Analysis: What Microsoft’s Q2 says about its Channel, AI, Cloud and DeepSeek

‘DeepSeek has some real innovations. Obviously, now that all gets commoditized. And it’s going to get broadly used. And the big beneficiaries of any software cycle like that is the customers,’ says Microsoft CEO Satya Nadella.

Microsoft’s disappointing results in its Azure cloud business in its latest quarterly earnings revealed that wins in the artificial intelligence space can have unintended effects on its cloud business. But don’t write off the AI revolution and don’t consider the door closed on cloud adoption.

The Redmond, Wash.-based tech giant’s leadership cast some of the blame on Azure’s performance of 31 percent growth year over year—the low end of Microsoft’s forecast and a deceleration from 33 percent to 35 percent the prior three quarters—on the vendor’s partner sales motion, with an imbalance of investments, marketing dollars and employees to make sure AI work didn’t take away from bread-and-butter migration and fundamentals work.

Microsoft CFO Amy Hood said to expect the issues to continue into at least the next quarter, forecasting Azure growth next quarter at 31.5 percent year over year, “modestly below expectations,” according to investment firm William Blair.

Microsoft’s “intelligent cloud” segment, which includes revenue contribution from Azure, missed Wall Street expectations by about US$300 million. Hood forecast the segment to bring in US$25.9 billion to US$26.2 billion next quarter, less than the Wall Street consensus of US$26.9 billion. It should be noted foreign exchange rates are also a factor for lower revenue guidance throughout the business segments.

Microsoft Q2

CRN has reached out to Microsoft for further comment.

Members of Microsoft’s 400,000-member partner ecosystem have echoed some of Hood’s observations in interviews with CRN, noting that so many Microsoft customers still have room for cloud adoption and upgrades in security and collaboration application Teams.

While it’s not clear that the executive move is related to Hood’s comments, Microsoft revealed after the earnings call a new partner organization headed by its former president of the vendor’s Europe, the Middle East and Africa business.

Concern over the partner program and performance in non-AI Azure shouldn’t distract from better-than-expected results in Microsoft’s AI wares and productivity applications suites—both areas with solution provider attachment.

And with Microsoft and its rivals spending billions to win the AI race, and AI looking to remake work for entire industries and transform what it means to be a salesperson, a programmer and a host of other job roles, it’s no surprise that this emerging technology area holds Microsoft’s focus.

Microsoft Chairman and CEO Satya Nadella (pictured) dismissed handwringing over China-based DeepSeek upheaving the vendor’s massive US$80 billion AI spending plans, saying in essence whatever advancements DeepSeek has found will be adopted by all AI vendors.

And DeepSeek’s possible success launching a competitive AI model at a fraction of the understood training cost could be a sign that AI is getting less expensive and that more users will have the ability to adopt the technology, perhaps marking faster mass penetration than previously thought.

Microsoft’s stock fall could be a reckoning brought on by DeepSeek hype. It could be Microsoft customers past the stage of getting dazzled by AI and wanting to see the return on investment and the killer application and apps that reward the early AI adopters.

Here’s more of what Microsoft’s results show about its channel partner program as well as its performance in AI and cloud.

Microsoft channel ‘challenges’

Enterprise and partner services revenue fell 1 percent year over year during the quarter, according to the vendor. Hood cited “go-to-market execution challenges” with “customers we reach through partners and through more indirect methods of selling,” attributing the challenges to the time it took to implement changes made in the summer—Microsoft’s fiscal year started July 1—around balancing AI work with “migrations and other fundamentals.”

“As you do that, you learn with your customers and with your partners on sort of getting that balance right between where to put our investments, where to put the marketing dollars, and importantly, where to put people in terms of coverage and being able to help customers make those transitions,” she said.

Hood told analysts on the call that Microsoft is “going to make some adjustments to make sure we are in balance because when you make those changes in the summer, by the time it works its way through the system, you can see the impacts and whether you have that balance right.” But she expects “impact” through the second half of Microsoft’s fiscal year.

While it’s not clear that the executive move is related to Hood’s comments, Microsoft planned to move Ralph Haupter to the role of president of the new Small, Medium Enterprises and Channel organization Feb. 1, according to a LinkedIn post Thursday.

It was not immediately clear how the new organization fits with the Global Partner Solutions organization headed by Microsoft Corporate Vice President and Chief Partner Officer Nicole Dezen.

Partners weigh in

One analyst firm understood Hood’s comments to mean that 150,000 partner customers have tried out and invested in GenAI, contributing to Azure AI revenue.

But those customers delayed implementing, rolling out and expanding traditional workloads based on CPUs, according to a Bernstein report Thursday. That showed up in the numbers as slower Azure growth and a deceleration in Azure revenue that didn’t come from GPUs.

Although enterprises creating separate, dedicated AI budgets this year instead of borrowing from other budgets for AI experiments should help cloud migrations and non-AI projects, Microsoft’s lower-than-expected guidance for the next reported quarter shows the vendor expects issues to continue for now.

Microsoft said to expect 31 percent to 32 percent growth ignoring foreign exchange, implying more than 19 percent growth in non-AI Azure, Bernstein said. The Wall Street consensus was for 32 percent to 33 percent.

“We would really be surprised not to see improvement and an Azure beat next quarter and a guide to Azure acceleration for Q4,” according to the Bernstein report. “Our comment to management: You have shown that you can deliver AI revenue better than competitors and better than we expected, but now you need to deliver on the core Azure revenue and we very much believe you can do it.”

Microsoft partners told CRN about a variety of issues with the partner program, including high bars for badging that change too often and an overemphasis on AI and Copilot over bread-and-butter work such as cloud migrations.

Customers feel that “sorely needed features, enhancements, and fixes to core products” have been put to the side “in favor of trying to build a Copilot into everything,” said one Microsoft solution provider who asked for anonymity to speak frankly about the tech giant.

“That is causing a lot of frustration,” he said. “Customers view Microsoft 365 Copilot as still being in its infancy and are not realizing the value that Microsoft continues to claim it provides.”

Microsoft needs a balance between AI innovation and services that have made it and partners essential to so many business customers, he said.

DeepSeek rivalry

Microsoft leadership sought to quell concerns about competition from DeepSeek, which appears to have created competitive AI models at a fraction of the cost of AI rivals.

“DeepSeek has some real innovations,” Nadella said on the call. “Obviously, now that all gets commoditized. And it’s going to get broadly used. And the big beneficiaries of any software cycle like that is the customers.”

Just before Wednesday’s call, Microsoft revealed in a blog post that DeepSeek R1 is now available in the model catalog on Azure AI Foundry and GitHub. However, multiple news outlets also reported that Microsoft and Microsoft-backed OpenAI were investigating whether DeepSeek improperly accessed OpenAI data.

In a Thursday report, one investment firm crowned Microsoft a superior AI vendor for its focus on model flexibility and inferencing over training. “Most other AI infrastructure providers are focused on training which is lower margins, less sticky and more impacted by the changing dynamics of the model business as recently seen with the DeepSeek announcements,” according to Bernstein.

KeyBanc praised Microsoft’s response to DeepSeek as well the vendor’s signaled flexibility in capital spending. “By embracing DeepSeek with open arms in embedding the model offering across Azure and Copilot use-cases … Microsoft was able to walk the tight line between beneficiary from potentially disruptive forces in its applications businesses and indifferently distancing itself from dependence on the brute force AI build-out in its infrastructure business where ROI grumbles have grown in decibels,” according to the investment firm. “We are taking the under on DeepSeek's long-term impact on AI and believe that AI capex will prove stickier than ROI hawks would hope.”

William Blair said that “Microsoft does not see DeepSeek fundamentally altering its capex outlook or the need for greater computing resources” in a Thursday report.

“In fact, as AI models become more efficient, AI consumption/inferencing should naturally increase, resulting in a subsequent uptick in computing demand,” the investment firm said.

AI product portfolio wins

Turning to Microsoft’s product portfolio, for partners going all in on the AI era Microsoft shared plenty of milestones to show the emerging technology is bringing in business, even as the vendor sorts out capital expenditures and tries a variety of business models.

Wins in the Microsoft AI product portfolio shared by leadership include AI services reaching a US$13 billion annual run rate revenue instead of the forecast US$10 billion. Copilot’s number of daily users doubled quarter over quarter and customers who bought Copilot in the first quarter of adoption have increased their seats tenfold over the past 18 months, according to the vendor. GitHub Copilot now serves about 150 million paid subscribers.

That US$13 billion of AI annual run rate is likely about US$10.2 billion Azure AI and about US$2.8 billion SaaS copilots—which exclude GitHub Copilot and Security Copilot. Bernstein estimated that Copilot has been sold to 2.4 percent of the commercial Office 365 install base.

Bernstein called the SaaS copilots money “a nice surprise” and guessed that Copilot Studio, Copilot for Sales and Copilot for Service were a “meaningful portion” of the revenue.

Azure AI contributed 13 percentage points to overall Azure growth, more than double year over year. Azure AI likely brought in US$1.8 billion in the quarter. Microsoft 365 Copilot likely brought in US$1.2 billion, according to William Blair.

“The broad application of AI copilots across different lines of business is helping expand Microsoft’s wallet opportunity beyond traditional IT spend as companies across all industries trial these productivity-enhancing solutions,” the investment firm said in a report Thursday.

Microsoft’s position as a gateway for capitalizing on the AI agent creation trend “is underappreciated,” Melius Research said in a report Thursday.

Some cloud, upsell positives

While Hood spent time laying out execution issues in non-AI Azure sales, the vendor did report positive momentum in some cloud wares and in productivity packages many Microsoft solution providers sell.

William Blair called Microsoft 365 Commercial Cloud growth of 15 percent year over year “ahead of expectations.” Greater adoption of the E5 license—work that sometimes comes from solution providers—helped fuel the growth.

Bernstein’s report added that “Commercial Cloud will become a larger and larger portion of the business going forward.”

“While the Cloud story is far from over for Microsoft, especially with Azure, and AI could become relatively quickly the next leg of growth for Commercial Cloud and overall Microsoft,” according to the firm.

Seat growth was 7 percent, driven by SMBs—a typical solution provider customer—and frontline worker offers. Bernstein called Dynamics 365 growth of 18 percent year over year ignoring foreign exchange “strong.”

The upcoming Windows 10 end of support appeared to help growth in Microsoft’s “more personal computing” segment, which saw flat growth year over year, better than the 3 percent to 5 percent decline Microsoft forecast.