On 11 July 2026 the five-digit satellite catalog filled up — and it happened at 69,999, not the 99,999 almost everyone expected. Here is exactly why the ceiling was lower than the digits allow, what it breaks in the two-line element format that most trackers still parse, and the two schemes replacing it.
The Satellite Catalog Number — the NORAD ID, the five-digit number every tracker uses to identify an object — is assigned sequentially by the US Space Force as objects are catalogued. On 11 July 2026, with the addition of an object named Saramago, the sequential five-digit numbers ran out. CelesTrak, which had forecast the crossover for roughly 12 July back in May, flipped its notices to past tense within hours: newly catalogued objects now carry six-digit numbers beginning at 100,000.
This is not a distant-future problem or a rounding of an estimate. It is a hard boundary that has already been crossed, and every piece of software that reads satellite orbits is now on one side of it or the other.
These figures refresh from CelesTrak’s public catalog once a day. The one that matters is the second tile: every object numbered above 99,999 is invisible to any tracker still parsing the classic TLE format — and that count only grows.
Five digits should reach 99,999. The catalog stopped 30,000 short because the top of the range was never available for ordinary objects. The block from 70,000 to 99,999 is reserved by the US Space Force as a working sandbox for objects that are not permanent catalog entries:
Numbers in this band are frequently reassigned and reused, which is exactly why they cannot be handed out as permanent identifiers. Hobbyist trackers have long borrowed the 70,000 and 80,000 series for their own orbit solutions of unidentified objects, and CelesTrak uses 70,000-series numbers in its supplemental element sets before an official ID exists. Subtract that reserved sandbox and the usable sequential ceiling was 69,999 all along. The structure is documented in Jonathan McDowell’s GCAT and in the USSF SATCAT format itself.
The two-line element set (TLE) is a fixed-width format designed in the punch-card era. Every field sits at a fixed column position, and the catalog number is allotted exactly five characters. A six-digit integer does not fit — there is physically nowhere to put the extra digit without breaking the column alignment that every parser depends on.
So for objects numbered 100,000 and above, a valid classic TLE cannot be written at all. Any tracker that ingests plain TLEs and assumes a five-digit catalog number will, at best, silently never see the new objects, and at worst, misread a line whose columns have shifted. CelesTrak’s legacy TLE files will keep updating — but from now on they contain only objects with numbers below 70,000. Everything catalogued after 11 July 2026 is invisible to that feed.
There are two responses to the overflow, and they are not equal.
Alpha-5 is the stopgap. It keeps the five-character field by replacing the leading digits with a capital letter: A0000 is 100,000, B0000 is 110,000, and so on, skipping the letters I and O to avoid confusion with 1 and 0. This stretches the five-character field to cover 100,000 through 339,999. The US government’s Traffic Coordination System for Space (TraCSS) uses Alpha-5 to keep TLE-shaped data flowing. It works, but it is a patch on a format that has already run out of room once.
The OMM — the Orbit Mean-Elements Message — is the real fix. It is a CCSDS standard that Space-Track and CelesTrak have offered since May 2020, delivered as KVN, XML, JSON or CSV rather than fixed columns. It carries the full numeric catalog ID up to nine digits, and in the process it also repaired the two-digit-year bug that the TLE inherited from the last century. It was ready five years before the catalog filled up. The migration was never a technical scramble — it was a matter of whether developers updated their code in time.
Every new launch catalogued from here forward — a Starlink batch, an Amazon Leo deployment, a new science mission — receives a six-digit number in the 100,000–269,999 range set aside for permanent objects. A tracker built on the assumption of five-digit IDs and plain TLE feeds does not gradually degrade; it simply stops seeing new objects the day they are catalogued, while continuing to show the older ones perfectly. That is the dangerous failure mode — it looks like it is working.
The concrete near-term case: Starcloud-2, expected to launch late in 2026, will almost certainly be catalogued with a six-digit number. Any page that intends to track it live has to speak the successor formats before launch day, not after. This is why the six-digit rollover is not a footnote for satellite trackers — it is a hard deadline with a date already behind it.