Military transhumanism is neither a cinema fantasy nor a bunker secret: it is a research reality, documented and partly public. For decades, agencies such as DARPA have funded work aimed at augmenting endurance, cognition and the interface between brain and machine. But between a funded capability, a laboratory result and a soldier actually equipped, the gap is immense. This guide sets out what is demonstrated, what is still being sought, and why the "invincible super-soldier" remains a myth.
Definition. Military transhumanism designates the body of research and technologies aimed at augmenting the physical, cognitive or sensory capacities of the combatant beyond the normal: stimulants, neuromodulation, brain-machine interfaces, exoskeletons, advanced prosthetics. It differs from mere equipment by its ambition: to modify the human itself, not merely to kit them out.
Every account of the augmented soldier crosses three letters: DARPA. The Defense Advanced Research Projects Agency was created in 1958 by the United States, in reaction to the shock of the launch of Sputnik (1957), with a constant mission: to avoid a new "strategic surprise" by funding high-risk research with high disruptive potential. Its official site publicly documents mission and programmes — it is not a secret outfit, but a research-funding agency.
Its civilian record is considerable and public: a founding role in ARPANET (ancestor of the Internet), GPS, speech recognition. Its operation rests on "programme managers" recruited for a few years, endowed with budgets to fund projects in universities and industry. DARPA does not own large laboratories of its own: it steers and funds. Its annual budget, on the order of a few billion dollars, is modest at the scale of the Pentagon, but its leverage effect is significant.
Several of the most publicised programmes in human augmentation — neural interfaces, advanced prosthetics, memory restoration — fall under its Biological Technologies Office, created in 2014. But the agency assumes a culture of risk (the famous "DARPA-hard") and claims its failures as the price of its breakthroughs. This is an essential point for what follows: many of its "successes" are measured in prototypes and proofs of concept, not in deployed equipment. Investigative journalist Annie Jacobsen, in The Pentagon's Brain (2015), documents this permanent oscillation between real breakthroughs and unkept promises.
This is the golden rule of the whole subject, and the source of almost all confusion. Naming a research capability is not asserting a deployed reality. There are three planes that must never be confused:
Faced with any spectacular claim, three questions suffice to sort the real from the sensational: which source? research capability or deployed reality? laboratory result or field operation? Applied systematically, they dissolve most of the "super-soldier" narratives — without detracting from the attention the real programmes deserve. The rest of this guide applies that grid to each front of augmentation.
This is the most spectacular front, and the one where the documented advances are the most impressive — provided one reads carefully what they establish. Brain-machine interfaces (BCI) have worked, in a medical setting, for more than a decade. In 2012, the BrainGate trial published in Nature (Hochberg et al.) showed people with tetraplegia commanding a robotic arm through the sole activity of their motor cortex, thanks to an intracortical implant. This is an established fact, reproduced since.
DARPA has massively funded this field for rehabilitation purposes. Its Revolutionizing Prosthetics programme led to the advanced "LUKE" prosthetic arm (DEKA, designed under the impetus of Dean Kamen), authorised by the U.S. FDA in 2014. Its Restoring Active Memory (RAM) programme explored "memory prostheses" aimed at restoring the formation of memories in brain-injured patients. Then, in 2018-2019, the Next-Generation Nonsurgical Neurotechnology (N3) programme explicitly targeted non-surgical interfaces, without implants, usable by able-bodied people — this is where the BCI leaves the strict medical setting to aim at the augmentation of the healthy subject.
The private sector has since taken over with major resonance. The company Neuralink announced, in early 2024, the implantation of its brain interface in a first human patient, able to command a cursor by thought. Other companies, such as Synchron with a less invasive endovascular approach, are advancing in parallel. The BCI leaves the academic laboratory alone to become an industrial and commercial object.
A distinction is capital here, and too often skirted: reading a motor intention (an established reality) is not reading thoughts (not established). Decoding the signal that commands a movement is not reading the content of a mind. The DARPA Silent Talk programme (2009), which explored "brain-to-brain" communication through detection of inner language, remained limited and exploratory: it illustrates a declared ambition, not a reached capability. The "telepathic soldier" does not exist.
The trajectory of BCIs — from the therapeutic (restoring a lost function) towards the potentially augmentative (adding a function to an intact body) — has given rise to an unprecedented legal debate: neurorights. As early as 2017, researchers led by Rafael Yuste proposed to enshrine in law the protection of mental privacy and identity. Chile legislated in 2021, and UNESCO is drafting a global recommendation on the ethics of neurotechnologies.
The inner forum, long protected by its sheer inaccessibility, must now be protected explicitly, before technical capabilities force access to it. This is a remarkable case: one legislates ahead of a capability that does not yet exist as a tool of control. Nothing, to date, constitutes an operational capability for "mind control": it is a risk hypothesis, not a fact. But that is precisely why neurorights are emerging now — and that is wise.
Before the brain, augmentation first targeted the body. The most publicised project, TALOS (Tactical Assault Light Operator Suit) of the U.S. special operations command (USSOCOM), launched in 2013 and sometimes nicknamed "the Iron Man suit," was officially abandoned in 2019 for failing to result in a viable integrated system. It is the perfect example of the gap between ambition and reality: a flagship, publicised programme that was never deployed.
Other prototypes have marked this history — Berkeley's BLEEX, Lockheed Martin's HULC, the XOS of Sarcos/Raytheon. All ran up against the same obstacles: energy autonomy, weight, bulk, cost. A recurring law of the field imposes itself: energy is the wall. A powered armour is useful only as long as it is powered; dependence on a battery turns an advantage into a vulnerability. The fantasy of the foot-soldier-tank runs up, again and again, against physics.
The successful applications of exoskeletons are, to date, more modest and often logistical or civilian: assisting the carrying of heavy loads, reducing musculoskeletal disorders (for example the Guardian XO from Sarcos). On the prosthetics side, by contrast, the advances are real and change lives: myoelectric arms, limbs controlled by residual nerve signals, sensory feedback (DARPA's HAPTIX programme). The therapeutic frontier is clear there — restoring a lost function — even if the most advanced prosthetics already pose the question of the threshold beyond which repairing becomes augmenting.
The first front of augmentation is the oldest: pushing back the limits of endurance and sleep. The use of stimulants by armies is nothing new — amphetamine and methamphetamine were distributed during the Second World War. The contemporary novelty lies in the shift to modafinil, a wakefulness promoter studied by several air forces as a countermeasure to fatigue for long missions, with, according to the studies (notably the work of John Caldwell for the U.S. Air Force), an effects profile different from that of amphetamines.
But these tools do not create energy: they redistribute a debt. Modafinil does not suppress the need for sleep; it defers its payment. Military research has moreover explored the reverse of the stimulant — the "sleep bank," preventive accumulation of sleep before a deprivation — an admission by science that the best augmentation of endurance remains, paradoxically, sleep itself.
On the brain side, transcranial direct-current stimulation (tDCS) has been studied by the U.S. Army for attention and learning speed. The published results are real but modest and debated: the literature is marked by strong heterogeneity of protocols, and several studies struggle to reproduce the effects. This is exactly the kind of field where media enthusiasm outpaces the solidity of the evidence. The promise of "downloaded learning" à la Matrix remains fiction.
Here is the most inflammable terrain, hence the one demanding the greatest rigour. Genome-editing tools, in particular CRISPR-Cas9 (Nobel Prize in Chemistry 2020 for Emmanuelle Charpentier and Jennifer Doudna), have made genetic modification infinitely more accessible — a major scientific fact, with applications above all medical and agricultural. DARPA moreover launched in 2017 a programme named Safe Genes, whose stated object is defensive: to control, counter or reverse genetic modifications. The name says the orientation: safe, safety.
It must be said clearly: the genetically augmented soldier does not exist, and the obstacles are not only ethical or regulatory — they are biological. Complex human traits (strength, courage, endurance) are polygenic, shaped by the environment, and are not "programmed" by editing a single gene. The He Jiankui case (2018), the Chinese researcher who announced the birth of babies with a genome modified by CRISPR, sentenced to three years in prison in 2019, sets the real limits: human germline editing is technically possible but ethically and legally proscribed, with unpredictable off-target effects. What is real is research on genome editing — not a regiment of supermen.
One can document everything of the "how"; there remains the "should one." The founding distinction of the bioethics of augmentation opposes therapy — restoring a normal function — and enhancement — carrying a function beyond the normal. The U.S. President's Council on Bioethics devoted a reference report to it, Beyond Therapy (2003). The report Enhanced Warfighters: Risk, Ethics, and Policy (Patrick Lin et al., 2013) poses concrete questions: is the soldier's consent free within a hierarchical framework? Is augmentation reversible at the end of service? Who answers for long-term effects?
The real ethical problem of the augmented soldier is not that he would become "too powerful." It is that it puts under tension three principles: consent (can one consent freely under military authority?), dignity (is the human a system to be optimised?) and responsibility (who answers for the consequences?). Added to this is "enhancement creep": a technique introduced as exceptional tends to become routine, then expected, then de facto obligatory. And these technologies are dual-use: the same interface, the same molecule, the same exoskeleton pass smoothly from the military field to the civilian field. The soldier is the scout of a question that concerns us all.
An honest subject must say what it is not, for this terrain is one of the most contaminated by confusion. The CIA's MKUltra programme (1950s-1970s), devoted to mind control and the use of substances such as LSD, really existed: it is documented by declassified archives and the Congressional investigation (Church Committee, 1975-1976). It is an established historical fact — and a recognised moral fault. But MKUltra belongs to the past, and it is regularly summoned to "prove" present-day capabilities of mind control which, themselves, are not established. The logic "they lied yesterday, therefore everything is true today" is a fallacy: a past lie founds vigilance, not credulity.
The "super-soldier serum," the invincible armour, the telepathic or genetically perfect combatant belong to fiction (from Captain America to video games) far more than to the laboratories. The public figures of transhumanist foresight — Nick Bostrom, who theorises the risks and promises of human enhancement, or Yuval Noah Harari, who describes a future of "Homo Deus" — offer positions and scenarios to be read as analytical hypotheses, not as verified prophecies. The real programmes, for their part, often fail (TALOS), produce modest effects (tDCS), or remain strictly medical (BrainGate, LUKE). There lies the honest measure of military transhumanism.
Does DARPA's augmented soldier really exist? Yes and no. The research exists, documented and partly public: advanced prosthetics (LUKE), brain-machine interfaces (RAM, N3), fatigue countermeasures. But no deployed "super-soldier" follows from it. DARPA funds proofs of concept; the gap with a soldier actually equipped remains immense, and several flagship programmes (TALOS) have been abandoned.
Does Neuralink create super-soldiers? No. Neuralink announced in 2024 a first implant in a human patient commanding a cursor by thought — a medical use, in the lineage of BrainGate (2012). Reading a motor intention is not reading thoughts nor controlling a mind. No operational military augmentation capability follows from it to date.
Are there genetically modified soldiers? No. Complex human traits are polygenic and shaped by the environment: one does not "programme" them by editing a gene. Human germline editing is technically possible (He Jiankui case, 2018) but largely proscribed. The DARPA Safe Genes programme is defensive. The genetic soldier remains a myth, for reasons that are first of all biological.
Should one be afraid of military transhumanism? The right posture is neither panic nor denial. The real danger is not a laboratory superman, but concrete ethical questions: consent under hierarchy, reversibility, enhancement creep, neurorights. These technologies being dual-use, what is authorised for the soldier prefigures what will be authorised for all. Useful vigilance is ethical and legal, not conspiracist.
To go further, our investigation The Soldier of Tomorrow gathers these real sources (DARPA, BrainGate in Nature, Lin et al., Yuste's neurorights), flags each claim by its level of certainty, and methodically untangles the demonstrated from the speculated — to understand where the frontier between healing and augmenting truly runs.
Yes and no. The research exists, documented and partly public: advanced prosthetics (LUKE), brain-machine interfaces (RAM, N3), fatigue countermeasures. But no deployed "super-soldier" follows from it. DARPA funds proofs of concept; the gap with a soldier actually equipped remains immense, and several flagship programmes (TALOS) have been abandoned.
No. Neuralink announced in 2024 a first implant in a human patient commanding a cursor by thought — a medical use, in the lineage of BrainGate (2012). Reading a motor intention is not reading thoughts nor controlling a mind. No operational military augmentation capability follows from it to date.
No. Complex human traits are polygenic and shaped by the environment: one does not "programme" them by editing a gene. Human germline editing is technically possible (He Jiankui case, 2018) but largely proscribed. The DARPA Safe Genes programme is defensive. The genetic soldier remains a myth, for reasons that are first of all biological.
The right posture is neither panic nor denial. The real danger is not a laboratory superman, but concrete ethical questions: consent under hierarchy, reversibility, enhancement creep, neurorights. These technologies being dual-use, what is authorised for the soldier prefigures what will be authorised for all. Useful vigilance is ethical and legal, not conspiracist.
Dossier : Transhumanisme militaire & DARPA
Accueil · Collection · Journal · Dossiers · Sources