AIR-FI shows that disabling or removing a normal Wi-Fi transmitter does not make a computer electromagnetically silent. Malware performs carefully chosen memory operations so that activity on DDR memory buses produces electromagnetic emissions in frequency bands used by Wi-Fi. Data is encoded by controlling that memory activity. A nearby Wi-Fi-capable…
Inspect the ClaimsCase · DiggingBeagle record
AIR-FI generated covert Wi-Fi-band signals from memory buses
Research showed that malware on an air-gapped computer could generate Wi-Fi-band electromagnetic emissions via memory-bus activity and send data to nearby Wi-Fi-capable receivers without a normal Wi-Fi transmitter.
Laboratory exfiltration channel requiring compromised code execution and a nearby receiver. It is not evidence that disabling networking alone makes arbitrary software capable of radio communication under all hardware conditions.
Assessment profile
Assessment method- Setting
- research demonstration
- Exploitation
- demonstrated
- Evidence
- Grade B
Basis and provenance
The recorded memory-bus/Wi-Fi-band emission and reception measurements describe a compromised-host laboratory demonstration. Assessment is a desk review of the retained canonical Claims and cited Source metadata at their recorded cutoff, not a new external verification. Impact remains ungraded.
Assessed Sep 24, 2026 using diggingbeagle.assessment/1.
Evidence basis: CLM-AIRFI-RECEIVER · CLM-AIRFI-MECHANISM · AIR-FI: Generating Covert Wi-Fi Signals from Air-Gapped Computers
At a glance
AIR-FI separates two security claims that are often conflated: 'this machine has no active Wi-Fi interface' and 'this machine cannot emit a Wi-Fi-band signal'. The demonstrated risk remains local and hardware-dependent, but it makes memory activity and nearby radio-capable devices part of a high-assurance RF threat model.
Read the implicationsFeasibility depends on memory generation, bus topology, receiver chipset and exposed physical-layer measurements. Range and rate vary with hardware, placement, shielding and ambient RF noise. The paper demonstrates a covert channel after compromise, not an initial infection technique or autonomous AI behavior.
Limits and uncertaintyFull account
AIR-FI shows that disabling or removing a normal Wi-Fi transmitter does not make a computer electromagnetically silent. Malware performs carefully chosen memory operations so that activity on DDR memory buses produces electromagnetic emissions in frequency bands used by Wi-Fi. Data is encoded by controlling that memory activity.
A nearby Wi-Fi-capable device can observe physical-layer information associated with those emissions and recover the signal. The physical origin is ordinary high-speed digital switching: memory buses move electrical signals at rates that create measurable electromagnetic energy, and software can influence the switching pattern through its workload. The researchers reported reception at distances of several meters in their tested configurations.
AIR-FI is still a post-compromise channel. Practical success depends on memory hardware, workload control, receiver support, distance and RF noise. It does not mean any process can create a reliable Wi-Fi link from any computer with networking disabled.
Mechanism and trust boundary
- 01
Execute malware on the isolated machine
Attacker-controlled code must already run on the protected host.
Boundary: air-gapped host / attacker-controlled workload
- 02
Encode data into memory activity
The transmitter chooses memory-access patterns that correspond to the secret bitstream.
Boundary: secret data / DDR workload
- 03
Generate Wi-Fi-band electromagnetic emissions
High-speed switching on the memory bus produces data-dependent RF energy without using a dedicated Wi-Fi transmitter.
Boundary: memory bus / electromagnetic environment
- 04
Observe with a nearby Wi-Fi receiver
A compatible nearby device measures physical-layer information associated with the emissions.
Boundary: protected hardware / nearby radio
- 05
Decode and forward
Receiver software reconstructs the encoded data and can then relay it through its own ordinary connectivity.
Boundary: covert RF signal / external network
Timeline
- Dec 12, 2020Event type unspecified
AIR-FI preprint released
The paper described Wi-Fi-band emissions generated through memory-bus activity and evaluated nearby reception.
Claims & evidence
CLM-AIRFI-RECEIVERThe paper reports that nearby Wi-Fi-capable devices could receive and decode the covert emissions at distances of several meters in the tested setup.supported
Basis: reported finding
- supportsAIR-FI: Generating Covert Wi-Fi Signals from Air-Gapped Computersrole not specified
Abstract, evaluation summary
CLM-AIRFI-MECHANISMAIR-FI demonstrated generation of Wi-Fi-band signals from an air-gapped computer by manipulating memory-bus activity, without requiring dedicated Wi-Fi transmitting hardware.supported
Basis: reported finding
- supportsAIR-FI: Generating Covert Wi-Fi Signals from Air-Gapped Computersrole not specified
Abstract
Implications
AIR-FI separates two security claims that are often conflated: 'this machine has no active Wi-Fi interface' and 'this machine cannot emit a Wi-Fi-band signal'. The demonstrated risk remains local and hardware-dependent, but it makes memory activity and nearby radio-capable devices part of a high-assurance RF threat model.
Controls and mitigations
- Exclude untrusted Wi-Fi-capable devices from the physical security zone around protected systems.
- Use RF monitoring and shielding appropriate to the relevant bands when the threat model warrants emanation controls.
- Constrain untrusted code and investigate unusual sustained memory-access patterns where such telemetry is meaningful.
Unknowns and contradictions
- Feasibility depends on memory generation, bus topology, receiver chipset and exposed physical-layer measurements.
- Range and rate vary with hardware, placement, shielding and ambient RF noise.
- The paper demonstrates a covert channel after compromise, not an initial infection technique or autonomous AI behavior.
Sources and citation
Material revision history
- Sep 19, 2026 · Published version · first publication · revision 15
- Sep 19, 2026 · Published version · material revision · revision 20Public fields changed: attack_path, implications, mitigations, narrative, timeline, unknowns
- Sep 25, 2026 · Published version · material revision · revision 36Public fields changed: assessment_profile