on Marvell. The stock is down 25% this year, even though the company has delivered good results. It has a large backlog that it announced on a recent earnings call. Marvell indicated that its ASIC [chip] business should grow more than 20% in 2026, and it expects to double that in calendar 2027.

Marvell isn't particularly expensive anymore. It has grown into its valuation. The company could earn close to $5 per share in 2027. The stock is $85. That looks pretty reasonable to us.

Nvidia is up 36% this year and more than 1,200% over five years. Do you buy, sell, or hold the stock now?

The narrative that Nvidia will inevitably lose share in AI processors is in the stock. But this is still a fast-growing market. Nvidia just grew revenue by 62% year over year in the latest quarter, and is guiding to 67% growth in its January quarter. It isn't as though the company's growth rate has slowed to a crawl or gone negative.

So, is it worth putting more money into Nvidia stock?

That is reasonable. The stock has cooled down. The company has grown into its valuation. Based on yardsticks such as the price/earnings ratio and enterprise value to free cash flow, the valuation isn't egregious.

Do investors need to protect themselves against the bursting of an AI bubble?

There are areas that feel higher-risk, including the so-called neocloud data-center companies such as Nebius Group and CoreWeave. This is something of a commodity business. It is inherently riskier because these companies don't have significant intellectual property.

You have been looking at investments in energy stocks. Why is it important for tech-focused investors to think more about energy than they have in the past?

Jensen Huang, the CEO of Nvidia, has said multiple times that power is the single biggest constraint on AI. As a result, there is strong demand for adding things like nuclear power. Old nuclear plants such as Three Mile Island are being recommissioned. There is strong demand for gas turbines at companies such as GE Vernova.

The rising tide has lifted a lot of companies that are speculative -- things like Fermi, which has land in Texas and plans to put in gas and nuclear power plants. The idea is that people will put data centers near Fermi's power plants. We have been skeptical of companies such as Oklo and NuScale Power, which are building small modular reactors. It takes a long time to get approval and build a nuclear plant.

Which energy stocks have the most potential?

Bloom Energy has a unique story and many attributes. It makes the most-efficient fuel cells in the world. You can get these installed next to data centers within six months, as opposed to waiting years for your local utility to provide power. Bloom got a 30% tax credit from Trump's tax bill that no one expected. It has been able to lower the cost of its fuel cells by 10% a year for many years. Conceivably, these products will be meaningfully cheaper than gas turbines over time.

Bloom can add capacity fairly easily. It has said it can add a gigawatt of capacity for a little over $100 million. We expect Bloom to grow at a high rate possibly for the next decade.

The stock is trading at almost 90 times the company's 2026 estimated earnings. Does that concern you?

We think next year's earnings estimates are too low. The consensus is $1 a share, and we expect Bloom to earn between $2 and $3. In 2027, the company could earn more than $7 a share. Bloom is going to be growing its top line by 50% for a few years. Next year, it could generate something like $3 billion in revenue. The company has a market value of $22 billion. It isn't particularly expensive.

Should people invest in quantum computing?

We are skeptical of the quantum computing space. Let's keep in mind that the four most prominent public-market quantum stocks were all created through SPACs [special purpose investment companies]. Quantum is an area that is perpetually five to 10 years away, and that has been the case for the past 15 years. It isn't a replacement for AI processors. It can't ingest huge amounts of information.

We feel that the total addressable market for quantum computing is small. It is useful for certain difficult mathematical problems and encryption. This is the reason governments are interested in quantum computing; they are nervous about people being able to break encrypted codes. But the commercial applications for quantum computing look like they are relatively trivial.

If you had to invest in one tech company worth more than $1 trillion, what would you pick?

Can I pick two?

Sure.

Alright. Broadcom's stock has sold off a bunch. In the next four quarters, the company will have AI-related revenue of $79 billion.

Google parent Alphabet would be my other pick. Google is the leader in AI. It makes Waymo self-driving cars. The company's YouTube and YouTube TV are both doing well. The new chief financial officer seems like she is running a very tight ship. The company has been buying back a lot of stock. It has the best free cash flow profile of all of the Magnificent Seven. There is a lot to like about Alphabet.

Thanks, Paul.

Write to Avi Salzman at avi.salzman@barrons.com

This content was created by Barron's, which is operated by Dow Jones & Co. Barron's is published independently from Dow Jones Newswires and The Wall Street Journal.

(END) Dow Jones Newswires

December 24, 2025 14:36 ET (19:36 GMT)

Copyright (c) 2025 Dow Jones & Company, Inc.