Blackfort Technology
Threat BriefingJuly 22, 2026·Christian Gebhardt

Open in Orbit: Why Half of the Observed Satellite Downlinks Carry No Encryption

A research team read geostationary satellite traffic with about 800 US dollars of receiving gear — 50% of the observed links ran in the clear. We reproduce the core in the lab, fully software-defined: why this is not an attack in the usual sense but simply missing encryption — and how the very same reception sees only ciphertext once the payload is encrypted.

Follow Blackfort on LinkedIn

Security incidents, technical analyses and hands-on insights — straight to your LinkedIn feed.

Follow now →

A research team from the University of California San Diego and the University of Maryland used about 800 US dollars of consumer receiving gear — a consumer satellite dish with a motor and a TV tuner card — on the roof of a university building in San Diego to read the IP traffic of geostationary satellites. The result, presented at ACM CCS ’25 in Taipei and awarded a Distinguished Paper Award: 50% of the observed GEO links carried their traffic in the clear. According to the study, this included phone calls and text messages from cellular backhaul, industrial control systems of utilities, military asset tracking, inventory data of global retailers, and in-flight Wi-Fi.

Short analysis: why half of the observed GEO downlinks run in the clear and why encryption is the only layer that holds in the footprint model.

What is troubling here is neither the equipment nor a clever attack. It is the absence of encryption. Anyone standing inside a transponder's footprint who points a dish at it receives the signal — that is all it takes.

Key figures

Cleartext share: 50% of the observed GEO links unencrypted

Scan scope: 39 satellites, 411 transponders, 25 longitudes

Effort: ~800 US dollars of passive receiving gear

Source: University of California San Diego & University of Maryland

Publication: ACM CCS ’25 (Distinguished Paper Award)

DOI: 10.1145/3719027.3765198

We reproduced the core of this finding in an isolated lab, fully software-defined in GNU Radio and with purely synthetic, invented frames. There was no real radio reception and no third-party data. We rebuilt the chain ourselves: our own unencrypted downlink, a software receiver that demodulates it — and, next to it, the same chain with an encrypted payload. This article explains why such links lie open, what threat model that implies, and how the safeguard works; backed by the real material from the run.

The problem: a strong signal that everyone in the footprint receives

A geostationary satellite sits fixed above a point on the equator and beams its downlink signals back to Earth through transponders. A single transponder footprint covers a very large part of the Earth's surface — up to about a third, according to the study. Anyone inside that area who runs a properly aimed dish gets the same signal as the intended ground receiver. Reception is passive: nothing is transmitted, nothing is disturbed, nothing in the connection is touched. From the outside, such eavesdropping cannot be detected.

Confidentiality therefore rests entirely on the encryption of the transmitted data. And that is exactly where the gap is. In satellite TV, encryption at the transmission layer has been standard for decades. On the IP backhaul links that connect cellular networks, corporate sites, utilities and others, it was often missing according to the study — and missing on both levels, the link and the network layer. The IP traffic then ran raw across the satellite.

So there is no broken cryptography and no exploit in the usual sense. There is an assumption that does not hold: that a signal you cannot see, and whose reception requires expertise, is therefore also confidential. The term for this is “security by obscurity.” It holds exactly until someone with an 800-dollar dish takes a look.

Scenarios: who can listen in — and what is at stake

The threat model is unusually simple. The listener needs no access to a network, no stolen credentials, no position “in the middle” of the connection. They need a dish in the footprint and patience. They are purely passive and stay invisible, because they change nothing about the connection. What this model cannot defeat: an encrypted payload — because then they receive the same signal but can do nothing with it.

What the team observed in unencrypted traffic over three years cuts, according to the study, straight across critical sectors: cellular backhaul with cleartext call and text content from several providers, industrial control of utility infrastructure (job scheduling, ICS/SCADA), military asset tracking, inventory management of global retailers, and passengers' in-flight Wi-Fi.

Individual affected operators were named and encrypted their links after the researchers' disclosure. We deliberately do not name them here. What matters is the pattern: sensitive internal links running openly over a satellite whose signal a large part of a continent can receive.

This is what the receiving side looks like in the lab. The software receiver picks up the unencrypted downlink; in the waterfall spectrogram the carrier band of the signal is clearly visible:

Waterfall spectrogram of the unencrypted downlink in the lab, from real IQ data of the software receiver: the carrier band of the signal is clearly visible
Waterfall spectrogram of the unencrypted downlink in the lab (from real IQ data).

After demodulation the symbols settle into four clean clusters — an intact QPSK constellation, so the signal is fully recovered:

QPSK constellation at the receiver in the lab: four clean clusters, the unencrypted signal is fully demodulable
QPSK constellation at the receiver: four clean clusters, the signal is demodulable.

And because the payload is unencrypted, the decoded frames then sit there in the clear — in the lab deliberately with invented content marked as synthetic. The cleartext is the leak:

Decoded frames in the clear: the unencrypted payload lies open — in the lab with purely synthetic content
Decoded frames in the clear: the payload lies open (synthetic content).

That final step would happen on a real, open downlink as well — only there it would be actual calls, telemetry or business data.

Solution: encryption at every layer, mandatory

The remedy is not a new technology, it is a decision: data in transit is encrypted, regardless of the medium. For radio and satellite links that means encryption at the transmission and the network layer, so the payload stays protected even if a single layer fails. This is defense-in-depth in practice: several independent layers, so that one fault does not lay everything bare.

Why that is enough is shown by the same lab setup with a single change — the payload is encrypted before transmission (AES-256), the framing stays the same. Afterwards the radio path is indistinguishable: waterfall and constellation look just as before. The receiver demodulates just as cleanly:

QPSK constellation with an encrypted payload: identical radio path, four clean clusters just as in the open case
QPSK constellation with an encrypted payload: identical to the open case.

The only difference is in the last step. The same receiver that just produced cleartext now sees only ciphertext — bytes that are unreadable without the key:

Decoded frames with an encrypted payload: only ciphertext, bytes that are unreadable without the key
Decoded frames with an encrypted payload: only ciphertext.

That is the whole point. The signal is just as receivable as before. The listener still gets it — they just can no longer read it. Encryption is the only layer that holds at all in the footprint model.

For organisations with a lived security management this is more than a recommendation. An effective ISMS(for example under ISO/IEC 27001) calls for a documented cryptography policy and the protection of transmission paths — the case shows how a blind spot, trusting in the “invisibility” of the medium, undermines a central control. Anyone connecting sensitive sites over radio or satellite links should anchor the encryption of those paths explicitly in policy and architecture.

What it means for operations

This case is a lesson against “security by obscurity”: a medium that seems hard to observe is no substitute for a control. Anyone carrying sensitive internal links over radio or satellite must establish confidentiality actively through encryption of the transmission paths — at the transmission and the network layer. That belongs in a documented cryptography policy and in the architecture, not in a silent assumption.

Note

This article explains a technical matter and is not legal advice. The material shown comes from a fully software-defined reproduction with purely synthetic data in our isolated lab environment; no real third-party traffic was received — such eavesdropping would be a criminal offence in Germany (secrecy of telecommunications, § 201 of the Criminal Code, § 148 of the Telecommunications Act). The study findings are taken from the research team's public release (University of California San Diego / University of Maryland, ACM CCS ’25, DOI 10.1145/3719027.3765198).

Kontakt aufnehmen

IT security for your organisation

Blackfort Technology supports organisations with NIS2 compliance, OT security and the protection of critical infrastructure – from analysis to implementation.