What’s Next for Microsoft and AI? Separating the Hype From the Reality

Ivanna Hampton: Welcome to Investing Insights. I’m your host, Ivanna Hampton.

Microsoft MSFT is leading the pack in the AI race. But how long can Microsoft stay ahead of the competition? Morningstar Research Services’ senior equity research analyst Dan Romanoff talked about the tech company’s partnership with OpenAI in part one, “Microsoft Bets Big on AI: What Investors Should Know.” So, check it out to hear his insights. Now, Dan talks about how Microsoft is sharpening its edge, and his outlook on the company’s stock.

Big Tech and AI

Hampton: Dan, startups typically pull the bigger companies forward when there’s new technology introduced. What has motivated Big Tech and other companies to jump on AI?

Dan Romanoff: Here you have the same thing happening. In the press, what you’re hearing is that this AI is new, but I would just point out that Google GOOG bought DeepMind, I think it was in 2015. OpenAI was started in 2016. So, you’re talking about eight-ish years ago where this was going on. And I think what makes it top of mind now is the release of ChatGPT to the general public late last year. So, that is a watershed moment. But my point here is that these startups were driving the interest in it, and that’s pretty typical in technology, as you point out. And what happens a lot of times is those companies just get swallowed up by bigger companies. Microsoft has definitely bought a number of companies along the way, as has Google. And I would guess that they tried to buy OpenAI outright, and OpenAI, with its mission as a nonprofit and wanting to benefit everyone, at its foundation anyway, prevented them from selling out, figuratively and literally, to Microsoft. But they obviously had a big influence. And realistically, Microsoft and Google should have been the ones that were making these advancements in AI, and they were making incremental advancements in AI. This is a massive step forward in AI. And I guess you need a small, flexible team of basically brilliant AI champions to move the ball forward. So, that’s what you get at OpenAI. They were solely dedicated to it, and the results, I think, are pretty obvious.

What Gives the More Established Companies an Edge?

Hampton: What gives the more established companies an edge? I mean, why can they do it better?

Romanoff: When it comes to creating the model itself, just a large language model like GPT-4, for example, I already see some of my companies like Salesforce CRM and ServiceNow NOW and Adobe ADBE, they’ve created their own large language models internally at this point. So, if you’re big enough and you have the resources like those companies do, then you certainly can come up with something. Amazon AMZN and Google have both created their own AI instances as well, their own large language models. If you’re big and you have the scale, you can do it. But for a lot of my companies, I don’t think that they can afford it. I don’t think that they have the engineering chops to do it at this point. That’s not to say that they couldn’t do it. If they put their mind to it in five years, they probably could, quite frankly. But there are a lot of challenges.

In all the research I’ve done, I read an interesting stat somewhere. I think it was one of the founding fathers of what we think of as the generative AI nowadays. One of the founding fathers of advanced AI basically said it costs more to have a true thought leader in AI development than it does to have a franchise quarterback in an NFL team. So, just to put it in perspective, you’re talking about actual geniuses, people who are capable of changing the world. So, there’s only so many of them, right? That’s why Google goes out and hires someone and they pay them millions of dollars, and Microsoft does the same thing. At some point, if you are Microsoft and Google and any of these large companies, you have the scale within your computing power, you’re a hyperscaler, so you have the public cloud service and you have all those servers at your disposal, you need that to train a model and power a model. So, it becomes a limited pool pretty quickly.

Small Companies and AI

Hampton: How can smaller companies participate in AI boom?

Romanoff: The good news for the smaller companies is there is OpenAI, and you can just use the API and pay OpenAI a little bit and you therefore have access to ChatGPT or the underlying models, GPT-4, DALL·E, et cetera. I do think that the differentiation over time is going to be somewhat limited because everyone is going to have access to a model, whether it’s an OpenAI model or an Amazon model or a Google model, or if you’re a creative person, you’re going to use Adobe and you’re going to use the Firefly model or Stable Diffusion. I mean, there’s a bunch of them already. So, you’re just going to pay someone basically for access to their model. And your costs will go up a little bit. But on the back end, you’ll be able to charge your own customers a little bit more, so you can make up for it. So, that’s how smaller companies are going to access this AI boom, and they’re already doing that.

What’s Next for Microsoft and AI?

Hampton: Right. Microsoft offers a long list of products, including Microsoft 365. Where can users of Microsoft products see AI, this next wave?

Romanoff: You could still go sign up on OpenAI’s website and just get the free account to use ChatGPT, which is fun enough. And I would encourage all the listeners to go do that. But realistically, pretty soon it’s going to be pervasive, I think, in everyday products. The most visible product, from my perspective, is going to be Microsoft Office or what is now Microsoft 365. You’re going to have Copilot. They branded it as Copilot. But that’s going to be infused. It’s already in beta. It will be infused and available soon for everyone to use. So, the ability to say to your computer, “Hey, can you analyze this spreadsheet and write a 500-word summary and a memo that I’ll send out in an email?” It can do that for you. That would be pretty helpful, I think, and increase your productivity. The danger side of that is, at Morningstar, for example, what we don’t want is me telling Microsoft Word, “OK, ingest this press release from Microsoft and write a 2,000-character note about the quarter for Microsoft.” We want to avoid that. But, theoretically, that is going to be possible. And incidentally, in my ChatGPT trials, I actually did that for a company, and it was Nvidia NVDA. It was pretty fascinating in that a lot of the analysis, I would say, was incorrect. But having a couple of decades of experience in this business, if I didn’t know what the numbers were and I didn’t really know the company that well, I would have believed it. It was very well written, and it sounded like it should have been right. So, it was pretty fascinating. But you can already see AI instances like this in Amazon product searches and Google and Bing searches. You’re already dealing with AI on some level. So, it’s out there.

How to Integrate AI

Hampton: If you’re a company, what kind of investment do you need to make to really integrate AI into a product or products?

Romanoff: I think it varies depending on what you’re doing. So, in all of the hubbub that’s been going on in the last six months or eight months since ChatGPT was released, everyone is saying, “OK, AI, how do we infuse ChatGPT?” One quiet thing that happened was Adobe in March, they said we just released our new large language model for imagery and video content generation that’s called Firefly, and it will be widely available very soon. And then, they gave a demo of it. And it was amazing, and it was way more powerful, I would say, or at least in the use instances I was able to garner myself anyway of DALL·E from OpenAI. But the ability to have that in the creative workflow for Adobe is going to be a monumental shift in productivity and creativity. Because you can just into a microphone, if it’s hooked up to your computer, you can describe the image you want created, it will create it for you. If you want it to be a realistic photo, they have a massive library of stock photos to draw from. So, the power of that is incredible to me. And the efficiency gains from an employee standpoint I think are going to be great as well. So, I don’t know. It depends on what functional job you’re doing, but you can already see AI instances, depending on what you’re doing, in your software applications that you use every day. And I think it’s already having a big impact on the way people work.

One example of this is GitHub Copilot from Microsoft. It basically writes software code for you. So, in a natural language, you say, “Write me a piece of code that inserts a Buy Now button onto a website.“ And it can do that for you. You can specify the language you want it written in. I can do some basic Pascal programming. I don’t think those languages have been used in 20 or 30 years. But the ability for me to go say, “Well, write this in C++.” That would be pretty cool for me to play around with. I’m sure it helps productivity quite a bit. I mean, that’s certainly what my conversations with software developers and CIOs and company management teams say. It’s definitely increasing productivity pretty massively. So, it is pretty exciting. You can see it even if it’s not visible to everybody at this point.

Will People Have to Pay to Interact With Artificial Intelligence?

Hampton: OpenAI offers a subscription service called ChatGPT Plus for $20 a month. Should people expect to pay to engage with AI?

Romanoff: For now, ChatGPT is still free. There’s still a free instance of it. I’m not sure how long that’s going to last, quite honestly. I think at some point they will probably pull the plug on it. Because in the back end, they are paying for the compute power to have those conversations, and it’s not necessarily cheap to do that. So, I assume that at some point they will pull the plug or dumb it down and limit it even further. Because right now, if there are too many users, it will kick you off. Sometimes the response times can be slow if there are a ton of users. That stuff happens right now on the free version. Should you expect to be able to interact for free? I mean, you can always just go deal with your Verizon or Comcast customer service people and you can interact with ChatGPT all you want. Maybe that’s not anybody’s idea of fun. So, I suspect at some point the free stuff will go away. But there is open source available out there. I guess, I’m sure in some way you probably will always be able to mess around for free with somebody’s ChatGPT equivalent.

AI Hype vs. AI Reality

Hampton: The AI journey appears to be a long one. I mean, you’re saying that we’re eight years in, since 2015, OpenAI launching their nonprofit. How can investors separate AI hype from AI reality?

Romanoff: This is a conversation I have with everyone. And I would say that right now it’s basically all hype. Because there’s not a ton of products out there that are generating revenue. There’s a few of them. There is Google searching and Bing searching. Is that really the AI that you’re expecting? I think the answer to that is probably no. So, it is pretty limited at this point. So, I don’t know. The problem with all of this is the best way, I think, to play a lot of the AI investment theme is from the big companies, the hyperscalers like Microsoft, like Google, like Amazon. And I don’t think that is a popular answer with investors. They want it to be OpenAI. And OpenAI is not public, and there really is no OpenAI or equivalent that is public. All of the biggest beneficiaries, I think, at this point are the hyperscalers, including Microsoft. And basically, it’s a very small part of their portfolio. So, there’s not a ton of revenue that is directly coming from generative AI. There’s the Azure AI machine-learning platform, and that’s been around for years already. So, that’s not a new product. But it definitely will benefit from this AI movement. That’s kind of the thing that is almost painful, it seems, for people to hear. But that’s the answer right now. Maybe OpenAI eventually shuns a full Microsoft takeover, and they try and go public, and then that will be obviously a great way to invest in this theme. In the meantime, Microsoft owns 49% of it. So, if they go public and they get $1 trillion valuation, then Microsoft gets $500 billion to then go invest in the next OpenAI and the next Nvidia. The cycle continues.

The Chase in the AI Race

Hampton: Right now, Microsoft is leading this AI race. Who is the closest to them to catch up?

Romanoff: I would say Microsoft is a leader. And I’ve talked about this with Ali Mogharabi, our Google analyst here internally. And we agree on this. People want there to be a leader also. And I think Microsoft is a leader. They’re not the leader. Google has made a lot of advancements. I think in a lot of instances, they were there first with their DeepMind acquisition and infusing that into the Google search. I mean, it’s pretty advanced AI going back for years already. The problem they ran into is they’re not the ones that popularized it. And that was a ChatGPT moment in October and November last year. So, Google is a leader. Microsoft is a leader. And Amazon is a leader. And that’s kind of what I would expect to continue over the next few years, for sure. I mean, in this sort of industry, I think five years is a stretch to say they’re going to be a leader in five years because this is really evolving pretty rapidly. But those companies are all, I think, realistically leaders, and they’re all bringing their own products to the table and their own models to the table, and they’ve all done great work. So, it’s pretty exciting to cover any one of them. I guess I’m fortunate that I get to cover two of them. So, maybe I’m just sort of tooting my own horn here for Amazon and Microsoft. But, yeah, I think that they’re both leaders. And Microsoft is probably a little more advanced than Amazon at this point. But it’s a matter of degrees and not necessarily miles.

Microsoft Stock Outlook

Hampton: As we wrap this up, what’s your long-term view on Microsoft?

Romanoff: Right now, I think the stock is approximately fairly valued. But I think of it as a core holding for investors, a core tech holding, a core software holding. From my perspective, what I want to do with that is I want to trade around it. So, if Microsoft starts to run up a little bit much, then I want to trim back. And if it starts like 2021, where these stocks are all running, I want to trim back. But last year, when the stocks are all down 50%, then I want to add to my position. So, Microsoft is a name, they benefit from public cloud with Azure, they benefit from AI, they’re going to be one of the beneficiaries from Meta if metaverse ever actually takes off. Plus, they have all the productivity tools like Office is a monopoly, Windows is a monopoly. So, Microsoft is very well positioned for the long term. That is a wide moat. It’s one of the widest moats I think that I cover. It’s basically an unbreakable moat just between Office and Windows. And I think with Azure, there’s only two other players realistically in Google and Amazon that offer a public cloud service. So, they’re very well positioned. I’m very bullish on Microsoft over the long term. Great free cash flow generation, good growth for as big as they are. They still grow revenues at 10% plus. So, it’s a great story. I’m excited about it. I think it’s something that probably should be owned. Like I said, I think it’s a reasonably close to fairly valued right now, but that doesn’t mean you shouldn’t want to own it.

Hampton: Thank you, Dan, for this in-depth discussion on the AI boom. It’s going to be a long journey, everyone. So, buckle up.

Romanoff: Thanks for having me. It was great to talk about something as exciting as AI.

Hampton: That was part two of “Microsoft Bets Big on AI: What Investors Should Know.” Make sure you check out part one if you missed it. Thanks to Dan for sharing his outlook on Microsoft’s role in the AI boom. Thanks to video producer Daryl Lannert. And thank you for checking out Investing Insights. We appreciate it. Subscribe to Morningstar’s YouTube channel to see new videos from our team. You can hear market trends and analyst insights from Morningstar on your Alexa devices. Say “Play Morningstar.” I’m Ivanna Hampton, a senior multimedia editor at Morningstar. Take care.

Read About Topics From This Episode

After Earnings, Is Microsoft Stock a Buy, a Sell, or Fairly Valued?

Microsoft and AI: Believe the Hype?

Microsoft Earnings: Solid Performance, With Azure Strength Offset by Margin Pressure in 2024

Before Investing in AI Stocks, Consider This