If you build modern software for scale, speed, and reliability, the ground beneath your stack is shifting faster than ever. Recent data from two recently released comprehensive Cloud Native Computing Foundation (CNCF)-sponsored research reports provide insight into how developers are responding to these rapid changes and which technologies developers are beginning to trust the most in cloud native and artificial intelligence (AI) environments.
In this report, we will share the insights we gained from both the "State of Cloud Native Development Report" and the "Cloud Native Computing Foundation Technology Radar", both of which were developed in conjunction with SlashData and include feedback from a significant number of developer participants. When you combine these reports, you will get a very clear and complete picture of how developers are changing their infrastructure by utilizing the latest tools, technology, techniques, processes, and personnel that are influencing the overall architecture of software development in general.
Cloud native becomes the backbone of modern development
Cloud Native computing is no longer considered a subset of available tools. It has quickly developed into a foundation for building, deploying, and scaling modern software applications. An estimated 32 percent (approximately 15.6 million) of all software developers globally will identify themselves as Cloud Native Developers by approximately Q3 of 2025. Of this group, an estimated 9.3 million+ are back-end Developers and an estimated 7.1 million+ work in the area of Machine Learning/Artificial Intelligence (ML/AI).
The largest number of Cloud Native Developers, at 56%, are Back-End Developers, representing strong year-over-year growth, indicating that the way back-end systems are being developed relies heavily upon the use of containerisation, orchestration, and observability techniques. Additionally, Cloud Native Development is expanding beyond the traditional role of Back End Developers, as DevOps professionals now represent 58% of the Cloud Native population, and swiftly following suit are Web Developers (full stack).
Further illustrating this growth is the trend toward broader implementation of Cloud Native principles, beyond just microservices and Kubernetes, as Developers begin to implement other Cloud Native development concepts such as API gateways, Event Driven Architecture (EDA), Streaming Services, and Platform Engineering tools to build resilient and scalable systems.
Hybrid and multi-cloud reshape infrastructure strategy
The reports reveal a steady rise in hybrid and multi-cloud deployment models. Hybrid cloud is now used by 30 percent of developers, while 23 percent rely on multi-cloud strategies. These approaches reflect a shift away from one-size-fits-all cloud strategies toward more flexible infrastructure models that balance performance, cost control, and compliance.
Backend developers were early adopters of these models and appear to have stabilized their choices. This suggests that broader developer communities are now catching up as organizations rethink how they distribute workloads across environments.
The cloud native landscape is also evolving in how infrastructure responsibilities are shared. Many developers are insulated from underlying complexity through internal developer platforms and managed services. This abstraction allows teams to focus on application logic while platform specialists optimize orchestration and reliability.
Production-ready AI tools gain developer confidence
According to the CNCF Technology Radar, developer confidence is currently clustered around tools and technologies related to AI. With feedback collected from 300+ industry professionals who develop solutions using technology at a global level, we can assess how developers view different types of products across several factors - Technology Lifecycle (Adopt, Trial, Assess, Hold), Technology Adoption Gap (Low, Medium, High), and Developer Network Effect (high, medium, low).
Developers are depicting an increasing level of adoption for some of the technologies listed under ML Inferencing. These include NVIDIA Triton, DeepSpeed, TensorFlow Serving, and BentoML, all of which are recommended for further development, with NVIDIA Triton standing out as having received the highest score on both maturity and usefulness. Additionally, Adlik has received the highest recommendation rating from other developers who've used it.
For ML Orchestration, Airflow and Metaflow also led in developer confidence, with Airflow being the most likely product to be recommended and Metaflow having the highest level of maturity. Both of these tools indicate a need for structured, scalable workflows for organizations that need complex machine learning workflows.
Also relevant are Agentic AI Platforms. Currently, Model Context Protocol (MCP) and Llama Stack fall under the category of adopt. Based on metrics used to evaluate technology performance, it is likely these products will see further adoption moving forward. MCP (Model Context Protocol) represents the best example of maturity, usefulness, and overall recommended technology by other developers.
Agentic AI signals the next phase of intelligent systems
The emergence of agentic AI is one of the primary themes found throughout all the reports. Through the development of agentic AIs, developers are starting to trust and to build applications that can operate autonomously while working with other agents to coordinate and form intelligent behaviours.
Developers are developing projects such as MCP Llama Stack and Agent2Agent to proactively work with the evolving ways that AI systems interact with data services and the end users of production lines or supply chain systems (e.g., automated shipping, etc.). The major advantage of these types of tools is that they are built using an open interoperable architecture, which allows developers to be able to develop more robust, secure, and flexible integrations within redeveloping technologies (e.g., cloud-based applications).
Although agentic AIs are still considered to be evolving and the early stage in the life cycle of building new types of AIs, it is expected that agentic AI will soon become the cornerstone of new AI innovations in the future. Developers are currently implementing the agentic AI platforms into their existing production environments to provide the reliability and scalability that is needed for production operations.
A three-stage evolution of cloud native maturity
In the State of Cloud Native Development report, three levels of maturity for cloud-native infrastructure are described:
- Level 1: Foundational technologies (API Gateways, Microservices) are considered ‘Core’ technologies (largely accepted and supported) of your organization. They serve as the building blocks for modern cloud-native systems.
- Level 2: Advanced technologies (Kubernetes Observability Tools, Event Driven Systems) are available to your organization and can provide an increased level of Control and Resilience.
- Level 3: Specialized technologies (Immutable Infrastructure, Chaos Engineering) that improve Reliability have limited use by a larger percentage of Developers.
The progression from Level 1 to Level 2 to Level 3 indicates both Growth and Opportunity. While foundational Cloud Native Technologies have become Mainstream, there is still considerable potential for Growth/Expansion of Advanced/Enhanced Reliability Practices that truly create Resilient Systems.
What this means for organizations and developers
The Cloud Native Computing Foundation (CNCF) has published multiple reports that summarised a series of findings about the cloud native and artificial intelligence (AI) landscape, and concluded with a set of recommendations for organisations looking to develop products or platforms in the future using the cloud-native way of development.
To successfully adopt cloud-native technologies and build products or platforms that leverage them, organisations need to concentrate on identifying production-ready tools (the tools developers trust) like NVIDIA Triton, Airflow, MCP, and Metaflow (to name but a few), while also investing time to continue their exploration of emerging platform technologies to meet their specific business or architecture needs.
In addition, through investment in cloud-native training and platform engineering, organisations can derive more value from existing technology. It is critical to prepare developers to understand the principles behind cloud-native computing, as well as being able to utilise the tools available to them effectively.
The future is built on shared insight
The reports from the CNCF reflect the collective views of the cloud-native developer community from all over the world. As such, these reports can be used as a reference for any developer or organisation building cloud-native products or platforms to see what is working in the real world and what is being developed or deployed in the future, along with what they need to be preparing for.
Whether you are upgrading your infrastructure or scaling your AI systems, the findings presented in these reports offer a common theme that moving forward, the future of software will be built on a foundation of cloud-native infrastructure, paired with intelligent automation working together to create products and services that can operate and adapt to the changes in the marketplace.
Read the entire report to ensure that your strategy aligns with the platform technologies being developed and used today, to ensure you are informed, flexible, and taking a step forward toward the development of smart, resilient systems in the future.
