# GitOps Grows Up: What the 2026 Argo CD Survey and Flux’s 10th Birthday Reveal About the State of Deployment Automation

- Link: https://thestackobserver.com/gitops-grows-up-what-the-2026-argo-cd-survey-and-fluxs-10th-birthday-reveal-about-the-state-of-deployment-automation/
- Published: 2026-07-13T08:33:28-04:00
- Author: Stackxx

The 2026 Argo CD User Survey results are in, and they paint a clear picture: GitOps
has moved from early adoption to mainstream infrastructure management. With 269 
respondents—a record for the project—the data reveals a community that is maturing,
scaling, and increasingly grappling with the complexities of running Argo CD at 
enterprise scale. At the same time, the broader GitOps ecosystem is celebrating 
a major milestone: [Flux turned 10 years old](https://fluxcd.io/blog/2026/07/flux-turns-10/)
on July 7, marking a decade since Weaveworks engineer Peter Bourgon made the initial
commit that would eventually give the world the term “GitOps.”

## Argo CD’s Net Promoter Score: 73.4

Let’s start with the headline number. Argo CD’s Net Promoter Score sits at **73.4**,
with over 75% of respondents saying they would recommend the tool to others. For
context, most developer tools in this space struggle to crack 40. The community 
sentiment is overwhelmingly positive—comments like “Keep doing God’s work” and “
You all rock!” pepper the survey results.

This isn’t just feel-good data. It reflects a tool that has become the _de facto_
standard for Kubernetes deployments. Sixty-six percent of respondents have been 
running Argo CD in production for more than two years. Nearly 60% of seasoned adopters
run three-quarters of their production applications on Kubernetes, and 66% of those
manage those applications with Argo CD.

## Scaling Replaces Environment Modeling as the #1 Challenge

For the first time, **scaling and performance** (34%) has overtaken “modeling environments”
as the top challenge Argo CD users face. This shift is significant—it signals that
organizations have largely figured out _how_ to structure their GitOps workflows,
but are now hitting the ceiling of what a single Argo CD instance can handle.

The numbers back this up. Forty-two percent of respondents manage **500+ Applications**
per instance, and 25% are running more than 10 separate Argo CD instances. The median
organization is no longer experimenting with GitOps; they’re productionizing it 
across dozens or hundreds of clusters.

Other pain points include RBAC and multi-tenancy (24%), troubleshooting (23%), and
environment modeling/promotions (21%). Notably, “insufficient automation” dropped
from 22% in 2025 to 13% in 2026—another sign that the community is solving the basic
workflow problems and moving on to harder ones.

## AI/ML Workloads: The New GitOps Frontier

Perhaps the most striking finding: **80% of organizations that deploy or manage 
AI/ML workloads use Argo CD**, and 60% of those do so in production. Model inference
and LLM serving are the most common use cases, with 61% and 47% of AI/ML adopters
respectively using Argo CD for those workloads.

This makes sense. Model serving endpoints are long-running, declarative deployments—
the exact pattern GitOps handles well. The biggest AI/ML-specific deployment challenges
are GPU resource scheduling (52%), managing large model artifacts (36%), and deploying
across GPU/non-GPU node pools (30%). These aren’t GitOps problems per se, but they
illustrate how GitOps tooling is being stretched into new domains.

## ApplicationSets Surge as the Community Matures

**ApplicationSets** adoption jumped from 66% in 2025 to **79%** in 2026. This is
one of the clearest signals of community maturity: ApplicationSets are the recommended
pattern for managing multi-tenant, multi-cluster deployments at scale, and their
growing adoption suggests organizations are moving beyond ad-hoc GitOps implementations
toward structured, repeatable patterns.

Sync waves and hooks—used by 54% of respondents—also indicate teams are building
sophisticated deployment orchestration directly within Argo CD rather than relying
on external CD tools. That said, the sub-project ecosystem shows some consolidation:
Image Updater dropped from 35% to 26%, while the Argo CD Agent grew as organizations
look for lighter-weight cluster connectivity.

## Environment Promotion: Manual Processes Decline, but Gates Are the New Pain Point

The promotion landscape shifted significantly. **Manual manifest updates** dropped
from 49% to 35%—a huge shift in a single year. Kargo adoption surged from 2% to 
20%, while GitOps Promoter held steady at ~8%. Custom CI scripts remain the dominant
method at 52%, suggesting most organizations still glue together their own promotion
pipelines.

But here’s the twist: while traceability across environments _dropped_ from being
the #1 pain point to #3 (38%), **release gates** (50%) and **standardized pipelines**(
44%) emerged as the top challenges. The implication is clear: as teams automate 
basic promotions, they discover that automating _intelligent_ promotions—with proper
gates, approvals, and rollback controls—is much harder.

## Installation Trends: Helm Dominates, Argo CD Operator Emerges

Helm Chart installations remain dominant at ~73%, with Kustomize holding at ~32%.
The most interesting shift: the **Argo CD Operator** grew from effectively zero 
to **8%** of installations. This mirrors a broader industry trend toward operator-
based lifecycle management and could signal a shift in how organizations think about
GitOps tooling deployment.

## Flux Turns 10: The GitOps Tool That Started It All

While Argo CD dominates mindshare, it’s worth remembering that [Flux](https://fluxcd.io/blog/2026/07/flux-turns-10/)—
the project that literally coined the term “GitOps”—just turned 10. Over the past
decade, Flux has accumulated 1,076 contributors, 44 repositories, 17,946 pull requests,
and a staggering **30.2 billion container image downloads** on Flux 2 alone.

Flux v2.9, released in late June, introduces a **CLI Plugin System** that allows
extending the Flux command line with independently versioned plugins. The release
also brings server-side apply field ignore rules for fine-grained drift control,
SOPS decryption with Age post-quantum cipher, and Kubernetes Workload Identity authentication
for OpenBao and Vault. Flux runs everywhere from 5G towers to satellites, and its
recent focus on `flux-schema` and `flux-mirror` plugins suggests the project is 
preparing for an era of stricter supply-chain security requirements.

## Platform Engineering Is the Dominant Role

The Argo CD survey also reveals something about the practitioners themselves. **
Platform Engineers** are now the largest respondent group at ~38%, with DevOps Engineers
second at ~28% (down from 36% in 2025). SRE representation roughly doubled from 
5% to 11%. Whether this reflects actual role shifts or just title inflation is debatable,
but it aligns with the broader industry trend of organizations building dedicated
platform teams to manage internal developer platforms and GitOps infrastructure.

## The Bottom Line

GitOps in 2026 is no longer a question of “should we?” but “how do we scale it?”
Argo CD’s community is growing more sophisticated—moving from basic cluster bootstrapping
to managing hundreds of applications across complex promotion pipelines. The rise
of AI/ML workloads as a primary use case, the surge in ApplicationSets adoption,
and the shift in pain points from “how do I model this?” to “how do I make it fast?”
all point to a technology that has found its footing.

Meanwhile, Flux’s decade-long journey reminds us that GitOps is fundamentally about
_how teams work together_, not just the tools they use. As Stefan Prodan, Flux’s
core maintainer, put it: “The first decade of Flux was about machines applying what
humans decide. The next is about AI agents joining that loop.”

Whether that next decade belongs to Argo CD, Flux, or something entirely new, the
survey data makes one thing clear: GitOps is now simply how Kubernetes infrastructure
gets done.