0. Introduction: What Changed Since October 2025

In October 2025, I published "Is Bluesky's 'Decentralization' Real? The Empty Promise of DID and the Road to a WordPress-Style Ecosystem." The conclusion was straightforward: Unit A (Bluesky Social's full-stack integration) and Unit C (individual PDS operation) existed, but Unit B (partial infrastructure providers) did not. This absence was the root cause of a system that called itself "decentralized" while functioning as a centralized one.

Three months have passed. This article revisits that analysis. The short version: Unit B is emerging. Decentralization has not arrived. We are in a transitional period.

1. Unit B Realized: The Blacksky Case

1.1 Independent Infrastructure

The October article stated that running a relay requires enterprise-scale infrastructure. That claim needs correction.

Blacksky has built independent infrastructure without depending on Bluesky PBC. According to Rudy Fraser, Blacksky operates its own full-network relay (atproto.africa), built from scratch and compliant with Sync v1.1.

Source: https://bsky.app/profile/rude1.blacksky.team/post/3lo7xk2szvs2b

A New_Public article from August 2025 provides further detail. Blacksky independently operates:

  • Its own relay (atproto.africa): syncs hundreds of gigabytes daily, tracking over 36 million accounts. A "moderation relay" monitors moderation decisions across the network.

  • Its own PDS (blacksky.app): hosts user data on Blacksky's servers. Users agree to Blacksky's terms of service, not Bluesky PBC's.

  • Its own moderation services: self-hosted Ozone, automated labeling via rsky-labeler, and feed management through SAFEskies.

  • Its own app (blacksky.community): ships with Blacksky's moderation and feeds as defaults.

Source: https://newpublic.substack.com/p/how-blacksky-grew-to-millions-of

1.2 Toward an Independent AppView

Bluesky's official blog (October 2025) reports that Blacksky is developing a full-network Bluesky AppView. A separate project called "Cypher" is also underway, with plans for a "local-only posts" feature.

Source: https://docs.bsky.app/blog/protocol-checkin-fall-2025

Once the AppView is complete, Blacksky will operate PDS, relay, AppView, and moderation independently. This is Unit B, fully realized.

1.3 rsky: A Rust Implementation of AT Protocol

Blacksky's technical foundation is "rsky" (/ˈrɪski/), a Rust implementation of AT Protocol. The GitHub repository shows it aims for a full protocol implementation, including:

  • rsky-relay: full-network relay

  • rsky-pds: PostgreSQL-based PDS (a different design from the official TypeScript implementation)

  • rsky-feedgen: feed generator for the Blacksky community

Source: https://github.com/blacksky-algorithms/rsky

rsky proves that AT Protocol implementations need not depend on Bluesky PBC's official codebase.

1.4 Correcting the October Article

The October article stated that Unit B "barely exists." That was accurate at the time. As of January 2026, Blacksky functions as a concrete implementation of Unit B.

A caveat. Blacksky's success depends on the dedicated effort of developers centered around Rudy Fraser and a team of volunteer moderators. Bryan Newbold, a protocol engineer at Bluesky PBC, has provided technical support to both Blacksky and Eurosky. I confirmed this directly with Newbold.

Blacksky's independence was not achieved in opposition to Bluesky PBC. It was achieved with Bluesky PBC's support. This suggests that decentralization is an intentional design goal of AT Protocol, not an afterthought.

2. Governance Progress

2.1 PLC Directory: Toward an Independent Organization

The October article flagged the PLC Directory's centralization as a single point of failure for did:plc. Bluesky PBC responded in September 2025 by announcing plans to transfer the PLC Directory to an independent organization.

The new organization will be established as a Swiss Association. The stated reason: to provide "a trusted, neutral, and stable global home in a time of international uncertainty."

Source: https://docs.bsky.app/blog/plc-directory-org

The October 2025 protocol check-in reports that Bluesky PBC is working with lawyers to establish the Swiss Association. A WebSocket API to facilitate PLC Directory mirroring is also planned.

Source: https://docs.bsky.app/blog/protocol-checkin-fall-2025

2.2 IETF Standardization

In August 2025, parts of AT Protocol were submitted to the IETF as a Birds of a Feather proposal. Internet Drafts covering the repository format, data sync process, and architecture overview were submitted in September 2025.

Source: Wikipedia, "AT Protocol" (updated January 16, 2026)

The standardization process signals AT Protocol's transition from a private project to a public good.

3. Eurosky and Regional Sovereignty

3.1 European Independent Infrastructure

In July 2025, a group of European tech entrepreneurs announced the "Eurosky" initiative. The goal: build social media infrastructure within Europe, backed by government support, to reduce dependence on US tech companies.

Source: https://www.eurosky.social/

Eurosky's first priority is "CoCoMo" (Commons for Content Moderation), a shared moderation system for developers and startups building on ATProto.

Open Future's analysis makes the stakes clear: without independent moderation infrastructure, anything Eurosky builds remains dependent on Bluesky Social PBC's internal moderation system. CoCoMo is the foundation for breaking that dependency.

Source: https://openfuture.eu/blog/eurosky-dawns-building-infrastructure-for-sovereign-social-media/

3.2 Mapping to the October Framework

The October article proposed "Unit B1: Community Relay + Shared AppView" and gave examples like a Japanese-language relay or an academic researcher relay. Eurosky maps onto Unit B1 as a regional variant.

Eurosky remains at the concept stage as of January 2026. It has not reached the level of concrete infrastructure operation that Blacksky has. Worth watching, not yet proven.

4. Remaining Problems

4.1 User Distribution

Bluesky's official blog (October 2025) states that "99.99% of users use infrastructure operated by Bluesky PBC." Migration to independent PDSes is happening among technical users. It has not reached the general population.

mackuba.eu's analysis (August 2025) counts approximately 2,000 third-party PDSes. Most are small.

Source: https://mackuba.eu/2025/08/20/introduction-to-atproto/

4.2 AppView Concentration

As of now, Bluesky PBC is the only operator of a full-network AppView. This changes if Blacksky's AppView ships. It has not shipped yet.

The Futur project's Zeppelin once functioned as a full-network AppView. It is no longer operational.

4.3 Economic Sustainability

The economic sustainability problem raised in October is unresolved. Blacksky relies on crowdfunding. Eurosky expects public funding. The commercial hosting services envisioned in the "WordPress-style ecosystem" model have not materialized at scale.

One development: Blacksky's December 2025 anniversary post announced plans for "blacksky.tech," a one-click infrastructure deployment service. This hints at entry into PDS and moderation hosting as a business.

Source: https://blackskyweb.xyz/anniversary/

5. Conclusion: From "Early Signs" to "Transition"

The October article called Bluesky's decentralization "decentralization theater." That assessment needs partial revision.

As of January 2026, the facts are:

  • Unit B is becoming real. Blacksky operates its own relay, PDS, moderation, and app. Its AppView is in development.

  • Governance independence is advancing. The PLC Directory is moving to a Swiss Association. IETF standardization is underway.

  • Regional sovereignty has begun. Eurosky aims to build European-independent infrastructure.

  • Bluesky PBC itself is supporting decentralization. Bryan Newbold's technical assistance to Blacksky and Eurosky demonstrates that decentralization is a deliberate protocol-level goal.

The problems that remain:

  • Over 99% of users depend on Bluesky PBC. The gap between technical possibility and actual usage is wide.

  • AppView monopoly persists. Only Bluesky PBC runs a full-network AppView.

  • Economic models are unproven. No commercially sustainable model for independent infrastructure exists yet.

The overall assessment: AT Protocol's decentralization has moved from "early signs" to "transition." It has not arrived. But "decentralization theater" no longer fits either.

The "WordPress-style ecosystem" proposed in October has been partially validated by Blacksky's practice. Lower technical barriers (rsky, Sync v1.1), community-driven operations (Blacksky's moderation team), and a "local-first, global connection optional" design philosophy (the blacksky.community app) all align with the direction proposed in October.

The question going forward: does this transition converge toward genuine distribution, or does network effect lock in Bluesky PBC's dominance despite the technical possibilities? The answer depends on what happens in the next one to two years.