Looking back at 2025, I find myself reflecting on a year that was less about dramatic new features and more about something arguably harder: making what we have actually work reliably. This was the year of focusing on stability for Holochain.

The Foundation Underneath

It's easy to forget now, but we only integrated  Kitsune2 this year. If you experienced Holochain before this transition, you'll remember the frustration: peer discovery that didn't quite discover, sync that sometimes just... didn't. DHT synchronization that could take 30 minutes or more—if it completed at all.

Kitsune2 changed that. Not through clever optimization tricks, but through a fundamental rethinking of how we approach network reliability. The result: synchronization dropped to about a minute or faster in most cases. More importantly, it became predictable. Data shows up. Peers find each other. The network does what a network should do.

This might just seem like baseline functionality, but anyone who has tried to build distributed systems knows that reliable peer-to-peer networking is genuinely hard. Getting it right matters, getting it fast comes next.

Validation That Validates

Alongside Kitsune2, we completed a first pass on our workflow updates, particularly around validation and integration. The validation pipeline had accumulated enough edge cases and failure modes that behavior was, to put it charitably, inconsistent. We went through it systematically—not adding features, but making the existing logic do what it was supposed to do.

The warrants feature represents the logical completion of this work. Holochain has always validated data, but until now there were no real consequences when validation failed. With warrants, bad actors get blocked at the network transport level, and anyone querying a compromised agent's public key receives the warrant. The immune system, as we're calling it, is now functional. It’s not complete (membrane proof enforcement during network handshaking is still coming) but it’s functional.

Seeing What We're Building

Three pieces of infrastructure came into their own this year, each essential for building with confidence:

  • Wind Tunnel reached production readiness. For the first time, we can run real-world application tests across as many nodes as we need, with comprehensive metrics from both the OS and conductor levels. The framework now runs scheduled automated tests, executes scenarios in parallel, and gives us actual data about how Holochain behaves under stress. Several bugs we caught this year only manifested under load—the kind of issues that would have been mysterious failures in production. Now we see them before you do.
  • The Build Guide fills a gap that, honestly, we should have addressed sooner. A year ago, Holochain documentation consisted of API references in cargo docs and READMEs scattered across repositories. Now there's a comprehensive, beginner-friendly guide that walks through every feature of the framework with working code snippets. If you haven't looked at it recently, take another look. It's substantively different from what came before.
  • The Roadmap represents a commitment we made this year to transparency about where we are and where we're going. It's not a marketing document—it's a live view into our backlog, our in-progress work, and our predictions about completion based on actual velocity data. You can see what we're working on, what's queued up, and how our estimates compare to reality over time. If you want to understand our priorities or hold us accountable to them, this is where to look.

The Long List

Beyond the headline items, there's been a steady stream of fixes across tx5, K2, Holochain core, and our tooling. Connection handling that gracefully re-establishes after failure. DHT sync sessions that correctly calculate all missing data. Request management that handles high volume gracefully. WebRTC library flexibility for deployment scenarios that need it.

None of these individually sounds dramatic. Collectively, they represent a codebase that behaves the way developers expect it to.

The People Who Make This Real

No framework matters without people building on it, and I want to acknowledge some of the ecosystem partners whose work this year demonstrates what Holochain makes possible.

Volla continues pushing toward privacy-respecting mobile computing. Unyt is building value accounting infrastructure for decentralized organizations. Coasys is developing ADAM and Flux, showing what semantic interoperability looks like when you take it seriously. Sensorica keeps advancing open value network infrastructure, bringing Valueflows and hREA to bear on real peer production challenges. The Carbon Farm Network is applying these tools to regenerative agriculture and climate-beneficial fiber sourcing. Arkology Studio is building a Data Commons Stack for enhanced collective sensemaking. HummHive continues to build on Holochain and deploy on Holo to help people create and share stories with choice. And DADA explores the edge where art, value exchange, and the collective meet.

On the media and discourse side,  hAppenings serves as a crucial connector and information resource for the broader community, making the ecosystem legible to newcomers and advocates alike. Tthe  Entangled Futures podcast has been exploring questions of mutuality and collective action—the philosophical territory that gives this technical work its meaning. And this year we participated in the DWeb Seminar, engaging with researchers and builders across the peer-to-peer space on the challenges that still lie ahead.

These are a few among many. The value of what we're building comes from the network of people using it, extending it, and imagining what it could become.

What This Means

We made a deliberate choice this year to limit Holochain to features known to be stable and to fix known issues in core functionality. The temptation in any project is to keep adding—new capabilities feel like progress. But there's a different kind of progress in making existing capabilities reliable enough that people can build on them without worrying about the foundation shifting.

That's what 2025 was about. The networking layer works. Validation is consistent. We can test at scale. Documentation significantly improved. These aren't the kinds of things that generate headlines, but they're the kinds of things that let you actually build applications.

We're entering 2026 with a more solid foundation than we've ever had. What gets built on it is the next chapter.

---

*Eric Harris-Braun is Executive Director of the Holochain Foundation.*