Microsoft surprised Wall Street by showing growth in its cloud services. Google disappointed the Street by delivering less cloud growth than they had in the past. Amazon met its growth expectations, which were lower than those of Microsoft and Google. The media and Wall Street have attributed Microsoft’s success to a faster and more aggressive commitment to generative AI, but I think that ignores an important truth, which is that it’s too early for generative AI to be a major factor in cloud service revenues. Before you can capitalize on any incremental revenue from generative AI, you have to sustain your current cloud revenue sources, and that means understanding cloud demand.
For years, perhaps for as long as a decade, there have been signs that the public cloud market wasn’t working as it was being described by vendors, analysts, and the media. For the most part, the signs were ignored until this year, when economic uncertainties caused enterprises to take a harder look at their cloud projects. They found, according to 187 of 211 enterprises who commented on this to me, that they’d made sub-optimal cloud decisions and were spending more on the cloud than was justified. Cloud growth rates slipped immediately, and this cloud-suspicion dynamic is still in play today. In fact, it’s what’s driving the shifts in the fortunes of the cloud providers.
Amazon has always been the cloud leader, but enterprises have always told me that they found Microsoft’s cloud service strategies better aligned with their own interests. Yes, even Microsoft tended to over-hype the cloud, but they at least had a vision of cloud and data center cooperation when other providers saw that relationship as a fight to the death. That vision is serving to sustain their engagement with enterprises, even as economic uncertainties and a sudden shift in the way the cloud is being covered by the media, threatens enterprise commitments.
There has never been a business case for “moving everything to the cloud”. Of those 211 enterprises, three said they believed that “some companies” who had “unusual computing requirements” might find a cloud-only strategy the best choice. The rest believed that the cloud and data center were in a perpetual partnership. I’ve always had the same view, and even if you disagree with both the enterprises and my own opinions, there should be no question that viewpoint is both a shift in enterprise attitude and a barrier to cloud growth.
There are applications for generative AI today, but they pale in comparison to the applications already generating cloud computing revenue. Sure, there may be more AI revenue in the future, but if we look at Microsoft and Google in terms of Street expectations, Microsoft was supposed to deliver 27% growth in cloud revenue and Google to deliver 26%. Microsoft got 29% instead and Google 22%. Failing to get out in front of an AI opportunity should not have cut Google’s cloud revenue growth if any AI contribution to cloud revenues was a surprise to the Street. I think that, at the least, what we saw was that Microsoft was able to stem cloud revenue growth declines and add a bit with AI, and Google could not stem the losses.
Then there’s the fact that Microsoft indicated it would likely see growth slide slightly from the levels of this quarter. That suggests that there was a pop created by AI’s coming on the scene, but that underneath that was still the basic strength of the Microsoft cloud as an enterprise service. To me, Google is seeing a greater slowing of cloud revenues because 1) like Amazon, Google depends more on startups and social/content types who are suffering in the tech wreck, and 2) because it promotes a more cloud-centric strategy when enterprises want cloud/data-center symbiosis.
We have two other AI/cloud data points to consider, too. Oracle, who had a big AI/cloud announcement and saw their stock drop despite it. That would make no sense if AI were really the future of public cloud revenues, but it’s what happened. IBM, who saw their own revenues boosted when they took an aggressive hybrid cloud position, and who actually has more AI credibility with enterprises than either Microsoft or Google according to those enterprises I had input from, isn’t presenting AI as the future of their cloud. They’re still hybrid-focused, and enterprises say they slot AI into that positioning rather than leading with it.
Decoding cloud impact on IBM’s own earnings for the quarter is difficult in part because of that very hybrid cloud focus that I believe has helped them. Overall, their earnings and revenues beat, though, and that should be a signal that IT spending isn’t all in the toilet. Software and consulting revenues were both up Y/Y but infrastructure revenue was off slightly. I read this as a validation of the hybrid, symbiotic, model of cloud to data center relationships. IBM fits AI into that, rather than trying to say that the future is all about AI alone.
Google’s problem, IMHO, isn’t that it was too slow in AI, it’s that it believes that AI is the problem it has. They suggest their Gemini model will save their cloud revenue growth, but at the same time Google itself believes that their cloud growth problem lies in the fact that enterprises want to “optimize spending” on cloud services. Does this mean that Google is admitting that enterprises are questioning the value of the cloud, and hoping that their Gemini AI will open enterprise purse strings again? That could only be true if Gemini was a major leap forward, to the point where it justified a lot of new projects, and if Microsoft’s own AI solution couldn’t somehow fit those projects too.
The fact is that Google is more vulnerable to the cloud skepticism that developed this year. They’re seen by a majority (over 80%) of enterprises as the “most innovative” in terms of the cloud, but innovation tends to be valued most when a buyer is looking to advance their commitment to a technology. That’s the opposite of what buyers say they’re doing, and what Google admits that they’re doing. Optimizing the cloud really means reducing cloud spending, which means doing less there. Google might respond with new services that would lower costs, but how would that help their own cloud revenues?
The solution to all the cloud providers’ problems is a restoration of cloud spending growth. However, the cloud is for enterprise, to an extent, the same thing that 5G was to the network operators. It’s a finite budget created by a finite benefit case. If you want cloud spending to grow, you have to grow the business case. The old notion that “everything will move to the cloud” is one I’ve never accepted, and I think that the current cloud skepticism comes out of enterprises’ having blindly accepted that notion. Now they’re demanding a real business case, not only for new projects but also for projects already approved. Half my enterprises said they believed they had cloud projects that should be reviewed, and just slightly under a third said they had already reduced cloud spending in some areas. There’s obviously a need to level-set cloud planning, and until that’s been done there’s likely to be more uncertainty on the horizon.
