Microsoft Is the World’s Most Valuable Company. Can It Stay There?

“Do you guys make money?” she baldly asked about their AI endeavors.

It’s a good question—and one to which Nadella and Altman didn’t directly respond.

Altman started. “Obviously, Microsoft likes to make a ton of money,” he said. “Satya’s going to understand better than I will who’s gonna make money.”

Then Nadella: “You can’t have a successful platform company if the money made on the platform is not greater than the money made by the platform,” he said, referring to the business model where a company profits from transactions between multiple parties.

Nadella’s response underscores Microsoft’s ambitions to become more of a platform goliath like its Magnificent Seven brethren
Amazon.com,

Apple,
and Google parent
Alphabet.
But more important, it speaks to three critical questions for Microsoft shareholders (and would-be shareholders): How does Microsoft make money from AI, how much will it make, and when? 

It also suggests that the answers to those questions aren’t clear.

In case you haven’t noticed, Microsoft has just been killing it, with its shares closing Thursday at a record $404.87, hovering around a $3 trillion market cap, and going back and forth with Apple as the most valuable company on the planet. MSFT stock has run the table, outperforming the
S&P 500 index
over the past one, three, five, and 10 years. If you had bought $5,000 of the stock at its IPO on March 13, 1986, it would be worth over $22 million today. (Damn!)

That’s a lot of shiny, happy numbers, but it’s worth focusing on two periods in particular—the past decade and the past 14 months. First, over the past 10 years, Microsoft stock is up—poetically—almost exactly tenfold (or 1,000%, versus 161% for the S&P 500), which neatly matches the tenure of Nadella, who became CEO on Feb. 4, 2014. That run speaks to the remarkable work the then little-known Microsoft engineer has done reviving the company from its moribund state. (You also may recall that in late 2015, my Barron’s colleague Alex Eule presciently anticipated the Nadellaissance in his article “Microsoft: Faster, Stronger, Better.”)

Nadella brought Microsoft back mostly by building out the company’s Azure cloud platform, which, according to Jefferies analyst Brent Thill, has grown market share from 18% in 2016 to 37% this year, and now has $64.5 billion in revenue—mostly at the expense of Amazon Web Services. Soaring morale and buzz have ensued. (“We haven’t had this kind of excitement since I don’t know when,” said a longstanding Microsoft exec at the Davos event.) 

Nadella, 56, isn’t eager to harp on anniversaries—“I haven’t really reflected on the 10 years,” he says. “This is my 32nd year at Microsoft, and it’s the second year of AI. My years have been punctuated by three other paradigm shifts: PC client-server, the web, and mobile and cloud. Now the fourth is AI. And so I’m trying to go back to learn how to operate in year two of the other three paradigm shifts.” (Nadella declined to speak with Barron’s. All comments from him in this article come from public appearances in Davos.)

Which brings us back to the past 14 months. Since ChatGPT was released on Nov. 30, 2022—an event Nadella and others say is commensurate with the introduction of the first widely adopted internet browser, Mosaic, in 1993—Microsoft stock has been riding an AI rocket ship. Over this period it has soared 66%, versus 22% for the market, a stunning move for such a big market-cap stock.

But this latest leg up is really in anticipation of a significant jump in incremental revenue and profits coming from AI. It might pan out, but no one really knows. 

The stock certainly isn’t in the bargain basement. According to Jefferies’ Thill, who has Microsoft as his top pick for 2024 with a $450 price target, the company’s forward price/earnings ratio now stands at 32.6, versus 27.1 for the
Nasdaq Composite,
historically high for the stock. (The company will announce its fiscal 2024 second-quarter earnings this coming Tuesday.)

Microsoft investors with long memories may remember facing an eerily similar dilemma 25 years ago, after the stock climbed dramatically during the internet bubble—should they sell their position? Back then, they would have been wise to do so. After peaking in December 1999 at a split-adjusted $59.96, the stock bottomed out in March 2009 at $15.15, losing nearly 75% of its value. Shares wouldn’t close above that 1999 peak until October 2016—truly a lost epoch for MSFT shareholders.  

Things are different today, though. In 2000, Microsoft was poorly positioned for the mobile phone cycle, which was about to kick off. (The BlackBerry phone appeared in 2002, followed later by the iPhone.) “We totally missed mobile,” a senior Microsoft executive acknowledged to me. Another contrast is that in January 2000, right as the stock topped out, Bill Gates stepped down as CEO. Nadella says he isn’t going anywhere.

In fact, one of Microsoft’s great strengths is its experienced management team even beyond Nadella. Chief Financial Officer Amy Hood has been at the company for 21 years, and co-founder Gates is still a technology adviser. Then there’s Brad Smith, a 30-year veteran and the company’s top lawyer, who helped defend Microsoft against the federal government’s antitrust case in the late 1990s and early 2000s. Smith is now vice chair and president—the company’s No. 2 executive.

From those trials and tribulations, Smith and Microsoft learned the ways of Washington and beefed up the company’s lobbying efforts years before the other tech giants. Smith, a consummate insider (he’s on the boards of
Netflix
and Princeton University, an author, and a podcast host), was described as “a de facto ambassador for the technology industry at large,” by the New York Times. Attorneys representing other companies and Silicon Valley critics say that Microsoft’s Washington influence and fingerprints are everywhere—from Joe Biden’s executive order on AI, to Epic Games’ litigation against Apple, to a federal government case against Google brought by Jonathan Kanter, assistant attorney general for the Department of Justice Antitrust Division, who previously represented Microsoft in private practice. “Microsoft has this whole thing wrapped up,” complains an attorney for a competitor.

Still, Microsoft has attracted regulatory scrutiny recently in the U.S. and in Europe with its acquisition of Activision Blizzard and its relationship with OpenAI, which some maintain is controlled by Microsoft. “The companies are clearly separate,” Smith tells me. “That’s unquestionable. We are partners. If you look at the AI ecosystem, there’s one very strong and respectable competitor that is vertically integrated. That’s Google. I think the first question people should ask is, do we want to encourage partnerships? It’s critical that we do so because if we do not, there’s only going to be one company that does for AI what that one company did for search.”

There’s no question that Microsoft’s partnership with Altman’s OpenAI and its ChatGPT products has put the company in the catbird seat, surpassing Google, which has been working intently on AI for years. Google had developed critical AI technology that it shared publicly; tech that became key to the formation of OpenAI. Founded in 2015, OpenAI counted Reid Hoffman, Elon Musk, Peter Thiel, and Altman among its investors. Altman would become CEO in 2019.

AI development entails complex, cutting-edge software (OpenAI has that) and requires massive computing power (Microsoft has that), and those complementary capabilities became the basis of the symbiotic partnership. Microsoft first invested $1 billion in OpenAI in July 2019 and then anted up a staggering $10 billion a year ago. (Late last year there was a management kerfuffle where the OpenAI board sought to dismiss Altman. Microsoft supported Altman, and it was the board that essentially got the boot.) 

Microsoft is in the early stages of infusing OpenAI’s technology into all of its offerings. The company started with coding platform GitHub, and it is rapidly deploying AI enhancements into Azure and applications software such as Excel through a pan-platform rubric called Copilot, a sort of AI assistant to help customers analyze data, optimize inventory, or write a report. “Bill Gates said that the PC era was information at your fingertips,” says Nadella. “This [AI] is knowledge or expertise at your fingertips.”

Microsoft is a big company to transform. Like Alphabet, Apple, Amazon, and Facebook parent
Meta Platforms
Meta Platforms,
Microsoft has grown organically and through scads of acquisitions big and small. (The company lists 230 since 1994.) Both those internal and external sources of growth have informed the company’s direction and organization.

The company is now organized into three business units: 1) Productivity and Business Processes (subscription software for business, now called Microsoft 365—which includes Word, Excel, PowerPoint, and Outlook, as well as Microsoft Teams and LinkedIn); 2) Intelligent Cloud (Azure and other units, such as SQL server and GitHub); and 3) More Personal Computing (which includes Windows operating software, gaming platforms Xbox and Activision Blizzard, devices like Surface, plus search engine Bing and Microsoft News).

Four years ago, the three businesses were the same size (each produced around $47 billion in revenue), but not anymore, as Microsoft’s cloud business has grown some 22% annually on average since then. Last year, the company’s cloud business delivered $87.9 billion of its total revenue of $211.9 billion, and Thill is looking for it to do some $119 billion in 2025. 

“The golden goose for Microsoft is taking Azure and supercharging it with AI,” says Wedbush analyst Dan Ives, who calls this “Microsoft’s iPhone moment” and argues the company could book an additional $25 billion of cloud revenue from AI over the next two years. “We think for every $100 of [Azure] cloud spend, Microsoft can monetize an incremental $35 to $48 from AI,” he says.

How much is Microsoft making overall from AI right now? “They are in the infancy of the revenue recognition of AI because most of the Copilot AI tools are still yet to be generally available,” says Thill. “You’re going to see a slow, steady ramp-up. It will go from a low-single-digit percent of revenue to become a lot more meaningful in 2025.”

Microsoft says all systems are go. “We’re getting a tremendous amount of customer traction,” says Takeshi Numoto, Microsoft’s chief marketing officer. “Think about some of the customers we’ve announced:
Visa,

Honda Motor,

Accenture,
PwC, and KPMG. When we launched we had availability limitations, and we had a tremendous amount of customers saying, ‘Hey, look, I need much more unconstrained access to the product. Please get rid of these restrictions.’” Numoto says they did just that on Jan. 16.

Still, even Microsoft bulls offer caveats. “Microsoft [was] the first to hype this,” says Thill. “Because they pump the volume harder than anyone, they’re fully on the hook to deliver. The risk is they’ve pumped up the volume and can’t really dance that well. If you want to be contrarian, a lot of these Microsoft AI demos don’t look as good in the real world.”

Thill also believes that pricing for Copilot is going to be discounted, still providing for 10% to 20% price increases for Microsoft’s software but not what he thinks is a 100% list price increase. Numoto says the company sees no unusual pushback on pricing.

Then there are questions about how readily and how quickly customers are adopting AI.

“There’s a lot of hype about AI, and people are figuring out how to use it,” says Trish Damkroger, senior vice president of
Hewlett Packard Enterprise’s
AI Solutions. “There’s the data cleansing problem, or getting the data ready to actually use AI. And rehauling systems—changing how we’re going to redo our internal processes—that’s going to take time.” In addition, she says, finding enough people who have technical expertise is a constraint, and slowing deployment.

Last year,
Morgan Stanley
announced an AI tool based on ChatGPT that gives financial advisors access to the bank’s research. But an employee there told me that now it didn’t seem like a priority. “We got a bunch of emails about it, but nothing lately,” the source said. “No one really talks about it.” Other employees reportedly complained about the tool’s accuracy. “There are not many applications where a plausible answer is good enough,” says tech investor and sometimes skeptic Roger McNamee. “Facts matter, especially in business.” 

Morgan Stanley told Barron’s, “Feedback of the AI Assistant has been very positive, marked by strong engagement. We have seen significant increases in accuracy compared with our previous knowledge management process.”

Microsoft and Altman are circumspect about how quickly the AI revolution will unfold. 

When
Salesforce
CEO Marc Benioff asked Altman in Davos what he wanted to accomplish over the next five years, Altman replied that he hoped AI would be used to make scientific discoveries. “Whether you could do that in five years or eight to 10, we’re gonna get there,” he added.

Meanwhile, Nadella and Microsoft scientists touted an AI project that had created a battery which required 70% less lithium—replacing it with common sodium. It sounded like a major breakthrough, but when I asked the scientists about commercialization, it didn’t sound like that was in the cards anytime soon.

“I think 2024 will probably be the year where [AI] will scale,” Nadella said. Note he didn’t say “will monetize.”

Some on Wall Street have taken note.

“ChatGPT was a product demo that caught the imagination of the world,” says tech investor Glenn Hutchins. “Right now we’re in the AI hype cycle. AI will go through its trough of despair and come back out again with widespread implementation.”

What does this all mean for the rest of us? “If an investor has a one-year perspective, it doesn’t make a ton of sense to buy Microsoft here, because the expectations have risen,” says Gene Munster of Deepwater Asset Management. “This is not going to be a good year for Microsoft when it comes to AI. The breakout years are three, five years down the road.”

For now AI isn’t a “must have,” or even a “want to have,” for most businesses. It may achieve the ubiquity and promise that Nadella and Altman see for it, maybe even soon, but if and when that happens, there will be a long upcycle. Therefore, taking a wait-and-see approach makes sense. Better to be right than early. 

For kicks, I decided to ask ChatGPT whether Microsoft will succeed in the generative-AI business, and it replied: 

“I don’t have the ability to predict the future, including the success of specific companies in the generative-AI business. The success of Microsoft or any other company in this field would depend on various factors such as the quality of their technology, innovation, market demand, competition, and strategic decisions.”

Hmm. Maybe there’s something to this AI thing after all.

Write to Andy Serwer at [email protected]