In a piece three years ago here on AI-RTZ, I outlined one of the oldest tenets in business, as espousedd by Netscape CEO Jim Barksdale in 1995:
“Gentlemen, there’s only two ways I know of to make money; bundling and unbundling.”
It’s a lesson that Microsoft in particular has taken to heart with its core products and businesses, in its fifty year history. Again, as I outlined in another AI-RTZ piece two years ago, titled “Microsoft readiers the AI ‘Bundle Hokey Pokey’”
“Microsoft’s true value could be $3.8 trillion, or $500 a share.”
“Investors worry about AI spending ROI and software disruption.”
“Software unit boasts higher profit margins than cloud segment.”
“Signs that Microsoft is open to spinning off its struggling Xbox gaming unit have put the spotlight on the tech conglomerate’s vast mix of businesses. Investors typically don’t like conglomerates, and that’s as true for Microsoft as for any other such company.”
“We calculated what Microsoft’s three main businesses—cloud, software and personal computing—would be worth if each was trading at valuations in line with rivals in those sectors. The total is between $3.5 trillion and $3.8 trillion, a big premium over the $2.9 trillion enterprise value at which Microsoft is now trading. The $3.8 trillion value translates to a per-share basis of almost $500, compared to Microsoft’s closing price on Wednesday of $384.”
“While conglomerates typically trade at a discount to their value calculated on a sum-of-the-parts basis, as we’ve done here, Microsoft’s discount seems too great. It’s a result of the software and cloud giant’s stock price falling 20% so far this year, making it by far the worst performer of any of the big tech stocks. The sell-off has dragged Microsoft stock to its lowest multiple of forward sales since early 2023, according to Koyfin data.”
One reason of concern for investors down the road could be an issue I brought up in a piece last year, highlighting how AI Infrastructure spend is ‘bending business models’ in tech. Especially the ones based in software, which see margins typically north of 80%.
“The reason isn’t a secret. Investors are jittery about whether Microsoft will get a decent return on its heavy capital spending on AI. (News broke this week that to keep costs in line, Microsoft plans to trim its workforce in the coming week.)”
The other concern is of course AI driven ‘software apocalypse’, another issue I’ve discussed at length.
“And there are also doubts about how Microsoft’s software business will fare as AI-powered alternatives emerge, like Anthropic’s Claude Cowork, an AI agent that can complete tasks such as organizing files or creating documents on behalf of a user.”
“So far, however, there’s little evidence that the software unit will suffer. Growth of Microsoft’s software business has lately accelerated, even as its profitability has improved.”
So far, so good, but questions remain.
“Regardless, many investors are downbeat about Microsoft’s prospects. John Belton, a portfolio manager at Gabelli Funds, said he is trimming a large position in Microsoft that he started amassing a couple years ago, citing concerns that the company’s widely touted AI co-working assistant, Copilot, is technologically stymied.”
“He said Microsoft’s valuation is “very reasonable” given the questions about Azure’s revenue growth and the viability of the software business.”
““Where Microsoft is trading, it’s kind of the worst of both worlds,” Belton said. “It’s caught up in the software AI disruption narrative, and then it’s also, I think, caught up in the market skepticism” about whether the lavish AI spending at large cloud computing providers such as Microsoft is justified by the revenue AI generates.”
The other question is of course how Microsoft’s position changes vis a vis its core enterprise customers. Again, a topic I just wrote about at length. Where I highlighted how CEO Satya Nadella was picking a side with his enterprise customers vs the frontier model companies OpenAI and Anthropic, who are coming for his core customers.
“One big cloud over Microsoft’s future is how its revenue mix will evolve over time. Its Azure cloud operation is growing very quickly—close to 40% annually. Including other businesses that Microsoft reports as part of its Intelligent Cloud segment—such as Windows Server, which can run on customers’ own data centers or other clouds—the growth rate falls to 29%. Azure, however, accounts for the majority of Intelligent Cloud.”
“In comparison, Microsoft’s software business, which sells corporate subscriptions to software for a range of needs from business messaging to cybersecurity, is expanding at about 16% a year.”
“Software, however—reported in Microsoft’s productivity and business processes segment—is much more profitable than cloud. It had an operating profit margin of 58% in fiscal 2025, compared with 42% for the cloud segment.”
“Assuming these relative growth rates continue, Microsoft’s overall profit margin is likely to fall over time. So far, though, it has stayed healthy, edging up to 47% in the first nine months of fiscal 2026 from 45.6% in fiscal 2025.”
So it of course comes down to long-term margin growth.
“It’s notable that the software unit’s growth rate has accelerated to 16.5% in the first three quarters of fiscal 2026 from 13% in fiscal 2025. It achieved that growth while expanding its operating profit margin, to 60.6%, up from 57.7% in the year-earlier period.”
Microsoft has ambitions for Copilot both in the enterprise and consumer fronts, as the Information notes in a separate piece on Copilot head Jacob Andreou, an EVP appointed by CEO Satya Nadella:
“Copilot still has far fewer consumer users than its competitors. Sensor Tower estimates Copilot has 38.5 million monthly active users compared to OpenAI ChatGPT’s roughly 1 billion.”
“Improving Copilot’s reputation is crucial for Microsoft, whose core Office and other applications are facing potentially existential threats from Anthropic’s and OpenAI’s AI for white-collar workers. Microsoft CEO Satya Nadella and his deputies told colleagues last fall that they were concerned the company’s Copilot AI in Office 365 wasn’t delivering on its promise to automate work meaningfully and that some companies weren’t using it much, even if they were paying for it.”
“Microsoft shares are down nearly 20% so far this year, the worst performance among the Magnificent Seven. Some large shareholders have dumped their positions, citing underwhelming Copilot quality compared to competing AI chatbots, as well as fears that new AI tools will weaken Microsoft’s productivity software juggernaut. Microsoft executives have also recently voiced concerns that the company’s AI coding tool, GitHub Copilot, is losing ground to rivals.”
“Of course, Microsoft has the advantage of serving millions of customers that have relied on its software for decades. And our financial columnist’s analysis suggests Microsoft could be worth $3.8 trillion, 30% higher than its current market capitalization. While Copilot initially saw patchy uptake among enterprise customers when it first launched in 2023, the tool—which starts at $30 per worker per month—has grown more rapidly in recent months. Microsoft disclosed in January that Copilot had 15 million paying users, and by April that number had grown 33% to 20 million; the company has also said Copilot helped accelerate revenue growth in the Office business in recent quarters. (In comparison, ChatGPT has more than 50 million paying subscribers, which include individual subscribers as well as people who access it through a corporate account.)”
“The revenue acceleration was thanks to sales of Microsoft’s AI assistant Copilot and its decision to bundle that assistant along with cybersecurity and other products in its updated enterprise software suite, E7, which it debuted in March. Microsoft executives said on its most recent earnings call, in April, that the company saw a 20% surge in the number of queries each Copilot user asked. That underscores the distribution advantage Microsoft has with its existing customers, many of whom are using Copilot within its other enterprise apps.”
“Still, Copilot’s continued growth acceleration is far from guaranteed. Alternatives from Anthropic and Alphabet are gaining popularity, too, and even if Microsoft’s Office suite retains its value as an interface through which people can interact with AI, it might lose overall pricing power if it can’t ensure Copilot keeps pace with other companies’ AI agents. Microsoft doesn’t say how much revenue its Copilot assistant generates on a stand-alone basis.”
Then of course there is the question of the software business vs the cloud business in Azure. And the ‘sum of the parts’ analysis with software and other businesses.
“To come up with a valuation for all of Microsoft, we looked at its three main businesses. Its software unit—which sells the Office 365 software and an array of other tools, and which also includes some of LinkedIn’s revenue—broadly can be likened to software purveyors Salesforce, ServiceNow and SAP. Their average operating profit multiple for the last 12 months is 27 times, according to data from S&P Global Market Intelligence.”
“That average is skewed by ServiceNow, which is growing faster than the other companies. Salesforce and SAP, which are growing more slowly, are trading at an average of 17 times. A more appropriate multiple for Microsoft might be 20. Using that multiple on what we expect Microsoft will report its software unit earned in the year that ended on Tuesday suggests the unit is worth $1.7 trillion.”
“Microsoft’s cloud business is harder to value because there aren’t any good comparisons. Its cloud unit most closely resembles corresponding units at Amazon and Google. Two of the only stand-alone publicly traded cloud firms, CoreWeave and Nebius, didn’t report profits last year, and both trade at extremely high revenue multiples, reflecting their startup-like growth rates.”
Of course, now Meta and Elon’s SpaceX/xAI are also potentially jumping into the AWS like cloud business, which means more conglomerates in the mix.
“Given the absence of a more precise comparable multiple for the cloud unit multiple, we can look at the company-level multiples of Microsoft and Google and apply that to Microsoft’s cloud unit to arrive at a ballpark price tag for the unit.”
“Notably, both Amazon and Google are trading at a historic operating profit multiple of 31. Microsoft trades at just under 19 times, the low end of its range over the past several years. Applying the higher multiple to Microsoft’s cloud business implies it is worth $1.7 trillion.”
That is one of the principal ways the author of this piece comes up with their conclusion. A methology with which I don’t concur.
They then go on from there.
“There is an argument that applying company-wide multiples to Microsoft’s cloud unit isn’t appropriate. Google, for instance, makes most of its money from advertising. Applying a discounted multiple of 25 implies it is worth $1.4 billion.”
“The smallest of Microsoft’s businesses is its personal computing unit, which includes the Surface hardware and gaming unit. We estimate it’s worth just $354 billion. That’s based on the average multiple at which Apple, HP, Sony and Electronic Arts are trading, applied to the personal computing unit’s expected fiscal 2026 profits.”
“The total of Microsoft’s three businesses is between $3.5 trillion and $3.8 trillion, depending on how the cloud unit is valued.”
“The most challenging part of assessing Microsoft’s valuation is calculating the worth of its software business, given the uncertainty about its future growth. Goldman Sachs recently valued that unit at around $500 billion, based on multiples of both revenue and profits. Its analysis suggests Microsoft shares are fairly valued at current levels.”
“What bearish investors might be missing, however, is that its software is embedded in very many big companies and institutions. Tearing that foundation out wouldn’t be simple.”
The ‘glass half full’ types come back to the traditional enthusiasts of Microsoft’s Windows franchises in the enterprise.
“I think the optimistic case is that Microsoft has so many built-in advantages,” Belton said. “It’s got a huge user base in Office; it’s got the right to win within that user base. Its users don’t really want to switch to another provider, and Copilot really should be the vehicle through which Microsoft delivers AI technology to those customers.”
The whole piece is worth a read for additional details and charts.
For my part, it’s notable that it’s one of the first ‘sum of the parts’ pieces on big tech recently that are discussing a ‘Spin-off’. The rest of the Mag 7 also discusses a closer look. Especially given the investment priorities around AI of hundreds of billions per year for the next few years.
Something that has not been a popular point of discussion until now. Maybe that is changing with current trends in massive AI Infrastructure investments in this AI Tech Wave.
And the decisions on what AI tech to bundle or unbundle around collections of businesses. As we discussed above with Microsof and Copilot AI.
The Bigger Picture, December 21, 2024 I’ve talked a lot in these pages about Small LLM AIs, even tiny ones, situated locally near users, playing a big role...
