A developer has finally put hard numbers behind one of the most repeated claims in the MCP ecosystem: that roughly half of all remote Model Context Protocol servers are dead or broken. The answer? About one in ten—significantly better than the doom-and-gloom figure that's been floating around developer communities and even appearing in published project READMEs.

How the Test Was Run

The author, who goes by parkerrrrr on DEV.to, decided to actually measure MCP server health after repeatedly seeing the 'half are dead' claim without anyone providing a source. Rather than relying on hearsay or spot-checking a few servers, they tested every single entry in the official MCP registry.

What Counts as 'Broken'

The methodology matters here. The test checked whether remote MCP servers were actually reachable and responsive—not whether their implementations were perfect or even production-ready. A server that times out, returns errors, or simply doesn't respond qualifies as broken under this definition.

Why the 50% Figure Took Off

It's unclear where the original 'half are dead' claim originated, but the author admits to repeating it themselves in documentation for a published tool. This kind of viral misinformation is common in fast-moving developer ecosystems where community perception can outpace actual data.

What This Means for MCP Adoption

A 90% success rate is actually encouraging news for developers building on or evaluating MCP. While no one wants broken servers in any ecosystem, the idea that nearly half of all MCP servers were DOA was likely deterring some teams from investing in the technology.

Key Takeaways

  • The 'half are dead' claim appears to be unfounded—actual failure rate is closer to 10%
  • Testing methodology focused on remote server reachability and basic responsiveness
  • Community members (including this article's author) had been spreading unverified statistics
  • MCP ecosystem health is better than feared, though registry maintenance remains a concern

The Bottom Line

The MCP community should treat the 'half are dead' narrative as a cautionary tale about how quickly misinformation can calcify into accepted wisdom—and maybe run a few tests before amplifying claims that could steer teams away from solid infrastructure choices.