Why Our CBRS SAS Uptime Stands at 99.986%
... and Why That Number Matters More Than You Might Think
At Key Bridge, we publish our SAS uptime without any asterisks, without any convenient rolling windows, and without any excuses. As of today, March 17, 2026, our lifetime uptime since commercial service began on March 9, 2021, is exactly 99.986%.
We calculate it the old-fashioned way: total hours of service offered divided by total hours of service actually delivered, from the very first day we turned on the system. No annual reset, no “last 12 months only,” no pretending the past doesn’t exist. That’s 44,016 hours of continuous commercial operation. The math is straightforward:
The entire difference between our number and 100% comes down to a single event in June 2021 — roughly three months after launch. Chicago, our only data center at the time, experienced a site-wide power outage. The facility was completely dark for exactly four hours. When power returned, it took an additional 2.16 hours for every system, database, and service to fully synchronize and resume normal operations. That one incident added 6.162 hours of counted downtime. Nothing else in five years has ever caused measurable service impact.
Today we run four fully independent data centers — two for the Key Bridge SAS and two for our Nokia-branded white-label service. They do not back each other up; each runs in its own isolated environment with full geo-redundancy inside its own pair. A power event in any single facility now would be invisible to our customers. But the early outage happened when we had only Chicago online, so it counted fully and permanently.
Some people might suggest we “move on” from an event that affected only a handful of early customers. I refuse. In fact, I insist we keep it on the record forever.
Here’s the philosophy I’ve carried with me since the beginning: early grades count just as much as later accomplishments. A single B on the first quiz of freshman year still appears on your final transcript. It doesn’t disappear because you earned straight A’s afterward. The same principle applies here. That June 2021 outage — when we were small, when only a few pioneers had trusted us — remains part of our permanent record because it should. Every minute of downtime, no matter how small the customer base at the time, is a serious issue. It is a permanent reminder that vigilance never gets a day off.
I keep that number visible internally for exactly this reason. It is not a mark of shame; it is a mark of seriousness. It tells every engineer, every operations person, and every new hire that we do not grade ourselves on a curve that conveniently forgets the past. It tells our customers that when we promise carrier-grade reliability, we are not measuring it the easy way. We are measuring it the honest way — the same way we will still be measuring it on the day we reach five nines, six nines, or beyond.
So when you see our uptime reported as 99.986%, know that the missing 0.014% is not hidden, not forgotten, and not excused. It is remembered on purpose. That memory is what drives the diligence that has kept the system flawless ever since.
We are not perfect. But we are honest about where we started, and we are relentless about where we are going. That combination, more than any raw percentage, is what I believe our customers deserve.
Pages