Seven proprietary content management systems built by media companies have died, been orphaned, or entered terminal decline in the last eighteen months. Four are dead. One is dying in public. Three survived. The pattern that separates them is not budget, not headcount, not whether they were loved by their editors. It is what they were built to do.
The CMS platforms that delivered content to screens are gone. The CMS platforms that understood what the content meant are still here. Distribution is commodity. Understanding is moat.
The Dead
| Platform | Company | Built For | Cause of Death | Date |
|---|---|---|---|---|
| Chorus | Vox Media | Multi-brand publishing | Disaggregation killed the parent | 2023–2025 |
| Vx | Vice Media | Multi-language distribution | $5.7B → $350M. Vice.com shut down. | 2024 |
| SuperPoster | BuzzFeed | Viral content production | Going concern. $8.5M cash. Nasdaq delisting risk. | 2023–2024 |
| Tribune stack | Tribune Publishing | Newspaper delivery | Alden Capital gutted all tech investment. | 2021+ |
Every one of these was a distribution system. Chorus routed stories across Vox Media's eight brands. Vx pushed content into 18 languages through Google NNT. SuperPoster optimized headlines for social virality. Tribune's stack moved newsprint to web. They were plumbing — well-built, sometimes elegant plumbing — but plumbing.
When the parent company died or got gutted, the CMS died with it. Because distribution infrastructure has no value independent of the distribution. No acquirer wants your publishing pipeline. They already have one, or they can rent WordPress.
The Dying
Approximately 72 employees remain from a peak of 200+. Two rounds of cuts: 54 in September 2024, ~75 in February 2026. Still serves 2,000+ sites. Only three Gold Partners worldwide. A former principal architect publicly warned of security breach risk. The Washington Post lost $77 million in 2023.
Arc XP is the most instructive death because it tried to be a product — to externalize the CMS as a business. That bet required the parent to subsidize platform development while simultaneously bleeding cash. The Post couldn't. Neither could anyone else.
Multiple agencies now market Arc XP–to–WordPress migration services. The obituary is being written while the patient is still technically alive.
The Survivors
| Platform | Company | Built For | Differentiator |
|---|---|---|---|
| Copilot | Condé Nast | Multi-tenant, multi-market content ops | AI translation, audio generation, 60 sites / 8 languages |
| Graphene-Selene | People Inc | Content intelligence / knowledge graph | Feeds D/Cipher, Inversion, five consumers |
| Vanilla | Future plc | Tight-coupling content + commerce | Seven-product ecosystem, affiliate-native |
These three survived for the same reason. Each CMS does something WordPress cannot replicate with a plugin. Copilot runs a single editorial workflow across 24 brands in 12 countries — and now generates AI translations and audio stories natively inside the CMS. Graphene-Selene doesn't just store articles; it builds a knowledge graph that D/Cipher queries for ad targeting and Inversion product teams query for content intelligence. Vanilla doesn't just publish reviews; it wires them to affiliate revenue in a way that's architecturally inseparable from the content model.
They all understand the content. They don't just deliver it.
The Absorber and the Alternative
WordPress VIP is the default destination for the dead. Vox.com migrated after Chorus. Rolling Stone and PMC brought 250,000+ articles. ADWEEK migrated and hit 100% uptime with an 82% traffic surge. Ask Media Group moved 12 sites. The Times UK. The list keeps growing. Starting at $25,000/year. Gartner Customers' Choice 2025. WordPress VIP is holding its Media Product Forum 2026 in New York, and the agenda is AI integration and editorial workflow — not basic hosting.
Brightspot is the enterprise alternative — LA Times (migrated off 19 legacy systems), POLITICO, US News, National Geographic, AP News (replatformed in six months, scaling to 80M+ daily views). Spring 2026 release shipped three features that matter: an MCP server enabling external AI agents to search and interact with content repositories, a browser-based editorial toolkit, and native CMS onboarding tours. Brightspot is the only media CMS vendor that shipped agentic infrastructure in the first half of 2026.
The Pattern
Plot every dead CMS on two axes — distribution scope and content understanding — and they cluster in the same quadrant: high distribution, low understanding. They were built to move content at scale. They moved it. And then the scale went away, or the parent company lost the ability to fund the infrastructure, and there was nothing underneath.
The survivors cluster in the opposite quadrant. Lower distribution scope (internal-only, tightly coupled), but deep content understanding — knowledge graphs, semantic models, affiliate intelligence, multi-language AI. When the ad market shifts, when Google rewrites the algorithm, when search traffic declines by 40%, the understanding layer still has value. D/Cipher still needs Graphene-Selene even if organic traffic drops. Future's affiliate commissions still need Vanilla's product mappings even if Google sunsets featured snippets.
WordPress VIP sits in the middle — broad distribution, moderate understanding — because it's commodity infrastructure with plugin extensibility. That's not a criticism. It's the correct positioning for absorbing refugees who don't need proprietary understanding.
And Brightspot is making the most interesting play: enterprise-grade distribution plus understanding features, including the first MCP server from a media CMS vendor. If you're a publisher too large for WordPress VIP and too late to build your own Graphene-Selene, Brightspot is the only remaining commercial option.
What This Means for Graphene-Selene
People Inc's content platform sits in the survival zone. It understands content — that's its entire architecture. The knowledge graph feeds five consumers. It's the reason D/Cipher can do contextual targeting without cookies, the reason Inversion product teams can query content intelligence instead of scraping HTML.
But Graphene-Selene is also internal-only. It has no public engineering blog. No MCP server. No open-source components. No external validation of its architecture beyond what D/Cipher's ad performance implies. It lives or dies entirely on People Inc's ability to fund and staff it.
That's the same bet Arc XP made — except Arc XP tried to externalize, which created a different failure mode (subsidy collapse). Graphene-Selene's failure mode would be slower: attrition of the team that understands the knowledge graph, underinvestment during a budget squeeze, accumulation of tech debt in a system with no external users to create accountability.
Three things the engineering org should do while the die-off is still happening:
1. Ship an MCP server
Brightspot did it. Condé Nast hasn't yet. Graphene-Selene's knowledge graph is the single most valuable content intelligence asset in media publishing. Making it queryable by external AI agents — LLM search, agentic assistants, partner integrations — turns an internal tool into a platform. This is the move that separates "proprietary CMS" from "content intelligence infrastructure."
2. Document the knowledge graph publicly
Not the implementation — the capabilities. What Graphene-Selene can answer. What its content model understands. Every dead CMS on this list was unknown to the industry when it died. Nobody mourns infrastructure they didn't know existed. Engineering blog posts, conference talks, architecture papers — the kind of external signal that creates accountability and attracts talent.
3. Watch Brightspot's customer list
AP News, LA Times, POLITICO, National Geographic, US News. These are not small publishers. Brightspot is building the understanding layer as a commercial product — the thing Graphene-Selene does as a bespoke internal system. If Brightspot ships content intelligence features that approach Graphene-Selene's capabilities at $200K/year, the case for maintaining a proprietary platform gets harder to defend to the board every quarter.
The CMS die-off isn't a market correction. It's a species extinction. The environmental change — the collapse of search traffic, the death of third-party cookies, the rise of AI-mediated content — killed the distribution organisms and spared the understanding organisms. The media companies that built CMS platforms to know what they had are still here. The ones that built CMS platforms to move what they had are gone.
Graphene-Selene is in the right quadrant. The question is whether People Inc treats it like a survivor or lets it slowly become the next thing on the list.