-
Brain-to-Brain Communication: Where the Science of Direct Neural Links Actually Stands
In 2019, researchers at the University of Washington published a paper in Scientific Reports describing BrainNet—a system that allowed three people, seated in separate rooms with no ability to see, hear, or talk to each other, to collaboratively play a Tetris-like game using only their brain signals. Two “senders” could see the game board and decided whether a falling block needed to be rotated. They communicated their decisions to a “receiver” who couldn’t see the board but controlled the game. No words. No gestures. No screens shared between them. The senders’ decisions were extracted via EEG, transmitted over the internet, and delivered to the receiver’s visual cortex via transcranial magnetic stimulation, where they appeared as flashes of light—phosphenes—that the receiver interpreted as instructions. Five groups of three people tested the system and achieved 81 percent accuracy.
That’s the headline. Here’s the fine print: the information transmitted was binary. Yes or no. Rotate or don’t rotate. One bit of data per transmission cycle. The senders communicated their decisions by staring at lights flashing at different frequencies—15 hertz for one answer, 17 hertz for the other—which entrained their brain’s electrical output at the corresponding frequency, readable by EEG. The receiver experienced either a flash of light (rotate) or no flash (don’t rotate). The “brain-to-brain communication” was, functionally, a very elaborate way to send the equivalent of one binary digit from one head to another. IEEE Spectrum described an earlier version of this approach as “telepathic Morse code.”
This is what brain-to-brain communication actually looks like in 2026: technically real, scientifically genuine, and approximately as far from telepathy as a tin-can telephone is from a 5G network.
What exists
The field has produced a series of legitimate demonstrations, each constrained by the same fundamental bottleneck: you can get information out of a brain with reasonable resolution using EEG or implanted electrodes, but you can deliver information into a brain noninvasively only through crude channels—magnetic pulses that trigger phosphenes (perceived flashes of light) or vague sensations. The input side is the constraint. Reading a brain is hard. Writing to a brain is harder by orders of magnitude.
The 2014 Starlab experiment was the first reported human brain-to-brain transmission. A sender in India imagined moving his hands or feet to encode binary data through EEG. The signal was emailed to France, where a TMS device delivered pulses to a blindfolded receiver’s visual cortex, producing phosphenes. The receiver reported the flashes verbally, and the team decoded the message. The transmitted words: “hola” and “ciao.” The transmission rate was approximately two bits per minute. The entire process took over an hour.
BrainNet in 2019 scaled the architecture to three people and demonstrated something genuinely interesting beyond the binary channel: when the researchers injected noise into one sender’s signal, the receiver learned to preferentially weight the more reliable sender—a trust calibration process that happened entirely through brain-to-brain signals without any conscious strategy. The receiver’s brain was doing signal integration across two noisy sources, the same computation that underlies sensory integration in normal perception.
Invasive brain-computer interfaces—Neuralink, Synchron, Blackrock Neurotech—are advancing rapidly on the reading side. Neuralink implanted its first human chip in January 2024 under its PRIME study, enabling a paralyzed patient to type and control a cursor through thought alone. Synchron’s Stentrode sits inside a blood vessel near the brain, avoiding open surgery. The PRIME study has a primary completion date of 2026 and full study completion projected for 2031. These systems are brain-to-computer interfaces, not brain-to-brain—they translate neural signals into digital commands for external devices. But they represent the reading infrastructure that any brain-to-brain system would eventually need.
On the AI-assisted decoding side, researchers at the University of Texas in 2023 used fMRI scans and large language models to decode continuous thought into coherent text—not single words or binary choices but streams of semantic content, capturing the gist of what a person was thinking about during a story or imagined narrative. Meta has developed noninvasive brain-scanning systems paired with AI models that can decode silently spoken words from brain activity. These aren’t brain-to-brain systems, but they’re solving the bandwidth problem on the reading end: extracting richer, more nuanced information from neural signals than EEG-based approaches can achieve.
What doesn’t exist
Telepathy—the transmission of complex thoughts, images, emotions, or experiences from one mind to another—is not close. The demonstrations that exist transmit binary decisions through artificial sensory channels. The receiver doesn’t “hear” the sender’s thought. The receiver sees a flash of light and interprets it according to a pre-agreed code. The brain-to-brain interface is a translation chain: thought → EEG signal → digital encoding → internet transmission → TMS pulse → phosphene → interpretation. At every link in that chain, information is lost. What arrives in the receiver’s brain is not a thought. It’s a stimulus—a magnetically induced visual artifact that carries one bit of information about the sender’s decision.
The gap between this and actual telepathy is not a gap that incremental engineering will close, because the limiting factor isn’t the technology between the brains. It’s the fundamental problem of neural encoding: we don’t know, for any given thought, which specific neural firing patterns represent it, how those patterns vary between individuals, or how to induce a specific firing pattern in a target brain that would be experienced as the same thought. Brains aren’t standardized hardware. The neural code for “rotate the block” in one person’s motor cortex is not the same pattern in another person’s motor cortex. Translating one person’s neural representation into a stimulus that would produce the same internal experience in another person requires a mapping between two unique neural architectures—a problem neuroscience hasn’t solved and isn’t close to solving.
What BCI companies are building toward is not telepathy but increasingly high-bandwidth brain-to-computer interfaces that could, in principle, be linked: Brain A → computer → Brain B. Neuralink’s implant reads neural signals at thousands of channels. Future implants will read more. AI decoding systems are getting better at extracting semantic content from neural data. But the write side—delivering complex, precise, meaningful information directly into neural tissue in a way that the receiving brain interprets as a coherent experience—remains the unsolved problem. TMS can trigger phosphenes and crude sensory impressions. It cannot implant a sentence, an image, an emotion, or a memory.
The timeline problem
Coverage of brain-to-brain communication tends to imply a trajectory: binary transmission today, sentences tomorrow, telepathy eventually. The trajectory is real in the same way that the Wright Brothers’ 12-second flight in 1903 implied commercial aviation—the physics supports the possibility, but the engineering required to get from demonstration to deployment is measured in decades, not years, and the technical obstacles on the write side are qualitatively different from the obstacles on the read side.
Reading a brain is an information extraction problem: the neural signals are there, and the challenge is building sensors sensitive enough and algorithms smart enough to decode them. This problem is yielding to better hardware and better AI. Writing to a brain is an information implantation problem: you need to induce specific patterns of activity in specific neural populations at specific times, through skull and tissue, without disrupting the brain’s existing activity. Noninvasive methods (TMS, focused ultrasound, transcranial electrical stimulation) affect large regions of cortex with limited spatial precision. Invasive methods (optogenetics, direct electrical stimulation) can target individual neurons but require surgery, gene therapy, or implanted hardware.
The honest assessment in 2026: brain-to-computer interfaces are advancing on a trajectory that will produce clinically meaningful products for paralysis, communication disorders, and sensory prosthetics within the current decade. Brain-to-brain communication, in the sense of transmitting complex mental content between two people, requires solving the neural write problem at a resolution and precision that current technology can’t achieve and that current neuroscience can’t specify. The demonstrations are real. The extrapolation to telepathy is premature by a margin that is difficult to estimate because the bottleneck isn’t engineering velocity. It’s a scientific knowledge gap about how brains encode experience—a gap that better instruments may close but that no existing roadmap guarantees.
Neuralink named its first consumer product “Telepathy.” The name is aspirational in the way that calling the first automobile a “teleporter” would have been aspirational. The product lets a paralyzed person control a cursor with their thoughts. That’s extraordinary and useful. It’s not telepathy. The distance between the two is the distance between reading a book and writing one—and in neuroscience, we’re still learning to read.
We cover brain-to-brain communication alongside spinal cord stimulation, retinal implants, and the full landscape of neural interface technology across our Neuroprosthetics course—including why the hardest problem in connecting two brains isn’t getting the signal out. It’s getting the signal in.
-
Orangutan Self-Medication: How Great Apes Choose Plants to Treat Their Own Wounds and Infections
On June 22, 2022, researchers at the Suaq Balimbing research station in Sumatra’s Gunung Leuser National Park heard a series of long calls from the canopy—the vocalizations male orangutans produce during dominance confrontations. The next day, they noticed that a flanged male orangutan named Rakus had a fresh wound on his right cheek, just below the eye, probably from a fight with a neighboring male. Three days later, they watched him do something no wild animal had ever been documented doing: he selected a specific plant—a climbing vine called Fibraurea tinctoria, known locally as Akar Kuning—ripped off its leaves, chewed them for 13 minutes, and then spent seven minutes applying the resulting juice directly to his wound with his fingers. He didn’t swallow the leaves during the application phase. When flies began landing on the wound, he covered it entirely with the chewed plant material, creating a poultice. The next day, he returned to the same plant and ate more leaves. Within five days, the wound closed. By July 19—roughly a month after the injury—only a faint scar remained. No infection developed.
The paper, published in Scientific Reports in May 2024 by Isabelle Laumer and Caroline Schuppli of the Max Planck Institute of Animal Behavior, called it “the first known case of active wound treatment in a wild animal with a medical plant.” The emphasis on “active” is deliberate. Animals have been observed swallowing plants with medicinal properties before—chimpanzees chew bitter pith, gorillas and bonobos swallow rough leaves whole to mechanically dislodge intestinal parasites, Bornean orangutans rub chewed plants on their limbs. But those behaviors involve ingestion or generalized application. What Rakus did was topical, targeted, and sequential: he applied the plant’s juice specifically to the wound, on no other body part, repeated the application multiple times, and then covered the wound with plant material. He treated his own injury the way a human would treat a cut—clean it, apply medicine, bandage it.
Why Akar Kuning matters
Fibraurea tinctoria is not a random plant. It’s a climbing liana found across Southeast Asia—Indonesia, Malaysia, Thailand, Vietnam—and it’s used extensively in traditional medicine to treat dysentery, diabetes, malaria, and infections. Chemical analysis of the plant has identified furanoditerpenoids and protoberberine alkaloids with documented antibacterial, anti-inflammatory, antifungal, antioxidant, and analgesic properties. The plant also contains jatrorrhizine, which has antimicrobial and anticancer properties, and palmatine, which has anti-inflammatory and antiviral effects. This isn’t a plant that happens to have healing properties. It’s a plant whose healing properties are well-characterized enough that humans have been using it medicinally for centuries.
Rakus’s population at Suaq Balimbing rarely eats it. In 21 years and roughly 390,000 feeding observations at the site, Fibraurea tinctoria appeared in only 0.3 percent of feeding scans. This wasn’t a plant the orangutan was already eating when he happened to touch his wound. He selected it specifically, used it in a way that doesn’t correspond to normal feeding behavior, and applied it exclusively to the injury. The researchers were careful to note that in 21 years and 28,000 observation hours, they had never previously seen an orangutan use leaves to treat a wound.
How deliberate was it?
This is the question the paper addresses directly and honestly. The behavior appeared intentional: Rakus selectively treated only his facial wound, not other body parts. He repeated the application multiple times. He used both the juice (liquid application) and the solid plant material (poultice). The entire process—feeding on the plant, applying the juice, covering the wound—took a considerable amount of time and was sustained across two consecutive days. The sequence is difficult to explain as accidental.
But the researchers offer two possible origin stories, and they’re transparent about not being able to distinguish between them. The first is “accidental individual innovation”—Rakus may have been feeding on the plant, accidentally touched his wound while chewing, felt immediate pain relief from the plant’s analgesic effects, and then repeated the behavior because it worked. Under this model, the behavior was discovered by accident and reinforced by its consequences, which is how a lot of animal tool use and self-medication originates. The second possibility is social learning—Rakus wasn’t born at Suaq Balimbing. Male orangutans disperse from their natal area during or after puberty, sometimes traveling long distances. Rakus may have observed the behavior in his birth population, carried the knowledge across dispersal, and applied it when the situation required. If so, the behavior represents a cultural tradition transmitted between individuals, not an individual invention.
The researchers can’t determine which explanation is correct because they don’t know where Rakus was born or what behaviors are practiced in that unknown population. This ambiguity is frustrating but honest—and it’s the central challenge of studying animal self-medication in the wild. You’re observing rare behaviors in long-lived animals across vast landscapes with limited coverage, and the most interesting questions (was it invented or learned?) require data from populations you may never have access to.
The broader landscape of animal self-medication
Rakus’s wound treatment is the most dramatic documented case, but self-medication in animals—zoopharmacognosy—is a recognized field with decades of evidence across multiple species and continents.
Chimpanzees at multiple African field sites chew the bitter pith of Vernonia amygdalina, a plant with antiparasitic compounds, when they’re suffering from intestinal infections. The behavior is targeted: chimps eat it when sick and avoid it when healthy, suggesting they’re responding to internal cues rather than eating it as a regular food. Gorillas, chimpanzees, and bonobos swallow rough, hairy leaves from Aspilia and other plants whole and without chewing—the leaves pass through the digestive tract intact and physically dislodge intestinal parasites, which researchers have confirmed by examining fecal samples and finding parasites wrapped in leaf material. This is mechanical self-medication: the plant’s physical properties, not its chemistry, provide the therapeutic effect.
Bornean orangutans have been observed rubbing chewed leaves of Dracaena cantleyi on their limbs, producing a lather that may have anti-inflammatory or antiparasitic properties. Capuchin monkeys rub citrus fruits and certain plants on their fur, potentially as insect repellent. Some moth species lay their eggs on alkaloid-rich plants when infected by parasitoid wasps, effectively medicating their offspring by ensuring the larvae consume antiparasitic compounds. Even fruit flies preferentially consume alcohol-containing food when infected by parasitoid wasps—the ethanol kills the wasp larvae developing inside them.
The pattern across these examples is consistent: animals with no understanding of chemistry, pharmacology, or infection select specific substances with specific biological activity in response to specific health conditions. The behavior isn’t random foraging. It’s condition-dependent, substance-specific, and in many cases targeted to the affected body region. The question isn’t whether animals self-medicate. They do. The question is what cognitive mechanism enables it.
What it means for the origins of medicine
The earliest known human medical manuscript, from Mesopotamia around 2200 BCE, describes wound treatment with plant-based remedies. But if a Sumatran orangutan—separated from the human lineage by roughly 14 million years of evolution—independently applies a biologically active plant to a wound and covers it with a poultice, the implication is that the cognitive capacity for wound treatment predates the human lineage entirely. Laumer and Schuppli suggest that “medical wound treatment may have arisen in a common ancestor shared by humans and orangutans,” and that the behavior observed in Rakus may reflect deep evolutionary roots rather than a recent invention.
The alternative—that Rakus and the Mesopotamian scribe independently arrived at the same solution—is possible but requires the same cognitive prerequisites: recognizing that a wound needs treatment, selecting a substance with appropriate properties, applying it specifically to the injury, and sustaining the behavior long enough for healing to occur. Whether the common ancestor had this capability or whether it evolved convergently in hominids and orangutans, the conclusion is the same: medicine didn’t start with humans. It started with primates who paid attention to what made them feel better and repeated it.
Traditional healers in Indonesian Borneo have reportedly learned plant-based remedies by observing orangutan behavior—the knowledge transmission running from ape to human rather than the reverse. If Rakus learned his wound treatment from his natal population, and if human populations learned similar treatments from watching orangutans, then the same medicinal knowledge has been transmitted across species boundaries in both directions. The forest pharmacy has always been open. The question is who figured out the inventory first.
We cover orangutan self-medication alongside baboon politics, ant collective intelligence, and the full landscape of animal cognition across our Animal Culture & Knowledge course—including why the first pharmacist may not have been a person. It may have been a primate with a cheek wound and the sense to reach for the right vine.
-
Electric Eels and Electroreception: How Some Animals Perceive a World of Electricity Humans Can’t See
Every living thing generates faint electrical fields. Your muscles produce tiny voltages when they contract. Your heart creates a rhythmic electrical pulse detectable from outside your body. The chemistry between salt ions and cellular membranes generates fields that radiate into the surrounding environment. For most animals—including us—these fields are invisible, unfelt, entirely outside perceptual experience. For roughly 350 species of fish, a handful of amphibians, two groups of mammals, at least one species of dolphin, and possibly bumblebees, they are as perceptible as light is to a sighted animal. These organisms sense electricity the way we sense sound—through dedicated receptor organs that convert electrical signals into neural information the brain can interpret. They live in a sensory world that humans cannot access without instruments, and some of them have been doing it for over 500 million years.
Two kinds of electrical sense
Electroreception comes in passive and active forms, and the distinction matters because they represent fundamentally different relationships with the environment.
Passive electroreception is detection without emission. The animal senses electrical fields generated by other organisms or by the environment itself. Sharks are the canonical example. Their ampullae of Lorenzini—pores in the skin connected by gel-filled canals to nerve endings—can detect voltage changes as small as 0.05 microvolts per centimeter. That sensitivity is difficult to convey in human terms, but here’s what it means operationally: a hammerhead shark can locate a flounder buried under sand and completely invisible to vision, sonar, or olfaction, by sensing nothing more than the electrical field generated by the flounder’s beating heart and contracting gill muscles. Camouflage is useless against an electroreceptive predator. You can match the color and texture of the seafloor perfectly, and the shark will still find you, because you can’t stop your muscles from generating electricity while you’re alive.
The ampullae of Lorenzini evolved early in vertebrate history—they appear in both cartilaginous fish like sharks and in ancient bony fish like coelacanths and sturgeons, which means the basic architecture predates the split between those lineages, placing its origin at roughly 500 million years ago. Most modern bony fish have lost the ancestral electroreceptors, but the sense has been independently reinvented multiple times in different lineages using different tissue types—a pattern of convergent evolution that tells you the survival advantage is significant enough to be worth rebuilding from scratch.
Active electroreception is stranger. The animal generates its own electric field using a specialized electric organ—modified muscle or nerve tissue, typically in the tail—and then monitors that field for distortions caused by nearby objects. Anything in the environment that conducts electricity differently from the surrounding water—a rock, a plant, another fish, a predator—warps the field in a detectable way. The animal perceives the size, shape, distance, and electrical conductivity of objects in its vicinity without light, without sound, without physical contact. It’s echolocation with electricity instead of sound waves.
Two groups of freshwater fish have independently evolved active electroreception: the South American knifefishes (Gymnotiformes), which include the electric eel, and the African elephantfishes (Mormyridae). Both live in turbid water where visibility is low, and both use their electric fields for navigation, foraging, and communication. Weakly electric fish modulate their discharge patterns to signal to conspecifics—territorial claims, mating readiness, species identity—essentially talking through electrical pulses that other species can’t perceive.
What the electric eel actually does
The electric eel—Electrophorus electricus, technically a knifefish rather than a true eel—is the most famous electroreceptive animal and possibly the most misunderstood. Its high-voltage discharge (up to 860 volts in the sister species E. voltai, roughly half the voltage of a taser) has been known for centuries, but until recently it was understood purely as a weapon. Research published in Nature Communications revealed something more sophisticated: electric eels use their high-voltage discharge simultaneously as a weapon and as a precision tracking system.
The eel generates high-frequency pulses during a strike—reminiscent of the “terminal feeding buzz” that bats produce during the final approach to an insect—and uses the return signal to track the position of fast-moving prey in real time. When researchers separated the mechanosensory cue (water movement from a fleeing fish) from the electrosensory cue (a conductor in the water), eels initially struck toward the water movement but redirected their final approach toward the conductor. Strikes initiated in the absence of a conductor were aborted entirely. The eel doesn’t just stun prey and then grope around for it. It stuns prey and tracks its precise location through the same discharge, using a single pulse of electricity for two completely different functions—immobilization and radar—simultaneously.
The electric organ itself is a stack of electrocytes—modified muscle cells, each generating a small voltage. The cells are arranged in series, like batteries in a flashlight, so their individual voltages add up. An electric eel’s body is roughly 80 percent electric organ by volume. The animal is, functionally, a biological battery with fins.
The platypus: electroreception reinvented
Mammals lost electroreception entirely when they moved to land—the sense works through water, which conducts electricity well, and air, which doesn’t. The platypus lineage reinvented it after returning to a semi-aquatic lifestyle, but using completely different hardware than fish. Instead of ampullae of Lorenzini derived from the lateral line system, the platypus evolved electroreceptors from mucous glands in the skin of its bill—roughly 40,000 of them, arranged in front-to-back stripes.
The platypus hunts with its eyes, ears, and nostrils closed. Underwater, it sweeps its bill through river-bottom mud, detecting the tiny electrical pulses generated by muscle contractions of shrimp, insect larvae, and small crustaceans. The electroreceptors work in concert with mechanoreceptors (pressure sensors) on the bill, and the platypus appears to triangulate prey distance by measuring the delay between the arrival of electrical signals and pressure waves—the electrical signal, traveling at near-light speed through water, arrives before the pressure wave, and the time difference encodes distance. The platypus makes rapid side-to-side head movements called saccades—the same term used for the quick eye movements humans make when scanning a visual scene—to update its electrical map of the environment.
The echidnas, the platypus’s closest living relatives, retained a diminished version: long-beaked echidnas have about 2,000 electroreceptors, short-beaked echidnas around 400, both near the end of the snout. Long-beaked echidnas feed on earthworms in tropical forest leaf litter—wet enough to conduct electricity. Short-beaked echidnas eat termites and ants in dry environments, but the interiors of nests are presumably humid enough for the sense to function.
The Guiana dolphin—Sotalia guianensis—adds another independent reinvention. Hairless pits on its rostrum, originally associated with the whisker follicles that all mammalian embryos develop, function as electroreceptors sensitive to fields as low as 4.8 microvolts per centimeter. Research on bottlenose dolphins published in 2023 demonstrated passive electroreception in that species as well, suggesting the capability may be more widespread among cetaceans than previously recognized.
Bees, flowers, and the electrical channel
The most recent expansion of the electroreception story moved it from water to air, where it shouldn’t work—air is a poor conductor. But bumblebees carry a positive electrical charge accumulated during flight, and flowers hold a slight negative charge. When a bee approaches a flower, the electric fields interact, and tiny mechanosensory hairs on the bee’s body deflect in response. The deflection carries information: a flower that has been recently visited by another pollinator has a different charge profile than an unvisited one, because the previous bee’s charge partially neutralized the flower’s field. The bee can detect whether a flower is worth landing on before it arrives—an electrical “occupied” sign that saves energy and time.
This finding—that electroreception functions in terrestrial arthropods through air, using mechanisms entirely unrelated to the aquatic electroreception of fish and mammals—suggests the sense may be more widespread than the aquatic bias of early research indicated. Aerial electroreception is, as one researcher noted, an emerging field. The pun is unavoidable and the science is real.
What it tells us about perception
Electroreception is the clearest evidence that the human sensorium is not the default model for perceiving the world. We see light, hear sound, feel pressure, detect chemicals as taste and smell. We assume this is what the world is. For an electroreceptive animal, the world also contains a continuous electrical layer—fields radiating from every living organism, distortions created by every conductive object, signals modulated for communication between individuals of the same species. That layer is as real as light. We just can’t see it.
The philosophical implication—raised by biologist Jakob von Uexküll’s concept of the Umwelt, the species-specific perceptual world each organism inhabits—is that reality as perceived by any animal is a filtered subset of physical reality, shaped by the sensory equipment evolution happened to provide. The shark’s reality includes the heartbeat of a buried fish. The bee’s reality includes the charge state of a flower. The platypus’s reality includes an electrical map of the riverbed, constructed with closed eyes in complete darkness. Ours doesn’t include any of these things, and until we built voltmeters, we didn’t know they were there.
We cover electroreception alongside cuttlefish camouflage, octopus distributed cognition, and the full landscape of comparative neuroscience across our Neurozoology course—including why the most important thing about the electric eel isn’t the voltage. It’s the fact that the same pulse that stuns its prey also tells it exactly where the prey is.
-
Water as a Strategic Resource: Which Countries Control the Rivers & Infrastructure Other Countries Need
On March 7, 2026, Iran’s foreign minister accused the United States of attacking a freshwater desalination plant on Qeshm Island in the Strait of Hormuz, disrupting water supply to 30 villages. The next day, Bahrain reported that an Iranian drone had damaged one of its 103 desalination plants. Iran’s parliament speaker then warned that if the coalition occupies an Iranian island with regional support, “all the vital infrastructure of that regional country will, without restriction, become the target of relentless attacks.” The vital infrastructure he meant was water. More than 400 desalination plants line the shores of the Arabian Gulf. They produce over 40 percent of the world’s desalinated water. Qatar gets 99 percent of its drinking water from desalination. Kuwait and Bahrain get over 90 percent. Without these plants, roughly 100 million people in the Gulf region would have no regular access to potable water. The petrostates are, as one scholar framed it, saltwater kingdoms—societies whose survival depends on converting seawater into drinking water at industrial scale, powered by the same fossil fuels that made them wealthy. The Iran war has turned that dependency from an engineering fact into a military vulnerability.
This is the version of water conflict that the 21st century actually produces: not armies fighting over a riverbank, but missiles aimed at the machines that make seawater drinkable.
The rivers that run through other people’s countries
Two hundred and sixty international river basins account for approximately 60 percent of the world’s freshwater. They cover nearly half of the earth’s surface and serve 40 percent of the global population. No formal agreement guarantees equal shares in 60 percent of those basins. The geopolitics of water is determined by a single structural fact: rivers flow downhill, which means the country upstream controls the water that the country downstream needs to survive.
Ethiopia’s Grand Ethiopian Renaissance Dam on the Blue Nile is the most consequential current example. Egypt depends on the Nile for 97 percent of its freshwater—a dependency so total that any upstream dam represents, from Cairo’s perspective, an existential threat. Ethiopia began filling the GERD’s reservoir in 2020. Egypt has framed the issue as a matter of national security. The Arab League’s May 2025 Baghdad Declaration elevated “Arab water security” to a shared strategic imperative, explicitly championing Egypt’s position—despite the headwaters of the Nile originating in non-Arab Ethiopia. Diplomatic negotiations have stalled repeatedly. The dispute has been ongoing for over a decade, with no binding resolution, and Ethiopia’s position—that it has sovereign rights to develop hydropower on a river within its borders—is as legally defensible as Egypt’s claim that historical usage entitles it to the Nile’s flow.
Turkey’s Southeastern Anatolia Project on the Tigris and Euphrates is the second flashpoint. Turkey’s dam-building programs have reduced Iraq’s water supply along both rivers by 80 percent since 1975. The Ilisu Dam on the Tigris generates less than half its potential energy output—climate-driven precipitation drops in the watershed caused reservoir levels to fall below operational thresholds in 2022—but it functions as a geopolitical lever regardless. Turkey uses water infrastructure to extract economic and political concessions from Iraq, a dynamic that will intensify as climate change reduces precipitation across the basin.
China’s cascade of dams on the upper Mekong—known in China as the Lancang—gives Beijing disproportionate control over water flows that Cambodia, Vietnam, Laos, and Thailand depend on for agriculture, fisheries, and hydropower. The Mekong River Commission exists as a platform for dialogue, but China is not a member. On the Brahmaputra, Chinese diversion projects raise fears in India and Bangladesh. The Tibetan Plateau—sometimes called “Asia’s water tower”—is the source of rivers that sustain billions of people across South and Southeast Asia, and the glaciers feeding those rivers are melting at rates that will fundamentally alter flow patterns within decades.
The Indus Waters Treaty between India and Pakistan, signed in 1960, has survived multiple wars—but India reportedly placed it in abeyance in May 2025, and the Ganges Treaty with Bangladesh expires in 2026. Both instruments were designed for hydrological conditions that climate change is rendering obsolete. Fixed allocation quotas don’t work when the total volume of water in the system is declining.
The desalination solution and its limits
Desalination is the technology that allows countries without rivers to exist at modern scale. Saudi Arabia has invested at least $53.4 billion in desalination infrastructure since 2006 and plans to invest roughly $80 billion more. Eight of the ten largest desalination plants in the world are on the Arabian Peninsula. The Ras al-Khair plant in Saudi Arabia produces roughly 264 million gallons per day. These facilities are engineering marvels that convert seawater into potable water through reverse osmosis or thermal distillation, enabling cities like Dubai, Doha, and Kuwait City to support populations that the natural water supply couldn’t sustain at any scale.
The limitation is that desalination plants are stationary, energy-intensive, and targetable. More than 90 percent of the Gulf’s desalinated water comes from just 56 plants. During Iraq’s 1990 invasion of Kuwait, Saddam Hussein’s forces released hundreds of millions of barrels of oil into the Persian Gulf, contaminating the seawater that desalination plants depend on. Kuwait had to import water by tanker. In the current conflict, Iranian strikes on March 2 hit Dubai’s Jebel Ali port roughly 12 miles from a complex with 43 desalination units. Debris from intercepted missiles reportedly damaged facilities in Kuwait and the UAE. The Hudson Institute’s assessment is blunt: unlike disruptions to oil markets, which primarily trigger economic consequences, striking desalination facilities “directly threatens daily survival.”
The Gulf states have built contingency infrastructure—pipeline networks, storage reservoirs, protective barriers for intake valves. The UAE maintains 45 days of water storage under its 2036 water security strategy. Saudi Arabia has geographic depth and Red Sea facilities that provide resilience. But Qatar, Bahrain, and Kuwait have minimal strategic reserves and near-total dependence on Gulf-shore plants within range of Iranian missiles. If Iran were to systematically target desalination infrastructure—which it has threatened but not yet executed—millions of people would face acute water crisis within weeks.
Desalination as a moonshot technology
The vulnerability exposed by the Iran war is also a technology problem with a technology roadmap. Current desalination is expensive—roughly $0.50 to $1.50 per cubic meter depending on the technology and energy source—and energy-intensive enough that the plants themselves are tethered to fossil fuel infrastructure, creating a circular dependency: oil powers the machines that make water that supports the populations that produce the oil.
Next-generation desalination aims to break that loop. Solar-powered reverse osmosis plants, already operational in small deployments in the Middle East and North Africa, decouple water production from fossil fuels. Forward osmosis, membrane distillation, and capacitive deionization offer potential efficiency improvements over conventional reverse osmosis. The broader moonshot vision—desalination powered entirely by renewable energy, at costs low enough for agricultural irrigation rather than just municipal drinking water, deployable at scales that could make arid regions self-sufficient in freshwater—would fundamentally alter the geopolitics of water by removing the scarcity that drives conflict. Studies project a potential 40 percent global shortfall in freshwater resources by 2030 while demand increases by more than 20 percent. Desalination at scale isn’t optional for the species. It’s the engineering requirement for sustaining 10 billion people on a planet where freshwater distribution doesn’t match population distribution.
What the map actually shows
The geopolitical map of water in 2026 has three layers. The first is the ancient layer: rivers that cross borders, with upstream countries holding structural power over downstream countries—Ethiopia over Egypt, Turkey over Iraq, China over Southeast Asia, India over Pakistan and Bangladesh. These conflicts predate the modern era and will outlast it.
The second is the industrial layer: desalination plants that allow countries without rivers to function as modern states, concentrated in the Gulf and now exposed as military targets in a way that their designers never intended and their populations are only now confronting. A technology that was supposed to solve water scarcity has created a new vulnerability—centralized, targetable, dependent on energy infrastructure that is itself a target.
The third is the technology layer: the moonshot question of whether desalination can become cheap, renewable, distributed, and resilient enough to decouple water supply from both geography and geopolitics. That’s a decades-long engineering problem, not a policy fix, and it belongs in the same category as fusion energy and space-based solar power—transformative if achieved, speculative on timeline.
The common thread across all three layers is the same insight: water is not a commodity. It’s a strategic resource whose control determines which populations survive, which economies function, and which governments maintain legitimacy. Oil made the Gulf rich. Water keeps it alive. The Iran war is making that distinction impossible to ignore.
We cover water geopolitics alongside the Darién Gap, forbidden zones, and the hidden geography that shapes the modern world across our Off The Map course. We also cover next-generation desalination as a civilization-scale engineering challenge across our Moonshot 2169 course—including why the most important technology for the next century might not be AI or fusion. It might be a cheaper way to remove salt from seawater.
-
The Global Helium Shortage: Why a Party Balloon Gas Is a National Security Concern
In March 2026, Iran struck Qatar’s largest liquefied natural gas facility. The damage knocked helium production lines offline—lines that could take years to rebuild. Qatar produces roughly one-third of the world’s helium supply, approximately 63 million cubic meters out of a global total of 190 million in 2025. That output is now functionally zero. About 200 specialized containers used to transport liquid helium are stranded near the Strait of Hormuz. The World Economic Forum estimates that conflict-related disruptions have removed approximately one-third of the global helium supply from the market. Spot prices have doubled since the war began. QatarEnergy issued a force majeure declaration on March 4, 2026, triggering cascading contractual mechanisms across every industry that depends on a gas most people associate with birthday balloons.
Helium is not a rare earth element. It’s the second most abundant element in the universe. It is, however, vanishingly scarce on Earth in usable concentrations, impossible to synthesize economically, and—unlike every other industrial gas—cannot be recaptured once it escapes into the atmosphere. It floats up and is gone. Every cubic meter of helium vented, leaked, or released from a party balloon is helium that the planet’s industrial base will never use again. The global economy runs on a nonrenewable gas with no substitute for its most critical applications, produced as a byproduct of natural gas processing in a handful of countries, and one-third of that supply just went offline because of a conflict that has nothing to do with helium.
What helium actually does
The party balloon market accounts for a negligible fraction of global helium consumption. The applications that matter are the ones where no alternative exists.
MRI machines require approximately 1,500 to 2,000 liters of liquid helium to cool their superconducting magnets to operating temperature—near absolute zero. There are roughly 40,000 to 50,000 MRI scanners installed worldwide, each requiring refills every two to six weeks. Healthcare accounts for roughly 32 percent of global helium consumption. When helium runs short, hospitals delay installations of new MRI systems, and existing systems face refill scheduling constraints. Each nonfunctional MRI scanner eliminates approximately 20 to 30 daily patient examinations.
Semiconductor manufacturing accounts for 24 percent of global consumption in 2025, projected to reach 30 percent by 2030. Helium cools superconducting magnets during chip fabrication, flushes toxic residue after wafer washing, and supports leak detection in the vacuum systems that advanced lithography depends on. EUV lithography—the technology that makes sub-5-nanometer chips possible—has driven semiconductor helium demand from roughly 6 percent of global consumption in 2015 to 10 to 12 percent by 2025. With TSMC, Samsung, and Intel all building new fabs under the CHIPS Act and equivalent programs worldwide, and 42 new fabrication facilities scheduled to come online by 2026, semiconductor demand for helium is growing 15 to 20 percent annually. In 2024, Samsung’s Vietnam fabrication plant experienced a 72-hour outage from helium supply disruption, resulting in approximately $300 million in losses.
Aerospace consumes 18 percent of global demand. NASA’s Artemis program alone requires 3.2 million cubic feet per Space Launch System launch. Quantum computing requires helium-cooled cryogenic systems to maintain qubits at millikelvin temperatures. The International Energy Agency has warned that helium shortages could delay quantum computing adoption by two to three years. Defense applications—missile guidance systems, surveillance technologies, and components manufactured using helium-dependent processes—consume classified but significant volumes.
The CHIPS Act allocated approximately $2.1 billion specifically for helium infrastructure to support domestic semiconductor production. The Department of Defense has established a target of maintaining a six-month helium reserve by 2026, up from the 83-day reserve that existed before the current crisis. Twenty-two countries now require special licenses for helium exports, citing national security concerns.
Why supply is this fragile
Helium is produced almost entirely as a byproduct of natural gas processing. You don’t mine helium. You extract it from natural gas fields where it occurs in concentrations of 0.1 to 7 percent, separated during cryogenic processing of the primary product—LNG. This byproduct structure creates a fundamental vulnerability: helium production depends entirely on natural gas production decisions. When QatarEnergy halted LNG operations, helium supply ceased automatically—not because the helium market changed, but because the primary revenue driver went offline.
Three countries dominate supply. The United States has historically been the largest producer, anchored by the Federal Helium Reserve in Amarillo, Texas—a strategic stockpile that the U.S. government began building in the 1920s for military airships. Congress passed the Helium Privatization Act in 1996, directing the Bureau of Land Management to sell off the reserve and wind down government involvement in helium markets. That logic—reducing government involvement in commodity markets—made sense when helium’s primary applications were party balloons and weather balloons. It looks catastrophically shortsighted in 2026, when helium is a strategic material for semiconductors, quantum computing, MRI systems, and defense.
Qatar became the world’s second-largest producer and is now offline. Russia’s Amur Gas Processing Plant was supposed to change the math—potentially supplying 25 percent of global demand at full capacity. Gazprom started helium production there in 2021, but the facility has been hit by explosions, technical setbacks, and Western sanctions. As of early 2026, Amur is running well below capacity. Russia has increased helium exports to China—up 60 percent in 2025 alone—but the volumes remain far below what was planned. Algeria rounds out the major suppliers, but production there has been flat.
New projects in Saskatchewan, Tanzania, and South Africa are in various stages of development. None are close to meaningful output. Greenfield helium developments typically require 7 to 10 years from exploration to production. The supply that’s missing today won’t be replaced by new sources for the rest of the decade.
Who gets it when there isn’t enough
Helium allocation in a shortage follows a predictable hierarchy. Essential medical uses—MRI machines, NMR systems—receive the highest protection. Defense and space applications sit immediately below. Semiconductors are high-priority industrial users but rank below medical and defense in a severe allocation scenario. Lower-value and more substitutable uses—welding, leak detection in non-critical applications, party balloons—face the sharpest cuts first.
South Korea is under the greatest near-term strain. The country produces roughly two-thirds of the world’s memory chips and sourced 64.7 percent of its helium imports from Qatar in 2025. Samsung is the most exposed major chipmaker, with an estimated buffer of six to twelve weeks. Taiwan entered the crisis with better short-term cover—one major supplier maintained stockpiles in both Japan and the United States—but remains exposed to cost inflation if the market stays tight for months. Chipmakers can store about six weeks’ worth of supply in specialized cryogenic containers, and once insulation is depleted, the helium warms, expands into gas, and escapes. You can’t stockpile it the way you stockpile oil.
The semiconductor equipment industry has responded by accelerating helium recycling system development. Current technology recovers 60 to 80 percent of helium used in fabrication, at installation costs of $2 to $5 million per facility. Semiconductor fabs achieve recycling rates of 95 percent or higher for some applications. But recycling reduces consumption; it doesn’t eliminate the need for fresh supply. And MRI machines—the largest single consumer—recycle at 70 to 80 percent, significantly worse than semiconductor fabs.
The pattern
This is the fourth major helium shortage since 2006. Shortage 1.0 in 2006 to 2007. Shortage 2.0 in 2011 to 2013. Shortage 3.0 in 2018 to 2020. Each one driven by the same combination: plant outages, demand spikes, and the structural fragility of having a nonrenewable, non-substitutable industrial gas produced as a byproduct in a handful of geographically concentrated facilities. The 2026 crisis is different in scale—one-third of global supply offline due to military conflict rather than equipment failure—but the underlying vulnerability is identical.
Helium is the material that makes the gap between “critical resource” and “national security concern” visible. It’s not scarce in the way rare earths are scarce—controlled by one country through deliberate industrial policy. It’s scarce in a more fundamental way: the planet has a finite amount, it cannot be manufactured, it cannot be recaptured once released, and the applications that depend on it—medical imaging, advanced semiconductors, quantum computing, space launch, defense systems—are the applications that define whether a country can function at a 21st-century technological level. A gas that lifts party balloons is now determining whether Samsung can make memory chips and whether hospitals can run MRI machines. The constraint was always there. It took a war to make it visible.
We cover the helium shortage alongside neodymium supply chains, semiconductor geopolitics, and the full landscape of critical materials that underpin modern technology across our Rare Earth Elements course—including why the most strategically important substance in advanced manufacturing is lighter than air and impossible to get back once it floats away.
