Hardware Analysis Series – Part 5: Electromagnetic Espionage – The Silent Attack

By Kike Gandia · Hardware & OT · 28 min read

We have built digital walls with unbreakable passwords, robust firewalls and encryption that borders on uncrackable. And yet a troubling paradox persists: even the most secure systems whisper their secrets into the void without anyone laying a finger on them. Our computers and devices give off phantom signals – faint electromagnetic radiation, acoustic vibrations or traces of heat – that can reveal confidential information to a patient spy with the right tools. How did we end up here, in a world where the simple act of running makes a machine betray its own confidentiality? The answer lies in decades of unbridled technological development, in an obsession with performance over containment, and in forgetting a fundamental physical truth: every active circuit is also a tiny transmitter. We have built a silicon civilization so electromagnetically noisy that today our secrets can leak out through the ether, with an adversary picking up, from a distance, what is happening inside our chips without injecting a single bit of malicious code.

This opening reflection forces us to question our assumptions. For years we focused on protecting the “official” digital communication channels – networks, ports, software – while inadvertently leaving physics' back doors wide open. As a result, an attacker does not need to physically touch a system to steal information; listening to the invisible emissions that system produces while it operates is enough. In an almost literal sense, we have allowed machines to become unwitting informants on their own internal processes. This article explores that silent, advanced form of espionage – electromagnetic attacks based on physical emissions – looking at how they work technically, their history through the Cold War and recent real-world cases, and reflecting critically on how the tech industry got to this point and what we should do about it.

What are electromagnetic attacks and why are they so dangerous?

Electromagnetic side-channel attacks (EM side-channel attacks) exploit the fact that every electronic device emits something while it runs. That “emission” can be electromagnetic radiation (radio-frequency waves), acoustic noise (sound or vibrations) or even dissipated heat. Unlike a conventional cyberattack, in a side-channel attack the adversary does not try to compromise the software or internal logic directly, but instead measures physical side effects of the hardware's operation. It's like eavesdropping on a conversation by listening to the echo bouncing off the walls instead of tapping the phone line.

Why is this dangerous? Because it lets the attacker extract sensitive information without altering the target system in any way and without leaving a digital trace. The emissions are involuntary and ubiquitous: every voltage change in a chip, every clock pulse, every bit sent across a bus generates a minuscule electromagnetic field, a vibration or a thermal pulse. With specialized (or even improvised) equipment, it is possible to capture those signals and reconstruct confidential data from them. In other words, hardware leaks its own internal state simply by operating, and a shrewd spy can “listen in” on those fleeting electronic murmurs.

Picture a scenario: a laptop encrypting a classified document. Even if it is disconnected from every network and physically protected, the rapid current changes in its CPU produce subtle radio waves. An attacker nearby, with a hidden antenna, could pick up that “clandestine radio” broadcast by the processor and, through advanced analysis, recover parts of the cryptographic key in use. Similarly, keystrokes generate unique electrical transients; by recording a keyboard's electromagnetic emanations or even the sound of the keys, an attacker can infer the letters typed. Even a computer screen, refreshing its content pixel by pixel, emits a radio-frequency pattern that a receiver can turn back into the image being displayed. These leaks require neither physical contact nor the installation of malware, so they go completely unnoticed by traditional defenses. They are silent attacks, literally out-of-band, that get around even isolated (air-gapped) systems because they exploit fundamental physical properties.

Part of the danger also lies in how hard they are to mitigate: it is difficult to “switch off” these emissions without switching off the device itself. Adding shielding or filters can raise costs and reduce performance, measures that are rarely taken outside the military sphere. That is why, in the civilian world, most of our devices are electromagnetic open books, ready to be read by anyone who knows how to tune in to their hidden signals. In short, EM attacks weaponize the hardware's own internal operation, achieving what once seemed like magic: spying on a secure system without even touching it.

The history of TEMPEST and espionage through physical emissions (waves, heat, sound)

Although they sound like cutting-edge technology, attacks based on physical emissions have deep historical roots. The first warning signs were documented as early as World War II. Bell Labs engineers discovered in 1943 that the cipher machines of the day were radiating their secrets: looking at an oscilloscope, they noticed disturbances synchronized with the encryption operations – and those signals made it possible to read the plaintext message from a distance! This initial finding was met with skepticism – it was hard to believe a teletype could act as a small, unwanted radio station – but later experiments confirmed it dramatically. In a famous test, Bell Labs managed to capture around 75% of the plaintext being encrypted by a machine simply by measuring its emissions from about 25 meters away, from an adjacent building. The military's reaction was telling: in the middle of a war, they preferred to proceed cautiously rather than redesign all their machines, so the first countermeasure was pragmatic – a 200-foot (roughly 60-meter) security zone around any cryptographic equipment. If you can't silence the machine, at least keep the snoops away.

That is how the concept of compromising emanations was born, and with it the code name TEMPEST. Originally just a code word, it would later be associated with the acronym Telecommunications Electronics Material Protected from Emanating Spurious Transmissions. During the Cold War, TEMPEST became a top-secret program dedicated to studying and mitigating these inadvertent emissions. Why so much secrecy? Because while everyone was racing to shield their own equipment, they also glimpsed an enormous espionage opportunity: if your enemy doesn't know that their devices emit intelligible information, you can listen to their secrets without them ever suspecting.

In fact, historical episodes abound. In 1954, the Soviet Union published surprisingly strict radio interference suppression standards for its teletypes and communications equipment – far tougher than those for ordinary industrial machinery – suggesting it had taken the lesson of compromising emanations to heart. In 1962, an American officer in Japan discovered a mysterious dipole antenna hidden under the eaves of a hospital across from a U.S. military crypto office; it pointed straight at the cipher room, probably soaking up its radiation on the sly. The antenna was hastily removed after it was discovered, but the message was clear: someone was trying to tune in to U.S. military machines from a distance.

A famous case came in 1964, when more than 40 hidden microphones were found in the U.S. Embassy in Moscow. The public scandal focused on the diplomatic conversations being spied on, but intelligence experts were asking a more disturbing question: what were those microphones doing right next to the embassy's cipher machines? Alongside the microphones, investigators found strange gadgets never publicly disclosed, such as a metal grid hidden in the ceiling of the communications room with extremely thin wires connected to it. Everything pointed to a sophisticated Soviet attempt to capture the cipher machines' electromagnetic emanations or to inject controlled interference. It was the silent war of electronics: while audio “bugs” were planted in plain sight, nets to fish for invisible signals were being woven in the shadows.

In parallel, acoustic and thermal attacks also emerged. The laser-on-the-window technique, for example, was perfected during the Cold War: a laser aimed at the window of a meeting room picks up the tiny vibrations in the glass caused by the voices inside, making it possible to reconstruct conversations from a distance. And while TEMPEST focused on electromagnetic (radio-frequency) emissions, over time it became clear that sound and heat could also leak information. A curious case from the 1980s was the eavesdropping on dot-matrix printers: agencies discovered they could reconstruct text by analyzing the impact sound of the printer's pins, since each character had a unique acoustic “signature”. Likewise, it was shown more recently that after someone types a PIN or password on a physical keypad, an attacker could use an infrared camera to read the residual heat on the keys and work out the sequence entered. The first keys pressed will be cooler than the last, revealing even the order of the digits. These are modern evolutions of an old principle: every action leaves a physical trace.

Around 1985, Dutch engineer Wim van Eck showed the civilian world what spies already knew: with relatively simple equipment he was able to capture a CRT monitor's video emissions from several meters away and reconstruct on another screen what that monitor was displaying. The technique is known as Van Eck phreaking, and it was so impressive at the time that many people considered it almost science fiction. Seeing a computer's display reproduced on a TV purely from waves captured out of the air set off alarms outside government circles. From then on, the technical community began to study side-channel attacks seriously. In the 1990s and 2000s, academic researchers such as Paul Kocher revealed, to general astonishment, that smart cards and cryptographic chips could betray their secret keys through measurements of their power consumption or EM emissions. Terms such as Differential Power Analysis (DPA) and Electromagnetic Analysis were coined, and the industry had to respond by building in countermeasures (for example, designing chips that emit less signal or add random “noise” to mask patterns).

The history of TEMPEST and physical-emission attacks is, at heart, a stealthy arms race: for every new attack discovered, new protections are tried (shielding, filters, masking mechanisms). But as technology advances, so do the attacker's tools. An NSA report concluded as early as 1972 that attack methods tend to stay ahead of countermeasures – and that is still true today. From the crude interference picked up in the 1940s to today's sophisticated algorithmic attacks, the discipline has evolved from a classified national-security matter into knowledge accessible to researchers (and potentially criminals) all over the world.

World War II — The first compromising emanations
It is documented that a piece of equipment can give away what it processes through what it emits. The concept of compromising emanations is born and, with it, the code name TEMPEST.
1954 — The Soviet standards
The USSR publishes radio interference suppression standards that are surprisingly strict for the time.
1964 — The Moscow embassy
More than 40 hidden microphones are found in the U.S. Embassy. The public scandal stops at the microphones; the underlying problem was broader.
Cold War — The laser on the window
Acoustic and thermal attacks are refined as well: the glass vibrates with the conversation and the laser reads that vibration.
1985 — Wim van Eck takes it public
The Dutch engineer shows the civilian world, with relatively simple equipment, that a monitor's emissions can be captured and its screen reconstructed.

A stealthy arms race: for every attack discovered, a countermeasure; for every countermeasure, a new attack.

Real-world cases: modern examples of invisible information leaks

To understand the current reality of these attacks, it is worth reviewing some demonstrated cases that highlight how easy it can be to steal information through physical emissions:

2015 · $20 Stealing RSA keys with a $20 radio: In 2015, a group of cryptographers from Tel Aviv University showed how they could extract a 4096-bit RSA private key in seconds by measuring radio waves near a laptop. Their setup consisted of a cheap radio receiver (a Funcube Dongle Pro+ connected to a small computer) and a simple antenna. Positioned just 50 cm from the target, they picked up emissions in the 1.6 MHz band given off during GnuPG decryption operations. With a bit of digital signal processing, they managed to reconstruct the secret key used by the encryption software. More astonishing still, they claimed that an ordinary AM radio might even be enough, with a smartphone's microphone used to record the signal. The subtlety of the attack lay in sending the victim computer specially crafted ciphertexts; when decrypting them, the RSA implementation produced distinguishable patterns in the surrounding electromagnetic field. Those patterns varied with the bits of the key, making it possible to infer the full key after careful analysis. During the test, the team hid the entire spying device inside a piece of pita bread, underscoring just how covert this method can be. The case showed that even robust algorithms running on up-to-date systems are vulnerable if someone with a small budget can get close enough with a radio receiver.
2014 · acoustic The sound of encryption: That work built on another surprising feat from 2014, when some of the same researchers (Genkin, Shamir, Tromer, et al.) managed to recover RSA keys by listening to the sound a computer makes during decryption. Every processor under heavy load generates tiny vibrations and high-frequency noise (phenomena such as the “coil whine” of voltage regulators). With a simple microphone placed next to the machine – even the microphone of a nearby smartphone – they were able to record that subtle hum and, after analyzing it, extract a complete encryption key as well. The fact that internal mathematical operations produce detectable acoustic differences is incredible, but real. This acoustic attack recalls another vector: keyboard eavesdropping. It has been shown that, from a recording of someone typing, machine learning algorithms can identify which keys were pressed from the unique sound variations of each key (due to mechanical differences). In other words, you don't need a camera to spy on your password; a directional microphone at a distance can “hear” your password on the keyboard with high accuracy.
Thermanator A revealing thermal camera: A more mundane but illustrative experiment is the thermal-camera attack known as Thermanator. Researchers showed that after you enter a PIN or password on an ATM, an electronic lock or even a phone screen, an infrared camera can “read” the heat traces your fingers left behind. The keys or digits touched light up in the thermal spectrum, and by analyzing their relative intensity (the first presses cooling down while the last ones are still warm) it is possible to deduce the order in which they were pressed. This attack is opportunistic – it requires a view of the keypad right after it has been used – but it is a physical side channel nonetheless. It shows that even heat, that unavoidable by-product of electronics and of our own bodies, can be turned against security if it isn't mitigated.
Modern Van Eck Modern phreaking on monitors and cables: Today's monitors (LCD/LED) can also be spied on, although their emissions are weaker than those of old CRTs. Techniques have been presented for capturing leaks from DVI/HDMI cables or internal wiring, reproducing low-resolution images of what is on screen from several meters away. Likewise, special radio receivers can detect radiation from a machine's DDR memory or USB bus, extracting patterns such as USB keystrokes or CPU activity. In academic settings, researchers have recreated attacks in which a drone flies over an office equipped with RF sensors and laser microphones, combining multiple channels (radio, sound, vibration) to spy on supposedly secure equipment.

Some are lab proofs of concept, others tangible threats. All of them confirm that the EM side channel is not theoretical.

These real-world cases – some lab proofs of concept, others tangible threats – confirm that EM side-channel threats are neither theoretical nor the preserve of superpowers. They are feasible today, on small budgets and with publicly available knowledge. One military expert commented on the Tel Aviv attack: “These attacks have been known for years; what matters is distance. If they can do it from 10 meters through a wall, that would be really impressive; from 20 cm, not so much.” Indeed, distance and ambient noise are usually the limiting factors. But the trend is clear: it is getting easier and easier to get close enough (physically or with hidden devices) and leak useful signals thanks to powerful signal-processing algorithms. Electromagnetic espionage has left the spy novels and entered the real world of information security.

Hardware talks: detectable emissions in chips, PLCs and smart cards

Personal computers are not the only ones at risk. Any complex electronic device can give away information through its physical emissions, from tiny smart cards to rugged industrial controllers:

Irony Dedicated cryptographic chips: In an ironic twist of fate, chips designed specifically to guard secrets (such as TPM modules, smart cards or HSMs) were among the first targets of side-channel attacks. These devices perform cryptographic operations (RSA, AES, digital signatures) and, in doing so, draw current in varying amounts and emit distinctive electromagnetic radiation. Power analysis attacks have shown that by measuring a smart card's power consumption over thousands of operations, its secret key can be derived. Similarly, with a small inductive coil held close by (an EM probe), it is possible to pick up the “field” around the chip and obtain traces that reveal bits of the key. Researchers have documented just how detectable encryption operation patterns are: for example, a 0 and a 1 can make a circuit consume different amounts of power or take different amounts of time, differences that are reflected in its emissions. Manufacturers have responded with countermeasures (shuffling operations, adding noise, shielding components), but it remains an ongoing race.
Industry Industrial PLCs and physical controllers: In industrial environments, Programmable Logic Controllers (PLCs) and SCADA systems run real-world processes – motors, valves, sensors – and every control action leaves a physical footprint. Think of a PLC switching on a motor: there will be a current spike on the power line, perhaps a small electromagnetic pulse from the change in load, or even an audible relay click. An attacker with close access to a power cable, or simply present in the room with RF instruments, could infer when the PLC is activating certain actuators or read a machine's operating cycles. If, for example, someone at a power plant could remotely measure magnetic field fluctuations in the cables, they might be able to deduce the active load and the commands being sent. There are historical precedents: Cold War espionage systems measured changes in buildings' power lines from a distance to infer when certain devices were in use (for example, teletype activity at an enemy base could be detected through slight fluctuations in the facility's power consumption). Today, with more sensitive sensors, a poorly designed PLC could “snitch” on its own operations to anyone who knows how to listen to the factory's electrical heartbeat.
Everyday Smart cards and IoT devices: Smart cards (chip bank cards, phone SIM cards, etc.) are tiny computers that fit in a pocket. Since they have no battery of their own, they draw power from the reader when inserted, which made the first power analysis attacks easier. But even battery-powered IoT devices, wearables or wireless sensors can give off telltale signals. For example, an attacker nearby could analyze a wearable's radio emission pattern to work out what data it is transmitting or when (without breaking the encryption, simply watching the packet cadence can reveal usage patterns). Another case: smart homes generate power consumption profiles and RF emissions (from Wi-Fi, Bluetooth, Zigbee) that an outside observer can monitor to deduce the occupants' behavior (when they turn on the lights, which TV channel they are watching from the HDMI cable's radiation, etc.). Although these examples go beyond strictly “security hardware”, they illustrate how any repetitive pattern in electronics can be detected externally. Ultimately, all hardware speaks a silent language – in the form of waves and signals – and different types of devices speak it with different accents. Whether it's the pulses of a cryptographic chip, the hum of an industrial controller or the heat traces on a keypad, the universal challenge is the same: can we silence these voices or, at the very least, keep unwanted ears from understanding them?

Personal computers are not the only ones at risk: any complex electronic device can give away information through its physical emissions.

The electromagnetic spy's toolkit: probes, radios and oscilloscopes

To exploit these physical side channels, attackers (or security researchers) rely on a variety of specialized tools, many of them commercially available:

Capture Electromagnetic (EM) probes: These are essentially small antennas or coils designed to pick up electric or magnetic fields very close to a circuit. High-sensitivity probes exist for different frequency bands. An attacker can move an EM probe over different points of a device (for example, near the CPU, near the power cable, etc.) and “listen” to the local emanations. Loop probes detect changing magnetic fields, ideal for capturing clock signals or data pulses on internal buses. Electric field probes (capacitive tips) pick up voltage variations in components. With these tools it is possible to map emission “hot spots” on a piece of hardware and get the best signals for later analysis.
Digitize Oscilloscopes and high-speed digitizers: An oscilloscope is the fundamental instrument for observing electrical signals over time. In power or near-field EM attacks, the probe is usually connected to the oscilloscope to record the waveform of the power consumption or radiated field while the victim device performs sensitive operations. Modern oscilloscopes can sample at several gigahertz, capturing even very brief transitions. By storing millions of samples, one can then apply filtering and statistical calculations to extract patterns correlated with secret data. For example, to break AES encryption through a side channel, a researcher might take 10,000 captures of the power consumption signal while random data is encrypted, then average and correlate those signals against hypotheses about subkeys. This is the classic Differential Power Analysis (DPA) technique that has compromised countless smart cards. Without a good oscilloscope (or a specialized ADC), this level of analysis would be impossible.
Tune Software-defined radios (SDR) and RF receivers: The Software Defined Radio (SDR) revolution has put wide-spectrum receivers in everyone's hands. Devices such as the HackRF One, the (very cheap) RTL-SDR or the aforementioned Funcube Dongle make it possible to tune in to and record signals across wide bandwidths, from kHz to GHz. For an electromagnetic spy, an SDR acts as a remote “ear”: it can pick up a device's emissions without having to be right next to it. For example, a HackRF connected to a simple directional antenna could collect radiation from a monitor or a cable from tens of meters away. The attacker then analyzes those radio recordings on a PC, looking for modulations that correspond to the target's digital activity. What makes it powerful is that SDRs come with flexible software for applying demodulation, Fourier transforms, filtering, etc., all with open-source programs. What once required expensive military receivers can now be done with a USB dongle costing under €300. There are even cases (such as the Tel Aviv attack) in which a traditional portable AM radio was enough to carry out the eavesdropping, given how simple the target signal was. In short, SDRs democratized emission espionage: anyone with basic knowledge can experiment in this field.
Interrogate Bus Pirate v6 and digital interface tools: Although they don't capture emissions themselves, devices such as the Bus Pirate are valuable allies for a hardware attacker. The Bus Pirate is a programmable gadget that can speak a multitude of protocols (UART, SPI, I2C, etc.) and act as an interface to integrated circuits. How does that help in physical attacks? First, it lets a researcher send specific commands or data to a victim device with precision while synchronizing the emission capture. For example, they could automate the encryption of 1000 messages on a cryptographic chip, triggering the oscilloscope measurement for each message, all coordinated by the Bus Pirate. Second, in an espionage scenario, a Bus Pirate could sniff a device's internal communications (if it manages to connect to an exposed internal port) while its EM emissions are monitored at the same time, adding context to the analysis. Version 6 of the Bus Pirate is especially powerful, with higher speed and broader protocol support, which also makes it useful as a hardware pentesting tool (e.g., for reading firmware from internal flash memory and correlating it with emissions). In short, it is a Swiss Army knife for interacting with devices at a low level, complementing purely passive measurements with the ability to inject or sniff supporting digital signals.
Support Supporting equipment and analysis software: Alongside all of the above, a serious attacker might use signal amplifiers (to pick up very weak emissions), band-pass filters (to isolate frequencies of interest), high-gain directional antennas (for operations at a distance, aiming reception at the victim) and even portable Faraday cages (for lab testing free of outside interference). On the software side, mathematical analysis and machine learning tools become important: with so much data collected, finding the needle in the haystack (the fingerprint of the leaked secret) sometimes requires training models or using large-scale correlation algorithms. The key point is that many of these tools are affordable. The biggest barrier used to be knowledge, but today there are tutorials, kits and even open-source communities dedicated to side channels. The 21st-century electromagnetic spy could just as easily be a hobbyist hacker in their garage, not necessarily a government agent with a briefcase full of exotic gear.

Beyond traditional cybersecurity: its limits against physical threats

Most cybersecurity strategies assume a purely digital attack model: protecting passwords, hardening networks, patching software vulnerabilities, and so on. But what happens when the attack is not an exploit in the code but in the physical nature of the device? That is where we run into a gap in traditional security coverage.

Side-channel threats expose several clear limitations:

Blindness Digital controls don't detect analog attacks: A firewall, an antivirus or an intrusion detection system has no way of knowing that an antenna nearby is picking up signals from your PC. These attacks generate no logs, consume no system resources and don't alter the software's behavior. That is why they are said to leave no trace – they are practically undetectable while they are happening. A system can be under active surveillance and neither its users nor its security administrators would have any clue, because every internal monitoring mechanism will keep seeing “all normal”.
Air gap Faith in the “air gap” fades: Traditionally, top-secret data is kept on isolated (air-gapped) computers with no network connections or external devices. This protects against remote intrusion, but it doesn't stop the machine from emitting radiation that can be picked up. As we have seen, even physically segregated systems can be breached by a well-aimed EM attack. Conceptual examples abound: an attacker could leave a hidden microphone in the room to record keystrokes, or a miniaturized, battery-powered RF sensor disguised near the target to relay its emissions. Again, no air-gap policy would detect this, because the leak doesn't travel over Ethernet or USB; it travels through the air or through structural vibrations. So physical separation is no longer an absolute guarantee of security.
Standards Gaps in standards and training: Many corporate security standards (ISO 27001, NIST, etc.) mention “physical security” in passing, but they focus more on controlling access to facilities than on controlling emissions. Outside the military/government sphere, almost no company tests its equipment against compromising emanations or trains its staff on these threats. This creates a false sense of security – we obsess over strong passwords and 2FA, while an industrial spy may literally be listening to our meetings through the wall with a laser, or capturing the data we process from the office next door. Classic cybersecurity rarely brings electronics engineers into its audits, and that is where we miss the chance to uncover these physical-side vulnerabilities before an attacker does.
Exposure Exposed critical infrastructure: Traditional IT environments are at least housed in controlled buildings; but what about dispersed embedded systems? Think of telecom antennas, ATMs, IoT sensors in smart cities… They are deployed in the field and relatively accessible. A malicious actor could approach these devices with receivers and extract information. Take an ATM: even if it encrypts all its communications, the internal bus between the card reader and the encryption module might emit a detectable signal that reveals card numbers. Or the sensors in a petrochemical plant: their wireless transmissions may be encrypted, but the sensor's own analog signal (a 4-20mA loop on a cable) may be interceptable and tell pressure levels to whoever intercepts it. These physical data flows usually fall outside the threat model of conventional cybersecurity.

Purely digital defenses fall short against physical attacks. It is a humbling reminder.

In short, purely digital defenses fall short against physical attacks. It is a humbling reminder: however sophisticated our systems are, they still operate in the real world, obeying the laws of electricity and magnetism. An attacker who thinks outside the box can target the layers nobody is monitoring. And although these attacks require proximity or local access, the history of security teaches us not to underestimate adversaries: with enough motivation, they will find a way to get close (whether by planting a spying device, bribing someone to install equipment, or using drones). Information security has to expand to cover these vectors; otherwise, we will keep reinforcing the front door while information slips out through the invisible cracks in our infrastructure.

The decline of design with purpose and containment

At this point, it is fair to ask: why do our devices leak information so easily? Is it inevitable, or were there design and industry decisions that led us here? This is where we take a critical, almost philosophical, look at technological evolution.

There was a time when building electronic equipment was an art of precision and purpose. At the height of the Cold War, engineers designed with military discipline: every twisted-pair cable, every metal chassis, every filter added served the purpose of containing the system, of making sure it did only what it was supposed to do and nothing more. A well-designed piece of equipment was not supposed to interfere with others or emit spurious signals; in a sense, it respected the principle of electronic silence. But as technology went mass-market, other priorities took over: speed, cost, ease of manufacturing, user convenience. Containment was sacrificed for economic efficiency. The extreme miniaturization of components made leaks harder to control (a 3 GHz chip is inherently a powerful radio-frequency emitter). And Silicon Valley's “move fast and break things” mindset never sat well with the caution and rigor that physical security demands.

Alexander Karp, CEO of Palantir, criticizes in his manifesto how the tech industry “has lost its way”, focusing on trivial consumer products without asking what is worth building or why. That philosophy applies here: we have flooded the world with IoT devices, gadgets and chips everywhere, but how many of those designs seriously took emissions and side-channel resistance into account? Very few, because that rarely scores points on the feature list that sells a product. The decline of design with purpose means we no longer reflect on why each device exists in terms of security and society, only in terms of the market. And the lack of containment shows up as hardware that “shouts” its operations to its surroundings.

We can also see it as a symptom of structural decline in how we train engineers and managers. Comprehensive security requires holistic thinking: a system is only as strong as its weakest link, physical or logical. But at some point we pulled the disciplines too far apart: software developers assume the hardware is trustworthy and silent, and hardware engineers assume security is software's job. This cultural disconnect feeds the problem. The old practice of defensive engineering, which anticipated failures and abuse even at the most basic levels, has gradually been lost. In early military computers, for example, strict TEMPEST standards were imposed – by design, the equipment came in rugged metal enclosures, with filters on every power line, with synchronized clocks so as not to give anything away, and so on. Today, the average PC is an electromagnetic cacophony: its buses spew interference, its plastic components block next to nothing, and its priority is computing speed at any cost. We have forgotten how to build with silence and discipline.

There is also an element of strategic decline: organizations have chosen to invest in glamorous cyber defenses (such as AI for intrusion detection or blockchain to secure records) rather than revisit fundamentals such as unwanted emissions. Saying “we're going to spend money on Faraday enclosures and conductive coating for our boards” is less visible, and perhaps less rewarding, than saying “we're moving to the cloud” or “let's adopt zero trust”. But this neglect could cost us dearly in the long run. We live in an era of fierce geopolitical competition, in which powers and corporations spy on each other by every means available. A well-resourced adversary will notice our design negligence. Some nations are already suspected of building offensive TEMPEST capabilities into their electronic intelligence units, perhaps taking advantage of the fact that most Western critical infrastructure is not hardened against it.

In short, the decline is not technological – we keep advancing in computing power – but a decline in principles. We need to recover the notion that anything not explicitly designed to be secure will probably be insecure. And as things stand, most devices have not been explicitly designed not to emit exploitable information. We have to acknowledge this uncomfortable reality: convenience and haste have made us neglect the silence of our systems. Unless this trend is reversed, we will keep patching with cryptography and software a shortcoming that is fundamentally a matter of hardware and physics, and we will therefore remain exposed.

...toward a culture of silence, discipline and purpose

At the end of this technical and philosophical journey, the question that remains is: what do we do about it? The answer is neither simple nor immediate, but it points to a cultural shift in how we conceive and build technology.

Recover silence: in practical terms, this means once again valuing zero, or at least minimal, emissions. It means designing equipment with an eye to isolating it electromagnetically, suppressing unnecessary emanations, and keeping every LED, every clock and every port from becoming a loudspeaker for data. In a more poetic sense, it means restoring silence as a technological virtue – machines that do their job without broadcasting it to their surroundings. This calls for investment in materials research (better, affordable shielding), updated emission standards for commercial equipment, and perhaps even physical security certifications for consumer products (who will be the first to sell a “TEMPEST-hardened” smartphone to the public?). Silence costs money, no doubt, but the cost of electronic indiscretion can be far higher when we are talking about industrial, financial or state secrets.

Impose discipline: here we are talking about engineering and operational discipline. Discipline to follow security principles built in from the start (Security by Design), which would include security against side channels. It means not cutting corners, not signing off on a design as “good enough” without testing it against physical attacks. Discipline in operations, too: for example, if you handle top-secret information, operate in controlled environments – SCIF-type rooms (Sensitive Compartmented Information Facility) with protected walls, protocols that ban unauthorized electronic devices nearby, and so on. These practices exist in the government world, but they should spread further into the private sector where appropriate. It takes organizational discipline to maintain those policies and to educate staff so they understand that someone else's smartphone in a confidential meeting could be the invisible spy. Karp and other leaders have pointed out that Silicon Valley spent years celebrating the “fail fast” philosophy; perhaps it is time to complement it with “fortify early” – so that engineers regain that almost military rigor in certain security contexts.

Rediscover purpose: in the end, it all comes down to asking ourselves why we build what we build. If the only purpose is to sell more gadgets or ship the next trendy app, security will keep taking a back seat. But if we recover a greater sense of purpose – protecting people's privacy, safeguarding critical infrastructure, preserving strategic advantages – then the extra effort of making our systems more trustworthy at the physical level will be fully justified. Karp talks about focusing technology on challenges that matter for our collective security and well-being. Well, closing these silent espionage channels matters. It is part of building technology worthy of our trust. A truly secure device doesn't just encrypt your data; it stays silent when it should.

At today's crossroads, we have two options. We can ignore these “invisible” threats and carry on until disaster strikes – picture a massive key theft or an act of industrial sabotage carried out by an adversary who exploited uncontrolled emissions. Or we can, starting now, build the accumulated knowledge (from TEMPEST in the 1950s to modern attacks) into the next generation of engineering. Every new chip, every embedded system, should go through not only functional testing but also an electromagnetic examination of conscience: what are you inadvertently revealing? Only with that kind of technical introspection will we get back on the path to comprehensive security.

Ultimately, this article closes with a firm call: let's recover a culture of technical construction built on silence, discipline and purpose. Let machines go back to being obedient, discreet tools, not electromagnetic snitches. Let engineers once again combine efficiency with prudence. And let the tech industry, in its race toward the future, not forget the lessons of the past about the importance of containment. Because in an ever noisier world, true power will belong to whoever can keep quiet.

Silence Value zero emissions again: Design equipment with an eye on how much it emits, not just how well it performs.
Discipline Security built in from the start: Engineering and operational discipline to uphold principles when deadlines get tight.
Purpose Ask why you are building it: If the only purpose is to sell more units, containment never makes it into the budget.

The decline the article describes is not technological — we keep advancing in computing power — but a decline in principles.

Topics

#Hacking-Hardware #Hardware #Hardware-Hacking #Karp #Palantir #Industrial-hardware-security

Keep reading

Talk to us about OT security