70 Years of PBS History Trapped in a Denver Data Center
Quick question: who actually owns your data when it lives in the cloud? Not a philosophical question — a legal one. Ask Nine PBS in St. Louis, which just found out the hard way that possession and access are not the same thing.
Here's what happened. Nine PBS stored more than 50TB of archival material — roughly 70 years of TV shows, videos, and program masters — with a cloud storage vendor called Open Source Storage (OSS). OSS, in turn, hosted that data on servers at Iron Mountain's Denver data center. Classic cloud arrangement: logical layer here, physical layer there, everybody's happy.
Until OSS just... disappeared.
In March 2026, Nine PBS tried to renew access. OSS stopped responding. No warning, no graceful handoff, no "here's your export window before we shut down." Just silence. The station's irreplaceable archive — decades of local history, produced content that can't be recreated — sat physically intact on Iron Mountain's servers, completely inaccessible because the contract was with a company that had gone dark.
So Nine PBS sued. On July 28, 2026, they filed against Iron Mountain in Denver District Court, trying to force recovery of files they never lost custody of, technically, but absolutely lost access to. A judge has since ordered Iron Mountain to cooperate with a third-party specialist chosen by Nine PBS — while making sure nobody accidentally tramples the data of other former OSS customers stuck in the same mess.
Think about that for a second. Iron Mountain isn't even the villain here, not really. They're just holding servers that belong to a company they had no direct relationship with the end customer about. It's a legal gray zone created entirely by the layered structure of modern cloud storage.
<> The broader criticism is of "cloud illusion": organizations may think they have direct, durable access, but in practice they may be dependent on a chain of vendors and legal relationships that can fail at any point./>
That line should be printed and taped above every CTO's desk.
Here's the uncomfortable part. This isn't a hacking story. Nobody breached anything. There's no ransomware, no leaked credentials, no dramatic exploit. The data just became a hostage of corporate vanishing. And that's arguably scarier, because there's no patch for a vendor ghosting you.
What engineering teams should actually take from this
1. Bulk export capability isn't optional. If your storage provider doesn't let you pull everything, on demand, without begging — that's a red flag, not a feature gap.
2. Contracts need teeth on offboarding. Retrieval rights, timelines, escrow provisions for what happens if the vendor implodes. Most SaaS contracts treat this as an afterthought.
3. Physical custody ≠ your custody. If you don't control the API, the credentials, and the billing relationship directly, you're one bankruptcy away from an OSS situation.
4. Run recovery drills. Not backup tests — actual "what if this vendor vanished tomorrow" drills. Most teams have never done this.
Hot take
Everyone's going to read this story and blame OSS for going defunct, or Iron Mountain for the gray-zone access dispute. Wrong target. The real failure is Nine PBS's own vendor diligence from 2019 onward — using a single provider (and its predecessor) for seven years without an independent backup or exit plan for archival material that is, by definition, irreplaceable. Public institutions treat cloud storage like a utility. It's not. It's a relationship, and relationships end badly more often than IT departments want to admit. If your only copy of 70 years of history lives with one company's goodwill, you don't have a backup strategy — you have a prayer.
