Cloud and DevOps InfoQ Trends Report 2026: AI, Resilience, Platforms, FinOps, and Sovereignty

InfoQ Editorial Team and Industry Trends

• QCon San Francisco is scheduled for November 16th through the 20th and will feature over 60 speakers across 12 tracks focused on production-level engineering practices 0s.
• The InfoQ cloud and DevOps trends report for 2026 examines current developments in the industry, specifically highlighting the impact of AI and non-AI advancements 42s.
• Matt Saunders serves as a veteran editor on the InfoQ DevOps queue and works as the VP of DevOps for Adaptivist, a company specializing in Atlassian services, platform engineering, and DevOps 1m25s.
• Mark Sylvester is an InfoQ editor on the DevOps queue and works as a platform and architecture manager for Griffith Weight, focusing on regulated enterprise environments 1m55s.
• Shrea has 24 years of experience in technology, holds multiple patents, and has authored two books, with a professional focus on technology transformation, platform engineering, software architecture, and the design of AI-native systems 2m22s.
• Renato has been an InfoQ editor on the cloud queue since 2021 and works primarily with AWS cloud technologies 3m5s.
• Stefan Wigers has been with InfoQ for nearly 10 years and currently manages the cloud queue, covering providers such as AWS, Cloudflare, and Google 3m25s.
• In his professional role, Stefan Wigers works as a domain architect for a major health insurance company in the Netherlands, where he focuses on AI-native solutions, integration platforms, and data initiatives 3m25s.
• Daniel Bryant serves as the InfoQ news manager and a product manager at Cintaso, where his current work centers on building platforms specifically for AI 4m5s.

AI Infrastructure and Agentic Systems

• Developers and their agents are increasingly interacting with cloud platforms, prompting a need to understand these new workflows. 0s
• The scale of AI infrastructure spending is unprecedented, with estimates reaching approximately $450 billion by 2025 and continuing to grow through massive investments in electricity and hardware. 42s
• Major technology companies, including AWS, Microsoft, Google, and SpaceX, are driving this significant build-out of AI infrastructure. 42s
• An "agent infrastructure arms race" is currently underway, characterized by hyperscalers integrating AI capabilities directly into their products. 1m25s
• Examples of this integration include the development of DevOps agents by AWS and Microsoft, the release of the GKE agent sandbox by Google, and Cloudflare’s dynamic workload offerings. 1m25s
• Frameworks such as Microsoft’s Citadel, which utilizes AI Foundry, are being developed to establish comprehensive governance layouts for AI operations. 1m25s
• The role of AI has shifted rapidly from an experimental tool for efficiency to a mandatory requirement imposed by leadership and boards of directors. 2m25s
• This sudden mandate for AI adoption has exposed organizational weaknesses, as many teams are not currently structured or prepared to integrate these technologies effectively. 2m25s
• Discussions regarding team topologies, such as those presented by Matthew Skelton at KubeCon London, highlight the challenges of managing AI agents within bounded contexts rather than relying on a single, universal agent. 2m55s

Cloud Reliability and Agentic Architectures

• The reliability of major cloud services has been lower than expected over the past year, with notable incidents including an AWS region outage and significant downtime in US Virginia last October 0s.
• There is a notable divide in the industry regarding the application of AI in software delivery, characterized by two distinct approaches 45s.
• One group advocates for the use of increasingly large, powerful models to solve complex problems, while another group focuses on smaller, agentic infrastructure 45s.
• The agentic approach involves utilizing specialized agents—such as coding, testing, and CI agents—that can run other agents, mirroring microservices-style architecture 45s.
• It remains uncertain whether these two paths will converge or if one will prove to be more effective for integrating AI into established human processes 45s.

Enterprise AI Adoption Challenges

• AI agents for cloud engineering were positioned in the "innovators" category of the diffusion of innovation model for 2025 1m35s.
• Enterprise adoption of AI is currently facing significant challenges related to compliance, security, and governance, which have become board-level concerns 1m35s.
• In regulated enterprise environments, the adoption of AI is often hindered by fragmented organizational structures where various teams operate in silos 2m6s.
• A common issue is that teams are accustomed to traditional ticket-based workflows and late-stage project involvement, which conflicts with the rapid implementation of AI 2m6s.
• Organizations face the risk of wasted resources when individual teams implement AI solutions independently without sharing knowledge or considering the needs of the wider organization 2m6s.

Regulatory and Organizational Silos in AI

• Organizations are increasingly integrating security and compliance teams into systems from the beginning of development to ensure all teams start with a solid foundation 0s.
• Health insurance companies face significant regulatory challenges, including requirements from the European Union such as DORA, which act as blockers for artificial intelligence initiatives 25s.
• A lack of coordination exists regarding AI initiatives, as different groups—such as classic machine learning teams and generative AI developers—often work in silos 25s.
• Some teams attempt to build their own language models or use specific platforms like Databricks, bypassing centralized services like AWS Bedrock, which complicates architectural control 25s.
• There is a tendency for teams to pursue AI solutions for problems that could be solved by existing infrastructure, such as input management systems, often driven by business pressure to adopt AI 25s.

Security and Compliance in AI Tooling

• Simon Wardley’s work is recommended for mapping ecosystems, specifically for identifying duplication and balancing innovation with commoditization 1m15s.
• The focus of AI adoption has shifted from individual productivity to team-level productivity, though significant challenges remain regarding security and compliance at the enterprise level 1m35s.
• Enterprises are struggling to manage security across a diverse landscape of open, proprietary, and cloud-provided models 1m35s.
• Companies are implementing restrictions on Model Context Protocol (MCP) to limit the exposure of tools and prevent unrestricted access to all available resources 1m35s.
• Tool consolidation is occurring within the industry, with most companies narrowing their focus to combinations of two or three specific tools such as Claude, Cursor, or Codeex 0s.
• The Agentic AI Foundation, established in late 2024, represents a significant initiative aimed at standardization, which is a critical component of security and compliance 25s.
• Managed services like Amazon Bedrock are currently favored for their ability to meet initial compliance and security requirements, serving as a convenient starting point for organizations 1m5s.

Governance and Cost Management of AI Services

• There is skepticism regarding the long-term viability of managed AI services, as the complexity of managing additional layers may not provide the same advantages as traditional managed services like databases 1m20s.
• Architects are increasingly required to perform total cost of ownership (TCO) analyses, as utilizing managed AI functionality can sometimes be significantly more expensive than self-managed alternatives 1m45s.
• Enterprises are currently at a turning point regarding governance, as widespread experimentation with cloud-based AI tools often leads to challenges when scaling 2m10s.
• Security and compliance issues arise when developers connect AI tools to internal systems via protocols like MCP, as these tools often inherit the permissions of the individual user who configured them 2m35s.
• There is a growing trend of organizations adopting AI, driven by a desire to remain competitive and avoid being left behind 0s.
• The Model Context Protocol (MCP) has faced challenges regarding adoption due to concerns about its potential to bypass existing internal organizational permissions and identity and access management (IAM) systems 0s.
• New enterprise-level tooling is emerging to address MCP security concerns, specifically by introducing centralized authorization (auth) to help manage compliance and governance issues 0s.
• Experts recommend prioritizing strong API governance by embedding security, compliance, and authorization directly into the API layer before adding secondary interfaces like portals, MCP, ServiceNow, or Jira 42s.

Platform Engineering in the AI Era

• Platform engineering teams are increasingly focused on becoming "AI-native enablers" to prevent themselves from becoming organizational bottlenecks 2m6s.
• There is significant pressure on platform engineering teams to establish company-wide standards, as inadequate internal developer platforms (IDPs) often lead individual teams to build their own "shadow platforms" 2m6s.
• Organizations are currently navigating the transition of platform engineering into the AI age, with ongoing discussions regarding the future role of internal developer platforms and agent-based systems 1m34s.
• Efforts are underway to establish platforms specifically for agentic AI, focusing on the management of language models 0s.
• Microsoft has introduced a concept referred to as "Microsoft citadel," which utilizes a hub-and-spoke model to facilitate the creation of AI platforms 0s.
• The hub-and-spoke model features a centralized AI gateway—functioning similarly to API management—and a centralized model catalog that dictates which models are permitted for use 0s.
• Spokes in this model are provisioned to enable individual teams to build their own AI solutions 0s.
• There is ongoing concern and exploration regarding organizations building in-house AI platforms to address sovereignty requirements 0s.
• Model routing is being utilized as a strategy for cost control, allowing organizations to direct simple prompts to less expensive models while reserving frontier models for more complex tasks 0s.

Evolution of Platform Engineering

• Platform engineering is currently described as being in a "holding pattern" because the focus has shifted from traditional infrastructure tasks, such as Terraform or hosting wiki pages, to data sovereignty, model hosting, and access control for frontier models 1m15s.
• Platform teams are attempting to add value at the platform level to prevent fragmented decision-making across organizations 1m15s.
• Innovation in platform engineering is currently almost exclusively related to AI, and the lack of non-AI-related innovation is viewed as a validation of the maturity of existing platform approaches 1m15s.
• Platform teams are struggling to keep pace with developers in the agentic AI space, often learning about new developments at the same time as the developers themselves 1m15s.
• The current state of agentic AI platform development is compared to the early days of platform engineering, with teams applying lessons learned from building cloud infrastructure to these new challenges 1m15s.
• Established principles from cloud and DevOps history remain relevant and applicable to current technological trends 2m25s.

Maturity of Platforms and Developer Portals

• At a recent KubeCon event, approximately 70% of attendees surveyed indicated they view platform engineering as a rebranding of DevOps 5s.
• Platform engineering is currently transitioning into the early majority phase of maturity, as the industry has moved past basic discussions regarding infrastructure vendors and Kubernetes-as-a-service options 35s.
• While serverless options like AWS Lambda exist, the complexity of integrating these services often leads users back to Kubernetes, which has matured alongside its supporting engines 55s.
• Internal Developer Portals (IDPs) have matured, with many organizations adopting tools like Backstage or various vendor-provided portals to manage the inherent complexity of the underlying systems 1m20s.
• A clear distinction is emerging between platform engineers, who remain grounded in infrastructure and Kubernetes, and developers, who operate at higher abstraction layers to improve user experience 1m40s.
• Agentic developer portals are an emerging trend, characterized by the rapid experimentation with skills, plugins, and hooks, which is expected to mature faster than previous iterations of IDPs 2m15s.
• The concept of "mechanical sympathy"—understanding the layer beneath one's own—remains relevant for developers, though the increasing number of abstraction layers makes it difficult to determine where to draw the line 2m35s.

FinOps and AI Cost Management

• FinOps is considered a strategic priority, particularly as architects must account for the costs associated with building systems and integrating third-party services 2m55s.
• The high cost of AI implementation has created new challenges, leading to practical questions about how to manage token budgets and operational workflows once those limits are reached 3m15s.
• Organizations are currently experiencing a shift in how they evaluate productivity, moving away from outdated metrics like lines of code toward a focus on outcomes and the value generated by AI tools 0s.
• There is a growing concern regarding the high cost of AI tokens, as companies transition from a period of unrestricted experimentation to a more scrutinized approach to spending on models like Opus, Fable, and Sonnet 0s.
• While FinOps tools can track specific expenditures on AI tokens, they currently lack the capability to correlate these costs directly with business outcomes or developer productivity 0s.
• Measuring productivity remains a challenge, though there is progress in defining the abstraction layers between developers and platform engineers, which helps clarify where developers add value 0s.
• A significant issue in cost management is the instinctive pressure to reduce expenses as soon as they are identified, often without a proper justification or understanding of the value being produced 0s.
• Two years ago, cost optimization focused on deterministic factors such as storage classes, data transfer, and CPU usage, which provided clear figures for analysis 1m25s.
• AI has become a major new expense, involving models, managed services, and agents that lack full visibility and often require non-deterministic, agent-based tools for cost control 1m25s.
• It is difficult to optimize AI spending because token usage is often a proxy for other activities, making it hard to determine the actual value provided by a specific team or developer 1m25s.
• A lingering uncertainty exists regarding whether current AI costs are being subsidized by large providers to encourage vendor lock-in or if these represent the actual, sustainable costs that organizations will face in the long term 1m25s.

Strategic AI Cost Control and Value Measurement

• Uncertainty regarding AI costs has become a significant focus, shifting attention away from other areas of cloud and DevOps management 0s.
• Cost control for AI consumption can be managed through AI gateway policies that monitor usage metrics and identify which teams are incurring specific costs 15s.
• Organizations can optimize spending by directing users toward appropriate models, avoiding the use of large, expensive models for simple tasks 15s.
• While cloud providers like Azure and AWS offer consoles that track costs and sustainability metrics, linking these expenditures to specific business value outcomes remains difficult and subjective 45s.
• Current tools are capable of reporting costs, but they lack the ability to definitively determine if an investment is creating actual business value 45s.
• Future FinOps agents may be able to predict whether an expenditure is valuable to a business, though these predictions are based on provided parameters rather than deterministic outcomes 1m15s.
• The FinOps Foundation notes that while major instances of waste in AI and cloud optimization have been addressed, new challenges have emerged regarding the management of numerous smaller AI agents 1m35s.
• A primary challenge currently facing organizations is the effort required to integrate and utilize these numerous smaller AI agents effectively 1m35s.

Managing Fragmented AI Resources

• Organizations are currently shifting from large-scale machine learning investments and MLOps toward managing numerous smaller, specialized coding agents and models. 0s
• The FinOps Foundation has initiated efforts related to "tokconomics" to help address the challenges of managing these fragmented AI resources. 0s
• Cloud billing complexity remains a significant issue, leading to suggestions that the industry should focus on simplifying billing structures rather than relying solely on FinOps agents to manage costs. 0s
• While FinOps tools exist to provide engineering teams with necessary cost information, these tools are often siloed between finance departments and asset owners, remaining largely invisible to the engineers who need them. 35s
• Engineering teams are increasingly concerned about costs, particularly because minor changes in AI models can lead to a tenfold increase in expenses, a risk that was not prevalent in previous computing environments. 35s

Digital Sovereignty in Europe

• Digital sovereignty has become a prominent topic of discussion, particularly within the European Union, as organizations evaluate their reliance on American hyperscalers and software systems. 1m15s
• Achieving full digital sovereignty is difficult for many organizations because they are heavily invested in American-based infrastructure, such as Oracle databases and Salesforce CRM systems, which lack clear, equivalent alternatives. 1m15s
• European cloud providers can effectively offer storage solutions, but providing equivalent platform services and software-as-a-service (SaaS) options remains a significant challenge. 1m15s
• Regulated companies in Europe are increasingly adamant about keeping data and systems within the region. 2m15s
• A notable trend among European clients involves a gradual migration of systems back to on-premises infrastructure to meet sovereignty requirements. 2m15s
• There is a growing market demand for technology platforms that explicitly support sovereign cloud capabilities. 2m35s
• Historical resistance to cloud adoption by sectors like banking, which previously insisted on using private data centers, was eventually overcome by cloud providers. 0s
• Companies are expected to resolve regulatory and compliance challenges, such as those involving US-based service providers like Salesforce, to ensure their services remain accessible to European clients. 0s
• The practice of attempting to build and host custom, localized alternatives to US-based cloud services in private settings is reportedly declining. 0s
• Improvements in the political climate and industry-led solutions are reducing the perceived severity of concerns regarding the use of US-based cloud services in Europe. 0s

Challenges of European Cloud Sovereignty

• Achieving 100% European-based sovereignty is considered difficult because dependencies exist across software, hardware, and various other infrastructure components. 35s
• While major cloud providers have introduced specific European regions, there is debate regarding the true extent of their sovereignty and whether these offerings are primarily marketing-driven. 35s
• Many alternative European cloud providers currently lack the service levels and scalability required to serve as genuine replacements for major global cloud platforms, similar to early, limited implementations of S3-compatible storage. 35s
• Data sovereignty has historically been managed through frameworks like GDPR and defined geographic boundaries. 1m25s
• The training of AI models on internet-wide data presents a unique challenge for sovereignty, as it is unclear how to secure models that are already built upon publicly shared information. 1m25s
• Major cloud providers, including AWS, GCP, and Azure, are increasing efforts to establish cloud infrastructure within Europe, though the ultimate objectives and the specific assets being secured—whether data or models—remain ambiguous. 1m25s
• The industry is currently in an early stage regarding cloud sovereignty, characterized by more unresolved questions than definitive answers. 2m6s

Critical Evaluation of AI Trends

• Predictions regarding the total replacement of junior or senior engineering roles by AI are considered overrated, as the industry is shifting toward discussions about how to operate effectively within a world of near-instant feedback loops 10s.
• The implementation of fully autonomous agents within enterprise environments is currently viewed as overrated, as there remains a necessary role for humans to stay in the loop 1m0s.
• Caution is advised regarding the "agentic" trend, as deeper architectural challenges—such as managing agentic meshes and harnesses—are more significant than the superficial, fancy features or portals often marketed by influencers 1m18s.
• Many cloud provider services, particularly those that undergo frequent renaming or rebranding, are considered overrated, with the expectation that several major services may disappear within the next year 1m55s.
• Developers are expected to shift their focus away from the specific underlying AI models being used, instead viewing them as standard assistants 2m10s.
• "Agent washing," or the marketing claim that agents can perform all tasks, is cautioned against, particularly in regulated industries like health insurance where human decision-making is mandatory 2m22s.
• Users should critically evaluate the actual value added by AI features integrated into platforms or software-as-a-service (SaaS) products, as some of these features may be redundant or less effective than existing alternatives 2m38s.

Fundamental Principles for Technology Adoption

• Selecting the appropriate tools, services, patterns, or architectural guidance remains a critical priority for organizations 0s.
• There is a notable trend of "AI washing," where existing tools are rebranded with an AI-focused veneer rather than offering genuine innovation, similar to past trends observed with DevOps branding 4s.
• A primary recommendation for navigating new technology trends is to maintain a strong focus on fundamental principles 15s.
• While new and "shiny" technology is inherently appealing, core fundamentals remain the most essential component of successful technical implementation 25s.
• The discussion concludes with the intent to distribute the insights shared across multiple formats for broader accessibility 32s.