Analysis: Microsoft’s F23 Q4 and the Impact of AI (Premium)

Yesterday, Microsoft reported its financial results for the final quarter of its fiscal 2023. Here’s a closer look at that quarter, with an emphasis on the consumer businesses we care about the most about here at Thurrott.com, and whether or how the cost of AI is starting to hit the bottom line.

First, the hard numbers: Microsoft earned a net income of $20.1 billion on revenues of $56.2 billion in the quarter ending June 30 and a net income of $72.4 billion on revenues of $211.9 billion in the fiscal year with the same end date. In both cases, Microsoft experienced double-digit gains in net income (profit) and single-digit gains in revenues (earnings). Microsoft is, of course, the second-biggest company in the world by market capitalization and it appears to be doing pretty well. Ahem.

As you probably know, Microsoft has three top-level business units, and their relative sizes and positions compared to each other have changed much: Intelligent Cloud read: Azure), its biggest business, and Productivity and Business Processes (Microsoft 365/Office), its second, are both on growth trajectories, while More Personal Computing (Windows, Surface, Xbox) continues its decline.

That requires discussion: I was hoping to see improvements to Microsoft’s Windows business in particular, with the idea that we’re surely starting to come out of the post-pandemic PC slump, but that is very much not the case. Surface tanked hard, with no explanation at all. And Xbox continues hovering in the same place as ever, with declines in hardware revenues and minor growth in content revenues. This is the part of Microsoft that I, and I assume most of you, care about the most. What’s going on?

When I wrote the news article about Microsoft’s earnings yesterday, all I had to go on was its public filings, which include a press release, a presentation with high-level details about each of its three business units, and some other documents. But this morning, I can reference the post-earnings conference call that Microsoft held with analysts and the press and provide a bit more nuance and detail. So let’s dive in.

Windows

Windows is top of mind here, and this business has been struggling along with everything else related to PCs, including PC makers like Lenovo and HP and chipset makers like Intel and AMD. As noted, I keep looking for a light at the end of this slump tunnel the PC market is in. And while there have been some hopeful notes from analysts and PC makers, in particular, we have yet to see those positive thoughts translate into real-world gains.

Case in point, Windows, which sits right at the heart of the PC business.

Windows revenues from PC makers fell 12 percent in the quarter year-over-year (YOY) “primarily driven by the PC market,” Microsoft noted vaguely. And it would have been worse had it not been for an “early back-to-school inventory build” that offset the loss by 7 points. In the previous sequential quarter, Windows revenues from PC makers fell by 28 percent, so one might argue that this is an improvement. But we don’t compare quarter-to-quarter for all kinds of logical reasons. Each quarter has its own unique dynamics, and so YOY it is.

How bad are Windows revenues from PC makers? In the year-ago quarter, they fell 2 percent, but in subsequent quarters the declines were 15, 39, 28, and now 12 percent. That’s not a correction, it’s a canyon. And the PC market that is emerging from this post-pandemic period will likely be weaker than the one that went into the pandemic.

And there is other relevant data to suggest that Microsoft’s 12 percent revenue drop this quarter maps directly to what’s really happening in the PC market: IDC reported a few weeks ago that PC sales dropped 13.4 percent in that same quarter. So here we are, over a year into the post-pandemic period, and the PC market is still experiencing double-digit declines. Over the past several quarters, we’ve been able to sugar-coat the bad news by noting that these quarters were still above pre-pandemic time periods (in both revenues and unit sales), but it’s getting to the point where that is no longer the case. The excuse is wearing thin, and I’m back to wondering when the PC market (reverse) plateaus as it continues its decline.

Looking elsewhere, we see that revenues from Windows commercial products and cloud services increased by 2 percent. That’s positive, of course, but below the 6 percent growth it experienced a year ago. Subsequent quarters saw growth of 8, -3, and 14 percent, respectively. And that was literally all the data we got from Microsoft about Windows in its press release. But the post-earnings call revealed more details.

The number of PCs running Windows 11 has “more than doubled,” Microsoft said, though it of course provided no hard numbers, as it did so often for Windows 10. And Windows 11 commercial (business) deployments saw “continued growth,” an even vaguer claim without any context. Given how poorly Windows 11 is doing with businesses, one might expect the growth rates to be positive. That is, they can only go up.

This one is interesting: Combined, Azure Virtual Desktop and Windows 365 delivered over $1 billion in revenues in a 12-month period for the first time. (I assume these businesses are part of that Windows commercial products and cloud services call-out noted above.) So I guess that’s another billion-dollar business for Microsoft.

Unrelated to earnings or hard numbers, Microsoft also called out Windows 11, hilariously, as “a powerful new canvas in this new era of AI” because of the introduction (to a limited set of users in the Windows Insider Program) of Windows Copilot, which my early peek demonstrated is borderline useless. Indeed, it’s just an instance of Microsoft Edge designed to force-feed Microsoft’s online services, tracking, and advertising on unsuspecting users.

Interestingly, Microsoft addressed my PC market concerns by noting that “the consumer PC market overall was in line with expectations [meaning bad], although the early timing of back-to-school inventory builds benefitted Windows OEM [which is good]. In other words, the PC market is exactly what I think it is, unfortunately. But overall, the performance of Windows revenues in the quarter was “better-than-expected.” Huh.

Looking to the current quarter, Microsoft expects Windows revenues from PC makers to decline yet again, in the low-to-mid teens, percentage-wise, and that little gain from early back-to-school inventory will have the reverse effect. Microsoft expects “no significant changes to the PC demand environment.” And revenues from Windows commercial products and cloud services “should drive revenue growth in the mid to high-single digits.”

Surface

As I first read through Microsoft’s financials presentation for the quarter, I was looking for certain key points—Azure growth, which has been a major driver of Microsoft’s share price and market cap, the Windows PC stuff, of course, and so on—but there was a single item in the entire that halted me in my tracks. I could almost literally hear in my brain the sound of brakes straining to prevent a fast-moving car from impacting against a brick wall.

That item was Surface.

Surface got only a single mention in the report, and it was obfuscated partially because Microsoft now refers to this part of the business as “Devices.” (This is because it also includes HoloLens now, and that business has no material impact. So I will continue to call it Surface.) This is what it said.

“[Surface] revenues declined 20 percent.”

That’s it. But a related chart provided some context: in the previous two quarters sequentially, Surface revenues declined by 39 percent and 30 percent, respectively, while the two quarters before that saw revenue gains of 12 and 2 percent. That quarter with the 12 percent gain is the comparable quarter YOY.

Ah boy. I’ve been worried about Surface for a long time. And I’m surprised, frankly, that these PC do not sell better, given their modern look and feel, quality, and, frankly, similarity to Macs. Surface is the closest thing our industry has, I think, to the Mac.

Microsoft added a little color to this in the call, noting only that this result was “roughly in line with expectations.” And it expects another revenue decline in the current quarter in the mid-30s percentage-wise, “due to the overall PC market and adjustments we made in our portfolio with an increased focus on our higher margin premium products.”

That is curious on any number of levels. But let me introduce two thoughts: one, Surface has usually been positioned as a family of premium PCs, so Microsoft has apparently shifted back in that direction again (without ever mentioning it before). And that likely means that its lower-end products, which I feel tarnish the brand, will go away or be revved less often. And two, higher-margin products almost always lead to bigger revenues. So Surface either isn’t resonated yet with premium PC buyers, or its shift back to this focus is ongoing and we haven’t yet seen the results (in the form of new revisions/upgrades/models).

Put simply, there is no reason not to be worried about Surface. This product line has never been successful or profitable from what I can tell, and it should never have come to market in the first place. I bet we see some Surface models never get updated over time. And that Surface shrinks ever further going forward. I don’t see a future here.

(The word “Surface” was not uttered once in that call.)

Xbox and gaming

The pending Activision Blizzard acquisition is on everyone’s minds, of course. And, as obviously, that didn’t impact Microsoft’s financials in the previous quarter, plus Microsoft can’t really discuss too much there.  It noted only that “we continue to work through the regulatory approval process and remain confident about getting the deal done.” But if you’re interested in how Activision Blizzard will impact Microsoft’s financials each quarter, check out my article Microsoft + Activision Blizzard: Show Me the Money! (Premium). It’s not as dramatic as you might think. You know, beyond the many tens of billions of dollars that it will pay for the company.

Back to reality: Gaming revenue overall was up just 1 percent in the quarter, so it was basically flat YOY. Xbox hardware revenue (console sales plus a bit of peripherals and so on) declined 13 percent. And Xbox content and services (games and subscriptions) revenues grew modestly by 5 percent (this experienced -6, -3, -12, and 3 percent growth in the four previous sequential quarters). In other words, Xbox is gonna be Xbox, with falling console sales and decent game and subscription revenues.

In the call, Microsoft issued a bunch of vague statements about the quality and scope of the platform, none of which helped the financials. It’s looking forward to Starfield, for example. Great.

But in the hard/soft numbers categories, we learned that it set a fiscal fourth quarter high for monthly active users (MAUs), which appears to be a general number for all Xbox users, and not specific to subscriptions. This was driven, Microsoft said, “by strength off-console,” whatever that means. (PC gamers?) Xbox also experienced “record [fiscal] fourth quarter engagement,” whatever that means, across Game Pass, with hours played up 22 percent YOY.

From a revenue perspective, Gaming’s 1 percent gain was disappointing, and “lower than expected due to weakness in first-party and third-party content performance.” (We’re looking at you, Redfall.)

And Gaming’s meager gain—or perhaps it just holding its position—somehow helped to partially offset the losses that Windows experienced. We don’t know what percentage of this business unit’s revenues is recorded by Windows and Xbox, respectively, but this suggests that as Windows falls and Xbox holds steady, the percentage goes down for Windows and up for Xbox. Even without the numbers, that seems pretty clear.

Looking to the current quarter, Microsoft expects overall Gaming revenue to grow in the mid-single digits, and Xbox content and services revenues to grow in the mid-to-high-single digits, “driven by first-party and third-party content as well as Xbox Game Pass.” But it offered absolutely no guidance on console sales. Which is troubling.

AI

Ever since Microsoft announced Bing Chat and its AI efforts for both businesses and consumers, I’ve been struggling to understand, well, everything. My rolling reaction to AI and its impact on Microsoft’s customers has been something akin to the 5 stages of grief, except in this case, it’s been shock, anger, bargaining, acceptance, and hope. AI is everything from a world-ending game-changer to a productivity booster that shifts jobs, yes, but also makes us all more efficient, which means that it’s pretty much no different from any other tech advance.

Here, none of that matters. What matters is that AI is expensive, and that Microsoft will need to pass those expenses along to its customers while it strives to make AI less expensive by inventing and implementing its own custom AI-focused chipsets in its datacenters. But that’s for the future: right now, AI is costing Microsoft a lot of money.

I worked through this issue just yesterday in The Biggest Danger of AI May Be One That No One is Talking About (Premium), but the short version is that no matter how Microsoft intends to get us to pay for AI, it is footing the bill now. Surely, the cost of AI will start showing up in its financial reports. And when that does happen, it will be interesting to see how investors react. After all, they were giddy at the thought of Microsoft dominating AI. What impact will its cost have on Microsoft’s share price and market cap?

Well, we’re not finding out today. Microsoft mentioned AI just twice in its quarterly financial materials, but both were in a quote about the future from Satya Nadella. If you were looking just at the numbers, you’d have no idea that Microsoft was even investing in AI.

The post-earnings conference call likewise at first provided few clues and no answers.

“Every customer I speak with is asking not only how, but how fast, they can apply next-generation AI to address the biggest opportunities and challenges they face – and to do so safely and responsibly,” Satya Nadella said in a mega-quote that combined something he’s already said previously with some new about responsibility that addresses the critics and regulatory worries. Can we get a golf clap for the PR flack who came up with that bit of content-free nonsense?

Anyway, Nadella focused on the future, of course, at the opening of the call. He noted that AI would help customers take the best advantage of the Microsoft Cloud (a made-up business designed to show that Microsoft is competitive with Amazon AWS). He said that the firm was “investing in the new AI platform shift by infusing AI across every layer of the tech stack.” And most vaguely, that was “driving operating leverage.” If anyone knows what that means, I’m listening.

For investors, Microsoft explained that “Azure AI is ushering in new, born-in-the-cloud, AI-first workloads, with the best selection of frontier and open models, including Meta’s recent announcement supporting Llama on Azure and Windows, as well as OpenAI.” Meaning, in other words, that AI will drive growth in Microsoft’s most important (to investors) business. Perhaps the Azure growth figures will start going up again?

AI was mentioned over 50 times in this call. It impacts everything that Microsoft is doing, so it’s not just Azure. It’s the rising sea that floats all boats. It hits on each business unit. It hits on developers. It hits on so many outside companies that will need AI services and, get this, only Microsoft is big enough and positioned correctly to handle that demand. With AI, Microsoft could become an essential service, like electricity, water, and sewer. I’m paraphrasing.

What Microsoft did not discuss, and I mean not once, is how AI will actually impact the bottom line. And that lands in two ways: the expense of delivering AI capabilities and the revenues it can expect to collect from customers for providing these capabilities. Yes, everyone will use it. Yes, it’s all over the Microsoft stack. But … where is it? What’s the impact?

There were a few vague hints first. And then a single hard number.

“Microsoft Cloud [again, an imaginary business] gross margin percentage increased roughly 3 points year-over-year to 72 percent, also slightly ahead of expectations,” CFO Amy Hood noted at one point. “Excluding the impact of the change in accounting estimate for useful lives, Microsoft Cloud gross margin percentage increased slightly driven by improvements in Office 365, partially offset by lower Azure margin and the impact of scaling our AI infrastructure to meet growing demand.” (Emphasis mine.)

So there it is. Gains in Office 365, which is part of Productivity and Business Processes (and not Intelligent Cloud, which houses Azure), are being partially offset by costs associated with AI. This tells us something important: the cost of AI, and presumably future gains, will be shared between all three of Microsoft’s business units, which makes sense given Microsoft’s statements about how AI capabilities will likewise appear in products in all three units. In other words, AI financials won’t be recorded only in the Intelligent Cloud (as Azure’s are).

Related, Hood also noted that “Azure and other cloud services revenue grew 26 percent, including roughly 1 point from AI services, as expected.” So there is a contribution there too, a minor one, and not just a cost.

And then that hard number.

Capital expenditures—defined as “the money an organization or corporate entity spends to buy, maintain, or improve its fixed assets—in the quarter included finance leases of $10.7 billion “to support cloud demand, including investments in AI infrastructure.” Of that, Microsoft paid $8.9 billion in cash. So this cost wasn’t placed against a business unit or spread between all three. (Microsoft generated $28.8 billion in cash in the quarter, with a free cash flow of $19.8 billion.)

Interesting. And the investment will continue, of course. Microsoft will “accelerate investment in [its] cloud infrastructure.” And, most tellingly, it expects “capital expenditures to increase sequentially each quarter through the year as [it] scales to meet demand.” It will manage “total cost growth and operating expense in line with the demand signals [it] sees as well as revenue growth.” And it expects operating margins to remain flat in the coming year. Overall, I think this suggests that Microsoft can afford the cost of AI, basically.

The Q&A at the end of the call was focused almost entirely on AI, as I expected. And here, we get to the meat of this conversation, finally.

It started with a question about the pace at which the firm expected customers to adopt AI. Nadella’s answer to that was long-winded and, I think, inconsequential in that it’s all fairly obvious. Developers can adopt GitHub Copilot, which is evolving towards a more sophisticated offering now called GitHub Copilot X. Microsoft 365 customers will soon be able to adopt Microsoft 365 Copilot. AI is being added all across the stack, from Azure to Windows. Etc.

But that same person also asked about the impact of AI on the financials. And Amy Hood’s answer is interesting.

“I expect gross margins here to transition over time, just like they did in the prior cloud transition,” she said, a smart callback to Microsoft’s previous trigger of explosive growth. “One thing that’s different than last time [Azure as a growth driver] is that we start out in a different place with more of a shared platform, which allows us to scale those gross margins a bit faster than last time. And we do expect the pace of this adoption curve … to be faster. You’re seeing the CapEx spend accelerate in Q4 and then again in Q1. And we’ve talked about what it should look like the rest of the year.”

“Now that being said, we’re talking about all that and going through that transition while delivering in FY24 over FY23, effectively a point higher operating margins, because if it’s flat year over year, as we guided, with the headwind from the useful life change, when you correct for that, it’s about a point higher. I think the real focus here is being able to be aggressive in meeting the demand curve and focusing on the transition, and growth and gross margins, and delivering the operating leverage.”

And that maps to what I wrote above. Microsoft will manage the costs and revenues related to AI as it grows these capabilities across the stack. It is in a unique position to afford these costs, provide this infrastructure, and grow this business.

Another person asked when AI will become a “tailwind,” in that it will impact the financials in a positive way (as opposed to a headwind). But Neither Nadella nor Hood had little of substance to say to that.

But the follow-ups continued, as I knew they would. Another person asked whether the coming capital expenditures would be driven predominantly by the cost of AI or through some combination of things that might include physical datacenters, servers, or whatever else. Predictably, Hood replied that it was all of those things.

“It’s both on the datacenters and the physical basis, plus CPUs and GPUs, and networking equipment,” she answered. “Think of it in a broad sense as opposed to a narrow sense. It’s overall increases of acceleration of overall capacity … You’re seeing both accelerations and normal Azure workloads plus some of the AI workloads, [which] is partially the reason … It’s both overall commercial cloud demand and building out capacity for AI; it’s both.”

Another person asked about AI and the future of Azure growth specifically. (Nice.) And how this shift would impact the top-line finances. Here, Nadella described the cost of AI as a percentage of GDP (gross domestic product).

“What’s the tech spend?” he asked, rhetorically. “Let’s say the 5 percent of GDP is going to go to 10 percent of GDP, maybe that gets accelerated because of the AI wave. Then the question is how much of that goes to the various parts of our commercial cloud? And then how competitive are we in each layer, right? … in the second half of the next fiscal year, we’ll start getting some of the real revenue signals from [AI]. We’re looking forward to it, but we think of it long term as a third pillar [after commercial cloud, now called Microsoft Cloud, and Microsoft 365], like we thought about something like, say, Teams or SharePoint back in the day, or what have you.”

“We are a $110 to $111 billion commercial cloud [business] that has grown in the 20s [percentage-wise],” he added. “And so, therefore, we do hit [the] law of large numbers. But that said, we do think that this is a business that can have sustained high growth, which is something that we are excited about.”

“Revenue is an outcome [of this investment],” Hood added, “but it certainly does require the demand signal, requires the capital expense, and then creates the opportunity. And that’s why I think in some ways, we’re spending a little more time talking about some of that investment, is because it is the demand signal.”

After a later question about AI and Microsoft 365 Copilot—seriously, the Q&A was all AI—Hood again highlighted how AI would help Azure growth.

“Lots of these AI products pull along Azure because it’s not just the AI solution services that you need to build an app,” she said. “And so, it’s less about Microsoft 365 pulling it along or any one copilot. It’s that when you’re building these, it requires data, and it requires the AI services. You’ll see them pull both core Azure and AI Azure along with them. And I think that’s an important nuance as well.”

To which Nadella added another of those “seas that lift all boats” reminders.

“[It’s] a classic line of business extension,” he said. “When I have a copilot plug-in, that plug-in uses Azure AI, Azure meters, Azure data sources, Azure Semantic Search. You’ll see, obviously, a pull through not only on the identity or security layer, but in the core PaaS services of Azure plus the copilot extensibility in M365.”

On the consumer side, Microsoft first promoted the expansion of Bing AI capabilities and availability. But the closest we got to AI-related hard numbers is a few details about Bing, which adopted an AI-based Chat service in February that has since expanded dramatically from capability and availability standpoints.

“To date, Bing users have engaged in more than one billion chats and created more than 750 million images with Bing Image Creator,” we were told. “And Microsoft Edge took share for the ninth consecutive quarter.” OK, let’s not get too excited there: Edge’s overall usage share today is 5.27 percent, and it was 4.23 percent in December. So yes, it “took” (I’d say “gained”) share. But Safari gained share too, and its usage is five times that of Edge. And Safari isn’t even the market leader here.

But this might be the most important (if vague) point (again, as I noted just yesterday):

“More broadly, we’re growing our ad network, which is now available in 187 markets, spanning search, display, native, retail media, video, and connected TV.”

As for the future, Microsoft “continues to be excited by Bing usage signals and the longer-term opportunity as [it] invests in AI.” That is a curiously vague comment on future financials, given how specific it was about other businesses. I’m not sure what to make of that, but in reference to that linked article yet again, ad-supported services are financially tenuous whereas paid subscription services are financially much more viable. Bing falls into the former camp, and it will need to see tremendous growth to realize the benefits. And that future is not certain: Bing didn’t fare well against Google Search and it’s likely it will similarly stumble in the face of Google (and OpenAI) AI advances as consumers continue to use trusted and new products and services instead.

As with all things in life, we’ll see.