The State of Federation 2026 Survey Is Open

We're running a survey to understand how organizations are adopting, operating, and scaling federated GraphQL. The goal is straightforward: build a vendor-neutral picture of what Federation looks like in production today, and why some teams haven't adopted it yet.

The required questions take 5–10 minutes. Optional sections cover governance and operations, AI and LLM usage, platform and migration decisions, incidents and risk, and open-ended lessons learned.

This survey is for platform teams, API platform owners, and engineers responsible for federated graphs. Whether you're running Federation in limited production with one or two teams, operating at enterprise scale, or still evaluating, there's a clear path through the survey for you.

If you haven't adopted Federation yet, there's a dedicated section on barriers and when you expect to revisit it.

If you've adopted and are reconsidering, that perspective is just as valuable as a success story.

The survey covers the full Federation lifecycle: adoption drivers, governance maturity, platform choices, and outcomes teams have actually realized. We're asking about everything from subgraph counts to incident patterns, across organizations ranging from early production to deployments serving billions of monthly requests.

The survey is designed to be vendor-neutral and all responses are analyzed and reported in aggregate.

This year, we're paying close attention to how AI workloads are intersecting with federated graphs.

Teams are exposing their graphs to LLMs for retrieval, function calling, and agent orchestration. Internal copilots querying your graph look different from external AI apps at scale. Some teams are running these in production. Some are piloting. And some are still figuring out what "ready" looks like.

The survey asks directly: Is AI traffic hitting your graph? What's the use case? How prepared is your architecture for a 10× increase in AI-driven query volume? Are you using MCP? Where is AI traffic stressing your architecture today?

These workloads are still early enough that there’s no clear industry baseline. We want to establish one so we can track how quickly AI adoption is accelerating, which use cases are moving to production, and where architectural gaps are showing up.

The more responses we collect, the more useful the results become.

We aim to publish a public report in Q2 2026, using anonymized, aggregate data only.

The report will include performance benchmarks by request volume, governance patterns that distinguish stable teams from those hitting limits, and data on how AI is changing the Federation landscape.

Your data directly shapes the report the whole community will use.

Frequently Asked Questions (FAQ)