Seventy five brief interference events over seven years, clustered on weekdays and inside business hours, point to a single spacecraft in a Molniya orbit and a threat model that extends well beyond ground stations.
For seven years, GPS receivers across Europe and North America have been losing signal at the same moment on the same days of the week. The events are short, often under ten seconds, and the signal drops on a narrow slice of radio spectrum centered on 1,577.5 MHz, the frequency that carries the civilian GPS and Galileo signals. What makes the pattern remarkable is its regularity. According to a preprint from researchers at the University of Texas at Austin, 75 of these interference events between January 2019 and April 2026 fell on Tuesdays, Wednesdays, or Thursdays, and clustered inside business hours. That is not the signature of an accident or a one-off test. It is the signature of scheduled work.
The team, led by Todd Humphreys of UT Austin's RADIONAV Lab, used a public network of GNSS reference stations to detect sudden power drops on the L1 band, then applied time-difference-of-arrival analysis to localize the source. The result sits above 1,200 kilometers in altitude, which rules out ground-based jammers of the kind that have proliferated around Ukraine, the Baltic, and Arctic shipping lanes since 2022. The candidate set at that altitude is small. Cross-referencing the events against published satellite catalogs, the researchers identify a single match: Kosmos 2546, part of Russia's EKS early-warning constellation flying a Molniya orbit, which traces a long, looping path over the northern hemisphere.
A report on the paper by Passant Rabie at Gizmodo notes that the interference is not aimed at the GPS L1 center frequency itself but slightly offset from it. To Humphreys and his co-authors, that detail reads as a deliberate choice: a test signal calibrated to disrupt the band a real attack would target, but tuned off-center enough to avoid permanent attribution. "I can no longer say this is accidental with confidence," Humphreys told Gizmodo. "I'm leaning toward this being a periodic test of a capability that would be very damaging if it was deployed in anger."
The pattern matters more than any single event. Most terrestrial GPS jamming is regional and opportunistic. A single satellite in the right orbit can sweep denial across a continent in seconds, and the seven-year record suggests Russia has been quietly running that playbook on a weekly cadence. A second, lower-frequency burst overlapping China's BeiDou band was also detected, ruling out the simpler explanation of a single faulty transmitter. The capability the team is documenting is general, not country-specific.
There are real limits to what the paper proves. It is a preprint, submitted to the Institute of Navigation journal NAVIGATION but not yet peer-reviewed. The attribution chain runs through statistical and positional matching, not a Russian operator's admission. Moscow has not commented on the record, and "scheduled" does not yet mean "authorized at the Kremlin level." A scheduled test, a unit-level exercise, and a deliberate attack look the same in a power-spectrum plot, and only two of those three are acts of war.
The BeiDou-band observation is also a single data point, not a confirmed second capability, and the business-hour clustering is consistent with a European or U.S. workweek rather than confirmed against a specific time zone. Independent confirmation from a second research group, or any statement from Roscosmos or the Russian Ministry of Defence, would change the weight of the finding materially.
What readers should take away is narrower than the headline suggests. Russian electronic warfare in orbit is not new, but a multi-year, scheduled, narrow-band jamming program traceable to a single named satellite is. The threat model that has dominated since 2022 is regional ground stations near a conflict zone. The new finding points to a different kind of asset: a single spacecraft, in a long loiter, denying GPS to anyone beneath it, on a Tuesday afternoon, for as long as someone keeps turning the transmitter on.
Watch for the peer-reviewed version of the paper, for any official Russian response, and for follow-up analyses that confirm or refute the Kosmos 2546 attribution using independent station data. Each of those would tighten, or weaken, the case that the most interesting GPS jammer in the sky right now is not a ground station at all.