5 Stocks to Buy to Invest in AI With Less Risk (2024)

Susan Dziubinski: Hello, and welcome to The Morning Filter. I’m Susan Dziubinski with Morningstar. Every Monday on The Morning Filter, we cover what investors should have on their radars this week, some new Morningstar research, and a few stock ideas to research further. Now, my colleague Dave Sekera is on a well-deserved break this week. So today’s episode is going to be a little different.

I’ll cover some of the ground that Dave normally does, though I’m certainly no substitute. Then we’ll share a couple of interviews that Dave conducted before he left about investing in AI today. And then finally, I’ll close out the show with some stock picks based on Dave’s interviews. So let’s get to it. What should investors have on their radars this week?

Well, it’ll be a quiet week in terms of economic reports, thank goodness. Last week’s jobs report showed weak job creation and the highest unemployment rate since late 2021. Now, that triggered fears of a recession in the US, which triggered a market selloff on Friday. And by the looks of the futures market before open here on Monday, it looks like the selloff is likely to continue at least today.

Now on the earnings front, we have a couple of companies reporting this week that have been at the forefront of the weight loss drugs story, and those are Eli Lilly and Novo Nordisk. Now even though both stocks pulled back recently, they still look about 50% overvalued heading into earnings. Now, while Morningstar’s healthcare team thinks both companies have a long runway for growth with their stable of weight loss and diabetes treatments, they think the market is overenthusiastic about these stocks, and they see significant price risk here.

We’ll also be hearing from both Disney and Paramount this week. Both stocks look undervalued heading into earnings. Morningstar’s senior analyst Matt Dolgin covers both companies. Wiith Disney, Matt’s not expecting a big quarter of subscriber adds to the Disney’s streaming service, but he is curious to see what Inside Out 2′s success at the box office contributes to results. And he’s also interested to see how much Disney’s experiences business has slowed.

And then over at Paramount, he’ll, of course, be listening for more details about the Skydance deal. And finally, several of Dave’s stock picks from prior shows report this week, including Realty Income, International Flavors and Fragrances, Fortinet, and Baxter, all of which remain undervalued heading into earnings. We’ll be listening to what management from each of these companies has to say during their earnings calls.

Now moving on to some new research from Morningstar. I’d like to run through several notable changes to the fair value estimates on the stocks of some big companies that reported earnings last week. Let’s start with Microsoft. Morningstar raised its fair value estimate on Microsoft by more than 12% after earnings to $490 per share. Morningstar senior analyst Dan Romanoff noted that earnings came in as expected, and he seemed most impressed with guidance calling for Azure revenue to accelerate in the back half of the year as for the company’s investment in data center capacity comes online. Microsoft stock has a 4-star rating today, which means Morningstar thinks it’s undervalued. The Morningstar also raised its fair value estimate on Meta Platforms by more than 12%, up from $400 per share just $450 per share. Now, Morningstar Director Mike Hodel observed that solid ad demand allowed meta to invest aggressively while generating strong cash flow.

Mike was nevertheless happy that Meta maintained and didn’t increase its expense forecast for the year. Meta stock is trading in 3-star range today and looks fairly valued, according to Morningstar. Morningstar also raised its fair value estimate on Amazon by $3 to $195 after earnings. Morningstar’s Dan Romanoff thought second-quarter results were solid. The change in fair value was influenced by continued enhancements to profitability in the near term.

Dan says the stock is increasingly attractive after pulling back during the past few weeks. In fact, Amazon stock is trading at 4-star levels today, suggesting that the stock is undervalued. Morningstar also raised its fair value estimate on Apple by nearly 9% after earnings to $185. Morningstar analyst Will Kerwin said in his note that the fair value boost was due to an increase in Morningstar’s medium-term iPhone revenue forecast.

Now, Morningstar is forecasting double-digit iPhone revenue growth in fiscal 2025 as users upgrade their iPhones to take advantage of Apple’s generative AI features. Now, even after that fair value increase, Apple stock looks about 20% overvalued. And then on a somewhat related note, news came out this weekend that Berkshire Hathaway had sold about half of its huge stake in Apple stock last quarter.

It’ll be interesting to see if that reveal has any impact on the stock this week. And then lastly, Morningstar slashed its fair value estimate on Intel by 30% to $21 after the chipmaker provided a disappointing forecast, announced layoffs, and paused its dividend. The stock sold off by more than 25%, and Morningstar strategist Brian Colello called the selloff justified and noted that he’s concerned about Intel’s competitive positioning after that earnings report. Viewers who’d like to read more details about what Morningstar’s analysts thought about the earnings of these companies, and, of course, others, can click on the link beneath this video to access Morningstar’s earnings coverage. Now moving on. A couple of weeks ago, Dave sat down with two of Morningstar’s strategists to discuss opportunities in AI. His first conversation was with Brian Colello, who covers Nvidia, among other tech companies.

His second conversation was with Travis Miller, who’s a strategist with Morningstar’s utilities team. Let’s have a listen.

Dave Sekera: Artificial intelligence burst into the limelight in November 2022 when OpenAI launched ChatGPT. Once investors saw what AI could do and its potential for rapid growth, stocks tied to AI have surged to new highs. So, the question is, what exactly is AI? How is it utilized? And who benefits the most? And most importantly, where is it going? To help me answer these questions is Brian Collelo, our equity strategist on the technology team at Morningstar Research Services, to talk about how AI is currently being used and where we see it going forward.

Well, Brian, thank you for joining us here.

Brian Collelo: Thanks for having me.

Sekera: Large language models have really garnered all the limelight in the media today. But could you explain what is exactly AI and what are some of the use cases we see it being used for today?

Collelo: Artificial intelligence has been around quite a while, but the excitement that we’ve seen over the past 18 months is really around generative AI, which is a subset of the broader AI category. Even if you take AI as a whole, which again has been around for decades, machine learning is just a piece of overall AI. And machine learning is really used to sort of detect patterns in data. Companies will feed it data that the humans basically just can’t detect those patterns on their own. So, machine learning would aid in that. Now within that is a subset called deep learning, and within that is a subset called generative AI. And that’s where these large language models come from. The key concept in generative AI, which is different from all the others, is that these generative AI tools, these large language models, create new and original content. And so that’s something we haven’t seen before.

Sekera: From an investor perspective, of course, the question really becomes, how exactly are these companies going to utilize AI to either one, generate additional revenue on top of what they’re already generating? And/or how is it going to make companies more efficient in order to try and improve their operating margins?

Collelo: The start of AI right now, the winners are the companies building out the infrastructure. So, Nvidia for sure. And then the cloud companies renting out those Nvidia GPUs and those data centers to their customers. So, if we put that aside for a minute, we could talk about enterprise software. It’s more to your question of where’s the revenue going to come from, how are they going to going to save? So, on the revenue side, or maybe taking a step back, in software we think AI will be infused in existing products. It won’t necessarily be just an AI tool but it will be AI baked into the products we’re already using. So, from a revenue side, for a best-case scenario for software, a company will be able to explicitly upsell AI by selling a different module or new product. Microsoft Copilot might be the best example of this trying to add on $30 per month per user to get AI tools. That’s probably the best case from a revenue perspective. The medium scenario in that would be a software company is able to raise prices on the product on the whole or maybe they stay flat, maybe they don’t really explicitly charge for AI, but it allows them to retain business. Maybe they reduce customer churn, maybe they hold on to those customers longer. The worst case scenario for software companies are those that don’t develop an AI or are subpar and they just get steamrolled by competition later on.

That’s what we think will happen on the revenue side. On the efficiency side, we’re seeing efficiencies in many different places. Computer coding is one where AI is helping computer programmers there. Customer service is another area. Just even, there’s a lot of concern, maybe long term, about job loss and things like that, but even in the interim, if you take your average customer service representative, they’re able to get information on their fingertips more quickly to better serve their customers, being able to, let’s say, look through manuals or know where to go in a software program, things like that. So, there are advantages to AI being deployed in those two areas already making businesses more efficient.

Sekera: When I think about investing in AI, there’s the old adage, during the gold rush, back in the day, it wasn’t necessarily the miners that got rich, but the people that did get rich were the ones that were selling the picks and the shovels. So, when I think of the picks and the shovels today, of course, Nvidia with its GPUs, by far the market leader selling the chips that are being used to train the AI models today. So maybe if you could just explain to us, what exactly is it about Nvidia’s GPUs that make them the leader today, but I’m also curious, who else is out there? Who else is making products in order to try and do similar types of things or are there other chips that are being made out there for specific AI use cases?

Collelo: I’ve used that analogy as well that Nvidia was basically the only pick maker in town, and they make the best shovels and they’re expanding into bluejeans and water bottles and everything that a miner can use. I think they’re doing a really good job of elbowing out into other AI applications. This all started, Nvidia, I think we’ve always thought of them as a best-of-breed GPU vendor to begin with probably for the past 20 years, but more so on the gaming side with their technology. But sort of the big bang moment in the industry was in 2012 when a host of researchers reprogrammed one of those Nvidia GPUs and discovered that you could get far better results with a large language model running Nvidia. The first part, I don’t want to call it luck, but certainly the GPU architecture that Nvidia has was well suited for these large language models. Where Nvidia was really bright was investing in that trend for the past 10 to 15 years on the software side, creating the software and compilers and tools and all of the things to help the AI industry build these large language models. So now they’re all built using CUDA software that is proprietary, that only runs on Nvidia GPUs, and that’s why they have such a dominant position today.

In terms of competition, AMD has GPU experience as well, both on the gaming side, they have their AI chips out to market, not nearly at the size and scale of Nvidia, and the software is going to be a bit of a challenge, but certainly they want to be the second general-purpose GPU vendor. Intel also has AI accelerator chips that are at play there. We think the biggest threat to Nvidia over the next few years will probably be in-house chips built by the cloud companies themselves. So, Microsoft, Google, Amazon, for example, if they can take workloads and run it on their own chips, they can reduce their reliance on Nvidia. I think they’re still going to buy plenty of Nvidia chips, but to the extent they can reduce their reliance or even just put pricing pressure back on Nvidia, it’ll be a positive for those companies as well. And then outside of that, you have companies like Broadcom and Marvell that are helping the cloud companies build those in-house chips. Broadcom and Marvell are just two of many that are supplying sort of peripheral chips into the networking architecture. And then finally all of these companies, there’s an entire supply chain around them. So, whether it’s Taiwan Semi as the world’s leading foundry building, many of these chips or whether it’s ARM with the architecture and basically the blueprint for designing many of these chips, there’s a whole ecosystem of semis that are supporting Nvidia, but also their competitors, too.

Sekera: Great. Thank you. Thinking about what’s coming up and looking forward, if you could gaze into your crystal ball today, I kind of want to end on an open-ended question here for you, but just thinking about maybe over the next two or three years, how do you really see this developing? What are maybe some of the additional use cases that investors should be thinking about today?

Collelo: Nobody has a great crystal ball on this. I don’t even think the cloud companies do, maybe Nvidia doesn’t either, to be perfectly honest. The big question we get from investors about Nvidia and others is how long can the spending keep going? We would continue to look at the spending plans of the hyperscalers, the Googles, Amazons, Microsofts, and Metas of the world to determine how much more Nvidia would be selling, how much more of an AI buildout is happening because they would only be doing so if their tens of thousands of customers are developing AI models or not. That’s what we would look at from a revenue and sort of trend perspective. In terms of where it’s going, we talked a little bit about enterprise software and how AI will be infused there. I do think there are plenty of startups coming, but I think it’s too early to be expecting significant revenue from there. We’ve seen some reports and opinions that where are all the tools, there’s been development for two years. I still think it’s a bit too soon. The Ubers and Instagrams of the world, let’s say those, that might be a five- to 10-year trend until we start seeing these AI applications. One to two years might still be a little too early, or at least that’s my opinion on it.

I would also say nation-state and government spending is an interesting sort of trend as well for security and defense purposes and things like that. And that spending might happen completely independent of revenue. We see that this is the future in AI. And then finally, there are these big sort of moonshot projects that are going on where it will take a big piece of R&D budget and maybe they work out or maybe they don’t, but I think the spending will be there. So, things like drug discovery, genetics, autonomous driving, robotics, those are problems that the technology industry and healthcare and a lot of other industries haven’t cracked yet. And now with all this computing power, they might be able to move the ball forward on that. And so, I think that spending is going to happen again, maybe regardless of revenue. I’m sure they all want to cure cancer or have self-driving cars or make billions of dollars off of that. But I think the spending will happen either way.

Sekera: Thank you very much. I really appreciate you spending your time and providing your expertise to us today.

Collelo: Happy to help.

Sekera: Many types of companies stand to benefit from the long-term secular growth in artificial intelligence. And this means that there are opportunities for investors interested in this theme beyond the obvious AI stock plays like Nvidia, Microsoft, and Alphabet. With me today is Travis Miller, our equity strategist on the utilities team at Morningstar Research Services, to discuss how utilities stand to benefit from the growth of AI.

So first, Travis, thank you for being here with us today. But of course, the first question has to be, why is there now this recent focus by the marketplace on utilities as a potential play on artificial intelligence?

Travis Miller: Data centers are huge electricity consumers. So, when utilities’ primary business is supplying, either distributing or transmitting or generating electricity, data centers and the growth there is going to be a huge business opportunity for utilities. That means earnings growth. That means more investment in infrastructure. So, we think utilities are a good way to play this data center growth. To give you one interesting example here, Texas right now, it’s the largest area for bitcoin mining. Bitcoin mining and the data centers involved with that alone equal the same amount of electricity on a hot summer day in the entire city of Austin.

Sekera: Wow, that’s incredible.

Miller: Some grid operators in Texas are projecting that generative AI data centers, if they grow in Texas, could be 20 times that amount of electricity. This is a huge growth opportunity for utilities.

Sekera: Wow, that’s really interesting. Thank you. So, really how do you forecast the additional power requirement for artificial intelligence, and how do you really incorporate that into your own financial models in order to determine what the proper valuation of utilities should be going forward?

Miller: We’re forecasting about double demand growth in data center electricity consumption throughout the year on an annualized basis. Now that’s going to be very regional. So only certain utilities are going to benefit from that. The utilities we think are going to benefit the most we incorporate that in our future projections, that flows into our earnings growth projections and ultimately in our valuation. So, utilities we think that are going to benefit from data center demand growth do get a premium valuation relative to the rest of the group.

Sekera: All right, well then, let’s name some names. Which ones should investors keep on their shortlist to watch in the marketplace today for any potential selloffs or ones that might be undervalued going forward?

Miller: We like three in particular: WEC Energy Group, it’s a Wisconsin utility; NiSource, and its subsidiary in Indiana; and also Entergy, which is located in the Gulf Coast. All three of those are huge centers for demand growth from data centers right now. What we’ve seen in Wisconsin is data centers that Microsoft has planned could increase electricity demand in Wisconsin by 10% in the entire state relative to one Microsoft data center development. In Northwest Indiana, the utility there, NiSource, has projected that one or two data centers could double the amount of demand in Northwest Indiana. And Entergy earlier this year got a commitment from Amazon Web Services for a $10 billion facility. We think that could lead to a 10% or more increase in Entergy’s capital investment growth plan. So, all three of those are projecting higher than average earnings growth because of the data center demand, and we think that could continue for a decade or more, giving them a higher valuation relative to the group.

Sekera: Wow, that’s some really interesting growth, especially for a sector that always has been considered a very slow but steady earnings grower. So, thank you very much for your time. Really appreciate your insight.

Dziubinski: It’s time to move on to the picks portion of this week’s program. This week’s picks are some of the names that Dave talked about in his interviews with Brian and Travis. Now we’re focusing on stocks that score well on three marks. First, they’re undervalued, according to Morningstar, which means they carry less price risk today. Second, the companies have somewhat reliable cash flows as measured by their medium or low Morningstar Uncertainty scores. And then finally, they’ve carved out economic moats, which means Morningstar expects these companies to remain competitive for a decade or longer. So the first pick this week is Microsoft, which is clearly emerged as an early leader in AI with its investment in OpenAI. The company is an excellent financial health, has pretty reliable cash flows, and Morningstar thinks Microsoft will continue to strengthen its already wide economic moat.

Microsoft is an all-around attractive core holding, and it looks about 17% undervalued relative to our fair value estimate. Now, our second stock pick this week is Taiwan Semiconductor Manufacturing. TSMC is the world’s largest dedicated chip foundry and maintains more than a 60% market share. Morningstar thinks the company has carved out a wide economic moat.

Second-quarter earnings and third-quarter forecasts were in line with our full-year estimates. Now, Morningstar’s analyst says that negative recent news, like US presidential candidate Donald Trump’s remarks on Taiwan and reports on possible new export restrictions, should really have limited direct impact on TSMC’s operations, as most of its direct customers are from the US or Taiwan. The stock looks 30% undervalued today.

Now, the last three picks this week are some of the second derivative plays that Dave talked about with Travis. The first is WEC Energy Group. Now we think this utility should trade at a premium to other utilities due to its above-average earnings growth and the constructive regulatory environments in which it operates. Also, Microsoft announced additional land purchases in southeastern Wisconsin, indicating further data center development, which will benefit WEC.

We assign the utility a narrow economic moat rating. The stock offers a nice yield above 3.5% and looks about 6% undervalued. NiSource is the second second derivative AI pick. Like WEC, NiSource will also benefit from Microsoft’s data center plans, but this data center will be in Laporte, Indiana, which lies within NIPSCO territory. NiSource also earns a narrow economic moat rating.

The stock yields more than 3% and is 6% undervalued relative to Morningstar’s fair value estimate. And our final stock pick this week is Entergy. Morningstar assigns Entergy a narrow economic moat rating. We’re projecting 7% annual earnings growth at least through 2029 on the back of accelerating electricity demand from data center development and Energy’s Gulf Coast service area. The stock is just 4% undervalued and yields more than 3.5%.

Viewers interested in researching any of the stocks that we discuss on The Morning Filter can visit morningstar.com for more analysis. I hope you’ll join me for The Morning Filter again next Monday at 9 a.m. eastern, 8 a.m. central. In the meantime, please like this video and subscribe to Morningstar’s channel. Have a great week!