I doubt there are many who don’t accept that enterprise use of the cloud is undergoing some sort of transformation. I blogged about the “why” of that last week, and so this week I want to look at what the transformation might mean, and might already be meaning, in the business of the cloud and competition for cloud spending.
The primary driver of cloud change for enterprises, recall, is cost. Users have found that the cloud is costing them more than they expected and budgeted, and so they’re looking to reduce costs. Some say this means cutting spending through some creative cost management practices, some think they’ll need to change the services that they consume, moving perhaps away from more expensive and easier-to-use web service features. A few recognize that they will need to rethink their own applications to really address costs, and to prevent them from rising again in the future.
Any time a set of market leaders has a cost problem, it means that others have a market opportunity. Cost differentiation is the easiest thing to apply; you just do some math on your own pricing. However, in this particular situation enterprises themselves are suggesting to me that the competitors that matter the most to them are competitors who offer some fundamental architectural-based differentiation, something that makes lower cost more than a simple (and very likely temporary) exercise in math.
Enterprises this year have been telling me that three cloud providers are their primary candidates for increased spending. The number one on the list is Oracle, closely followed by IBM and then Google. Since Google is already number three in the cloud revenue race, it has the best chance of actually gaining a top spot because of the transformation in enterprise views of the cloud, but IBM is seen by enterprises as being the most realistic in terms of the way the cloud is used, and Oracle is seen as having the most cost-efficient cloud tool options.
According to many enterprises, IBM was the first company, cloud provider or otherwise, who offered them a real vision of hybrid cloud, a partnership between the data center and the public cloud. This was true both of IBM mainframe customers and Red Hat customers, and in fact my chats with the latter make me think that the group was even more influenced by the hybrid cloud vision.
The decision to push a hybrid cloud story explicitly was a marketing master stroke, as far as enterprise impact goes. Red Hat is only one of multiple channels for open-source software, but the hybrid story not only gave it a special identity, it resolved the tension between cloud usage and data center reality. Very few enterprises really have any intention of moving everything to the cloud, and so the IBM positioning grounds them in a reality that enterprises have been gradually accepting themselves.
IBM is also viewed very positively in AI, which is at least for now a major factor for enterprises. Watson is a long-standing product, and in truth perhaps a bit old-hat, but enterprises say IBM has modernized their approach, creating generative AI models and a strong AI professional services practice. In fact, IBM is the only AI source that gets broad enterprise recognition, though there are some AI tools in areas like networking that also get positive marks. While enterprises are not saying that they’ll rush into cloud AI, they do believe that “AI-as-a-Service” is the form most likely to impact their operations overall.
Oracle is a story both similar and different. On the “different” side, Oracle really came at the cloud almost as a SaaS business, and that’s still their predominate strength. They had a strong set of enterprise applications and making them available in cloud form gave Oracle a natural on-ramp into the cloud business. As it happens, these applications are all those routine enterprise bread-and-butter elements that can be off-the-shelf with no more than a little customization and enhancement, and so the Oracle initiative dodged the “move-everything” drive of the major cloud vendors because it attacked things companies were willing to move and weren’t “mission-critical”.
SaaS has been effective in part because line organizations can initiate a SaaS cloud deal for the same reason that Oracle applications aren’t colliding with the data center. They can be integrated with core business activity but don’t have to be run in the data center, and in fact can present benefits if run in the cloud. Salesforce was successful for the same reason.
The “similar” piece is the AI side. Like IBM, Oracle has its hands on a lot of applications that generate company data, and that data gives them experience in decision-making and analytics, and also a fertile place to train AI models. Oracle also made an early commitment to NVIDIA for GPUs, so they have a good set of resources on hand to host AI stuff. They can also look to integrate generative AI into those applications, based on a combination of broad-level knowledge “lakes” and on private data. Their desire to exploit that opportunity, and even perhaps their recognition of it, is a big advantage. Oracle is a close second to IBM in enterprise ratings of desirable AI partners.
Google is a bit of a mixed bag here. On the one hand, they are viewed by enterprises as being the most technically advanced of the three major cloud providers, but on the other they lack any strong specific application ties to the enterprise in general and the data center in particular. While Google is the source of Kubernetes, arguably the key technology for container hosting and current data center platform software, Google doesn’t push that in sales discussions as cleverly as they could.
Enterprises say that Google’s strength in the current market is a combination of a generally better pricing structure and expertise in helping them optimize applications, either developing new ones or redoing ones that have proved inefficient or failing to deliver the performance needed. Apparently Google’s sales and technical support is superior to its sales/marketing.
On the AI front, Google is seen as being “technically ahead” of their two main cloud competitors, but behind in terms of enterprise and application knowledge. Since enterprises are still kicking AI tires, that balance of plus and minus doesn’t hurt them in the race for the Cloud Top Two, but it’s impacting their competition with both IBM and Oracle, who have that enterprise/application skill set. I also found that enterprises felt that Google’s sales/marketing fell short of that demonstrated by Amazon, Microsoft, IBM, and Oracle. In particular, enterprises said that Google wasn’t as capable of bringing top-level management along in support of cloud initiatives. Interestingly, Amazon was also criticized for this.
The most important conclusion I can draw from the current cloud dynamic is that there’s been a critical and largely unheralded set of shifts in that market. First, buyers are now looking more for providers who understand enterprise applications and less at those who understand the cloud. Second, buyers now understand that “cloud native” or “microservices” may well be cloud state of the art, but not all applications are suitable for them. Finally, buyers realize that any pay-as-you-go model, including “serverless”, is almost certainly going to drive up costs more than a well-designed, simple, VM or container model. The problem is that these realizations are unheralded, and so cloud sales/marketing hasn’t caught up with the buyer. IBM and Oracle have apparently seen this, and they’re taking advantage of the slow response of cloud giants to the new reality.
