It’s our last platform weekly of the year. In the last 52 weeks, over 100,000 of you have joined me every week to dissect our platform engineering world. I’ve had the chance to read 1000+ emails from you, conversations with many, and debates with a few.

At the same time, we’ve surveyed 518 for our State of Platform Engineering survey, 500+ more for the State of AI in Platform Engineering, and spoken to dozens of more about what they’re working on for 2026, and what they’re expecting.

Based on all that - and the work I am doing myself. These are our predictions for 2026.

1. Agentic infrastructure becomes standard architecture: AI agents will move from experiments to first-class platform users. Mature platforms will manage them like any other persona with RBAC, quotas, and governance. AI will autonomously run subsystems deploying, allocating resources, and evolving architecture from observed patterns. While the Platform teams will define “agent golden paths” that help agents learn and propose optimizations we can’t even think of ourselves.
   
   
2. Platforms become the safety net for AI-generated code: The “vibe coding” era will make platforms responsible for reviewing and auto-remediating AI-generated code, particularly infrastructure code (Terraform, Kubernetes manifests), because non-deterministic outputs can introduce subtle production failures or security gaps that make the speed of vibe coding basically useless.
   
   
3. Self-healing evolves into self-architecture: The most advanced platforms will go far beyond auto-scaling to self-healing, AI-driven self-architecture, dynamically re-architecting systems to meet cost and latency targets without human intervention. We (humans) will start to shift from architects to strategists, defining objectives and constraints while AI proactively handles implementation.
   
   
4. DevOps and MLOps converge into unified pipelines: Today, ML delivery is super fragmented with manual model handoffs, inference endpoints outside governance, and data science teams operating completely separately from platform engineering. By the end of 2026, mature platforms will unify app and model delivery into a single pipeline that serves devs, ML engineers, and data scientists (Likely through one shared experience).
   
   
5. FinOps becomes a hard requirement: FinOps will shift from silos, and reactive dashboards to dedicated (preventive) platform controls, with things like pre-deployment cost gates that block services exceeding thresholds before they hit production, and AI-specific budgets for token and inference spend.
   
   
6. The platform gap becomes existential: Orgs that neglect platform capabilities will accumulate “org debt” far faster than they can repay it, combining tech & operational failures that hold them massively back. The result is imploding DevEx, slower delivery, top talent will flee, security gaps, and total failure with AI, making platform engineering investment essential to success.
   
   
7. Platform teams pivot to business value engineering: Teams will finally move beyond “just” DORA, and measure and communicate ROI in business terms: revenue enabled, costs avoided, and profit center contribution. The best teams will  instrument revenue attribution, cost avoidance, and dev productivity in business terms.
   
   
8. Compliance shifts to governance-by-default: No more shifting load onto devs and pretending its progress. Compliance will be enforced at the infrastructure layer, making non-compliant deployments technically impossible. Policy-as-Code, standardized service templates, and automatic security-control injection will become baseline requirements, espeeeeecially in regulated industries.
   
   
9. Role specialization will accelerate: The “platform engineer” role will splinter into clear specializations spanning platform leadership, product management, infrastructure, DevEx, security, observability, and AI-focused roles. Clearly reflecting platform engineering’s expanding scope (and why it will increasingly shape how organizations build and run software in total).
   
   
10. Certification and professionalization emerge: Industry-standard certifications and formal training programs will continue to emerge to define baseline competencies for platform builders, accelerating the professionalization of the field. Structured curricula like our certifications will validate expertise, raise team quality, and replace ad-hoc learning with recognized, standardized platform skills and best practice.

My biggest prediction of all… things we can’t even begin to predict will happen and dominate everything. I can’t wait to see it.

What are your predictions? Any disagreements with mine? let me know!