Think AI Has Reshaped Cloud? Wait Until You See What’s Next.
Recently, I published The Future Of Cloud report and a corresponding blog post to talk about its future as abstracted, intelligent, and composable. Let’s take a closer look at the “intelligent” component and what it means amid the intense interest in generative AI today and long-term trends in cloud.
Cloud is nearly always mentioned in AI conversations and strategies, and for a good reason: Hyperscalers have the vast capacity and scope that’s needed to train large language models (LLMs) like ChatGPT and Google Bard. Such LLMs would have stayed on the whiteboard, unachieved, if cloud platforms and ecosystems did not exist to provide the inputs for training and a means of implementation. As physicist and computer scientist Stephen Wolfram points out, the neural networks that underpin ChatGPT were first described by researchers in the 1940s, decades before technology existed to create them.
Fast-forward to today. Microsoft’s investment in ChatGPT developer OpenAI positioned Azure to be the first of the big three US cloud providers to bring such capabilities to market, with the Bing search engine among the top priorities as Microsoft seizes the initiative from its rivals on a critical technology for the first time since the public cloud market emerged. Forrester estimates that data and analytics make up $40.3 billion in the cloud market in 2023. We expect it to rise to $89.5B by 2026.
The race to put LLMs into search engines involves big money, as it could divert advertising revenue that powers Google and parent company Alphabet. AWS is making its own play to be an enabler of LLM development by its customers. It is an exciting space to follow.
But what’s next?
As I point out in The Future Of Cloud report, AI is transforming not only what services are available to cloud customers but how the cloud operates — and will ultimately transform what we mean by “cloud.” To borrow a framework used by AWS to describe the shared responsibility model in cloud, it’s useful to distinguish between the AI in the cloud — the models trained by users — and the AI of the cloud, in which artificial intelligence drives automation to allow cloud user platform teams to assemble and, ultimately, compose their own clouds.
- Let’s address the latter point — AI enabling composability — first: Some of the changes may not be immediately visible to cloud customers, as the hyperscalers battling out for market share use AI to rationalize cloud operations (if any cloud provider product managers are reading this, streamlining the convoluted dependency trees among cloud services would be a good place to start). In short, AI can be used to make cloud smarter yet less complex for the user. This is the foundation that supports abstraction and better platform operations enablement. Kubernetes-based cloud-native technology will be served up via custom abstractions tuned to the needs of particular users and businesses. Even today, enterprise platform teams with solid Kubernetes experience can assimilate Kubernetes-based environments from the hyperscalers alongside Red Hat OpenShift and the Kubernetes-based platforms in Salesforce Hyperforce or SAP Business Technology Platform. Add automation and intelligence into the mix, and you have some of the initial elements of the abstracted, intelligent, and composable cloud.
- Circling back to the first point — the push for AI-powered cloud services: For the cloud providers, AI everywhere is critical, as it allows them to charge premium prices to offset the enormous investments needed to sustain and expand their global reach. They’ll have to contend not only with one another but also AI cloud startups with an inside track on NVIDIA GPUs and that can focus on highly profitable AI workloads without having to invest in low-margin core cloud services.
- Bringing it together: Customers can draw upon such offerings to build their own IT automation and AIOps while also using AI to innovate with their products and services. It is clear that AI will be a significant part of clouds, past, present, and future.
The transition to the abstracted, intelligent, and composable cloud is already underway, albeit in a vendor-specific, fragmented, and incremental manner. To learn more about how these developments will mature in coming years, read my report or book an inquiry or guidance session.