By Piotr Zaniewski, Loft Labs Head of Engineering Enablement

2024 was an eventful year in the realm of cloud-native development and Kubernetes, to say the least. Generative AI exploded and as companies rushed to leverage it, skyrocketing cloud bills made the need for cost management solutions painfully apparent. Platform engineering gained prominence as a means of standardizing workflows and unblocking developers. Not to mention, Kubernetes turned ten years old and feels almost ubiquitous in the cloud-native space. Next year we will certainly see its adoption continue to increase, but there will be exciting shifts in how it is used as teams gain enough expertise to put Kubernetes capabilities to the test in novel new ways.

Below are trends that I expect to come to fruition in 2025. I anticipate platform teams will only become more crucial to successful cloud-native development, the need for cost management will reach urgent new heights, and Kubernetes will continue to surprise us with its flexibility for diverse use cases.

Platform Engineering Becomes a Buffer for AI Adoption

Platform engineering will continue to rise in 2025, as an emerging discipline that fits the niche of needing a layer between infrastructure - whether cloud or internal services - and the developer. As platform engineers further realize their role of taking things from the outside world and molding them in ways that make sense for engineers, this will extend into acting as a buffer for AI and other new technologies. We will increasingly see AI services and infrastructure fall under the purview of platform engineering; already, these teams are questioning how to deal with colleagues using generative AI, how to expose AI to developers, and how to reduce risk.

Because platform teams are responsible for designing infrastructure for containerized workloads, adding and maintaining internal tools and scripts, 2025 will see their natural progression into a means of standardizing and making AI initiatives safer and more effective. Currently, there are really no teams in place to take on this task. We do not see a preponderance of dedicated "AI engineering" teams, so I believe platform engineers are the most natural candidate to expand into this use case. Companies will almost certainly create additional AI compliance roles, but for engineering initiatives we need a more structured approach. Thus, platform teams will see an influx of questions around what to do with AI - it will be their challenge to find solutions.

Cost Management Becomes a Race to Squeeze More from Existing Infrastructure

The need to optimize cloud costs is not new, but it has become a key concern through 2024. In 2025, the urgency of cost management will reach new heights, and the goal for cloud-native companies will be to squeeze the most possible value out of their existing infrastructure. In particular, Kubernetes is now running more workloads than ever, including AI - computational capacities are already strained as teams rush to get ahold of NVIDIA GPUs, CPUs rise in price, and cloud providers train more custom AI models.

2025 will see teams laser-focused on managing Kubernetes infrastructure better to derive more value. This may manifest in increased bin-packing on existing clusters, and utilizing tools that make Kubernetes simpler and more efficient. For example, virtualizing Kubernetes clusters with tools like vCluster is becoming more prevalent, and more companies will consider this alternative architecture in 2025. This will allow them to scale AI initiatives by encapsulating workloads in virtual clusters, which are better isolated than namespaces. Also top of mind is improving resource utilization; teams will look to restructure Kubernetes components that often sit idle, such as cert-manager and others that are only periodically used. Whether on-prem or in the cloud, "sleep mode" features that automatically scale down these unused resources will become critical. 

Kubernetes Further Solidifies As a Universal Control Plane

Kubernetes has become so widely used that questions about its basic functionality, installation, and best practices for getting started have faded from the spotlight in cloud-native circles. Through 2024, we saw a rise in engineers comfortable with Kubernetes - in 2025, Kubernetes as a tool to run containerized workloads will slip under the hood as a foundational technology, akin to how VMs are still everywhere, but are no longer a major topic of discussion. Already, there is a "hidden assumption" that when users discuss a cloud-native project or application, it is running on Kubernetes and this no longer needs to be specified.

At the same time, because of the Kubernetes API and Kubernetes Resource Model (KRM), it is so flexible that it lends itself perfectly to manage and orchestrate all kinds of things beyond pods and containers. Users are becoming confident enough to run new things on Kubernetes, and are beginning to explore exciting edge cases. For example, leveraging Kubernetes to manage existing VMs using KubeVirt; or using Crossplane to manage infrastructure; or maybe they have their own Postgres operator. In 2025, we will see a rapid expansion of Kubernetes going far beyond its original use case - this concept of having a control plane at the API level means Kubernetes will become more ubiquitous, but with its use evolving into other areas. I expect we will see some wild innovations allowing Kubernetes to manage something we would never have thought possible.

##

ABOUT THE AUTHOR

Piotr Zaniewski is the Head of Engineering Enablement at Loft Labs. He is a veteran software architect, engineering manager, and platform engineer with over 15 years of experience mastering software engineering, cloud solutions, and leadership roles. Piotr excels at architecting and implementing cloud-native technologies; leveraging tools like Kubernetes, Docker, Terraform and Crossplane to empower business operations and drive successful digital transformation initiatives. He is also a regular speaker and thought leader, and the founder of cloudrumble.net. Piotr holds a degree in IT and Economics from the University of Finance and Management in Bialystok, and a Master's in Management Psychology and Finance and Management from Kardynal Stefan Wyszynski University.