What Will Replace Smartphones

 The Smartphone Isn’t Dying. It’s Disappearing.

What comes after the device that became our camera, map, wallet, office, social life and remote control for the world? The answer may not be another object at all.

Woman wearing smart glasses and an AI earbud walking through a city while subtle augmented reality navigation and notifications appear in her field of view.
The smartphone may not disappear overnight. Instead, its screen, controls and intelligence could gradually migrate into glasses, earbuds and wearable interfaces around us.

The smartphone won because it absorbed almost everything that came before it. It replaced the pocket camera, MP3 player, paper map, alarm clock, calculator, boarding pass, much of the wallet and a surprising amount of the personal computer. That is why asking what will replace the smartphone is harder than asking what replaced the iPod. A successor has to do more than display messages. It has to inherit an entire digital life.

Yet the first signs of the next transition are already visible. In 2026, several of the world’s largest technology companies are building computers that sit on the face, listen through the ears, understand the surrounding world and accept input without a touchscreen. At the same time, AI is changing the logic of software itself: instead of opening an app and navigating menus, users can increasingly describe an outcome and let an agent decide which tools to use.

That shift matters because the most plausible future is not a dramatic moment when everyone throws away a phone. The phone may disappear in stages. First we stop taking it out. Then it becomes a background computer and identity token. Only later might some of its hardware functions migrate into glasses, watches, earbuds, clothing or other devices. The replacement for the smartphone may therefore be less like a new gadget and more like a new interface to the world.

There is a useful historical pattern here. New computing platforms rarely erase the old one in a single launch; they steal its most convenient jobs first. Smartphones did not kill the PC, and streaming did not make televisions vanish. The post-smartphone transition may actually reverse the smartphone’s original trick. Instead of one object absorbing the camera, map, music player and computer, the functions of the phone may be distributed again — across glasses, earbuds, a watch or wristband and a background compute device — while one software layer makes them feel like a single machine.

Before We Replace the Phone, What Actually Has to Be Replaced?

A modern phone performs several jobs at once: it is a screen, camera, microphone, secure identity device, payment instrument, communication terminal, navigation system, entertainment console, sensor package and network connection. It also has one underrated advantage: privacy. You can look down at a screen and type something without announcing it to the room.

Any serious successor has to solve most of those jobs without becoming heavier, hotter, more distracting or more socially awkward than the phone it is trying to replace. That is why many “phone killer” products have failed. A device can be futuristic and still lose if it makes a two-second task take ten seconds.

There are four particularly difficult engineering problems. The first is display: how do you show useful information without blocking normal vision? The second is input: if there is no touchscreen, how do you write, select, edit and control things quietly? The third is power: glasses and tiny wearables have little room for batteries and even less tolerance for heat. The fourth is trust: a device that can continuously see and hear the world is useful precisely because it is also capable of collecting extremely sensitive information.

There is a fifth problem that is easy to overlook: identity. A phone is not just a screen; it increasingly holds passkeys, payment credentials, biometric authentication, two-factor approvals, encrypted messages and the recovery path to other accounts. A true successor has to answer a mundane but critical question: how does a bank, employer, airline or messaging service know that the person behind a pair of glasses is really you? The secure identity layer may eventually move into a watch, ring, wristband, glasses or a small compute module, but it cannot simply disappear.

That is why the most believable post-phone system is distributed. One device may handle visual output, another private audio, another authentication and subtle input, while heavier computation is split between local silicon and the cloud. To the user it may feel like one computer. Physically, it could be several.

The Strongest Candidate Is Already on Our Faces: AI Glasses

Smart glasses have been promised for years, but the category is changing. The important transition is from “a camera you can wear” to a computer that can show information, understand context and accept natural input. Apple Vision Pro has already demonstrated what a mature spatial operating system can feel like, but its headset form factor also makes the central challenge obvious: a post-smartphone interface has to become dramatically lighter and more ordinary before most people will wear it all day.

Meta’s current display glasses are a useful snapshot of the direction. Meta Ray-Ban Display glasses combine cameras, microphones, speakers and an in-lens display with the Meta Neural Band, a wrist-worn surface-electromyography system that detects subtle muscle activity. By September 2026, Meta was adding hands-free navigation, mirrored notifications, calendar actions and an on-glasses app discovery layer, while expanding the product into more markets. These are not full AR glasses in the science-fiction sense, but they already move several common phone interactions into the user’s field of view.

Google is pushing the same idea through Android XR, but in stages. At Google I/O 2026, the company described two categories of intelligent eyewear: audio glasses and display glasses. The audio models from partners including Warby Parker and Gentle Monster were scheduled to arrive first in the fall, using Gemini for hands-free help, messaging, photography and other tasks; display glasses are the more ambitious visual layer that follows the same platform direction. Snap went further in June 2026, unveiling SPECS, standalone see-through AR glasses priced at $2,195 and expected to ship in the United States, United Kingdom and France in the fall.

The more ambitious version is Meta’s Orion prototype, which uses micro-LED projectors and silicon-carbide optics to create an approximately 70-degree field of view in a glasses form factor. Orion is not a consumer product, but it shows where companies want the platform to go: wide visual overlays, eye and hand tracking, spatial computing and AI that can use what the wearer sees as context.

This is the first genuinely convincing route away from the smartphone because glasses solve the “where is the screen?” problem without asking us to hold anything. Navigation can appear near the street. A translation can sit beside the person speaking. A recipe can remain in view while both hands are busy. A message can be glanced at without breaking a conversation to unlock a slab of glass.

The catch is that ordinary eyewear is an unforgiving design target. People will tolerate a heavy headset for a game or a flight. Apple’s current M5-powered Vision Pro shows how much processing, sensing and display technology can fit into a spatial computer, but it still uses an external battery and is designed as a headset rather than ordinary all-day eyewear. People will not necessarily tolerate thick, hot glasses for fourteen hours a day. Fashion, prescription support, battery life and social comfort matter as much as processor speed.

Snap’s 2026 SPECS make the trade-off unusually visible in numbers. The two sizes weigh 132 and 136 grams and are rated for up to four hours of mixed use, with a charging case extending total time away from an outlet. They are impressive as standalone AR glasses, yet those figures also explain why replacing an all-day phone is harder than demonstrating compelling AR. The optics, cameras and AI may be ready before the weight and battery disappear into something people forget they are wearing. Snap’s published specifications are less a failure than a snapshot of the remaining physics.

Commuter wearing lightweight AI smart glasses and a neural wristband using a subtle finger gesture to control navigation and messages on a train.
The next interface may require almost no visible interaction. Smart glasses could provide the display while wrist-based EMG sensors translate tiny finger movements into commands.

The Ear May Become the Invisible Half of the Interface

Glasses solve the visual problem, but they do not solve every interaction. The ear is arguably a better place for information that should be private, immediate and low-attention. Directions can be whispered instead of drawn. A message can be summarized without putting text in front of your eyes. An assistant can ask for confirmation quietly while you keep looking at the person in front of you.

Consumer earbuds are already moving in that direction. Apple’s AirPods 5, announced in September 2026, pair hands-free Siri AI with head-gesture responses and Live Translation. None of those features makes an earbud a phone replacement by itself. Together, however, they show how much of the phone’s conversational interface can migrate into a device with no screen at all.

Research on hearables goes further. A 2025 review in Wearable Technologies describes in- and around-ear systems that can combine audio with sensors for signals such as heart rate, oxygen saturation, temperature and even ear-EEG. That does not turn earbuds into mind readers. It does suggest that the ear can become both an output channel and a quiet sensor platform, giving a future assistant more context about the user and environment.

This is why “glasses versus phone” is probably the wrong contest. Glasses can provide sight and cameras; earbuds can provide private audio; a wrist device can provide discreet control and authentication; the AI can coordinate the whole system. The successor to the smartphone begins to look less like a product category and more like a personal network worn on the body.

The Real Revolution May Be AI, Not the Glasses

The smartphone is built around apps. You decide which app to open, find the right control, enter information and move the result somewhere else. AI agents offer a different model: state the goal, then let software choose the sequence.

This is the shift Next Horizon explored in Artificial Intelligence Explained: From Neural Networks to AI Agents: modern AI is increasingly being combined with tools, memory and multi-step action rather than used only as a text generator. The same trend appears in consumer services. In travel, for example, an assistant can increasingly move from “suggest a destination” toward searching live inventory, monitoring changes and eventually performing approved transactions — a transition discussed in The AI Travel Agent Is Coming.

If that model becomes reliable, a future user may not need fifty icons arranged in folders. “Find the photo Anna sent me last spring, brighten it, send it to Dad and ask whether he wants a print” could be one instruction. The operating system becomes less visible because the AI is doing the navigation.

The deeper consequence is that the app may stop being the unit people think about. Today, an airline, bank, taxi service or photo editor controls the interaction because you enter its interface. In an agentic system, the user may simply ask for an outcome and the assistant may call several services in the background. The visible “home screen” becomes less important; permissions, APIs and the agent’s judgment become more important.

That creates a new set of problems. If an assistant can book, buy, send, delete and sign in on your behalf, it needs more than intelligence: it needs a permission model. Low-risk actions can happen automatically, while expensive, irreversible or sensitive actions may require an explicit confirmation. Users also need to know which service the agent chose, why it chose it, what data left the device and how an action can be undone. The post-smartphone interface could be simpler on the surface while becoming much more complex underneath.

This also makes local processing unusually valuable. A wearable that sends every camera frame, voice fragment and location cue to a remote server may be powerful, but it creates latency, connectivity and privacy problems. Some perception, wake-word detection, identity checks and routine actions are likely to happen on-device, with larger models or demanding searches called only when necessary. The best successor may therefore depend as much on invisible computing architecture as on the thing sitting on your face.

This makes the 2013 film Her unusually relevant. Its most realistic prediction was not a sentient operating system. It was the disappearance of the screen as the center of computing. The protagonist talks to an assistant through an earpiece while moving through the world. We already have the microphones, wireless earbuds, speech recognition, generative models and cloud services needed for a crude version of that experience. What we do not yet have is an assistant reliable enough to control the rest of digital life with minimal supervision.

Why the first AI “phone killers” struggled

The failure of the Humane AI Pin is an important warning. Humane shut down the consumer service in February 2025, making most cloud-dependent features unusable. HP then acquired key Humane AI assets shortly afterward. Humane’s own shutdown FAQ is a reminder that removing the screen does not automatically improve the interface. Voice can be slower, less private and less precise than tapping. A wearable also has less room for battery, cooling and display hardware.

Rabbit’s r1 took a different path and remains on sale as a dedicated pocket AI device. But its existence also highlights the central question: if an AI companion still needs its own screen, camera, battery, cellular connection and pocket space, why not simply put that intelligence into the phone people already carry?

The lesson is not that AI hardware is doomed. It is that AI becomes most disruptive when it removes steps rather than merely moving those steps onto a different object.

Woman using smart glasses and an AI earbud while a digital assistant quietly coordinates navigation, messages, calendar events, transport and payments.
The biggest change may not be a new screen at all. An AI agent could replace much of today’s app-by-app interaction by understanding what we want and coordinating services on our behalf.

The Missing Keyboard May Be Your Wrist

A screen can be miniaturized. Input is harder. Speaking to glasses works well for many commands, but nobody wants to dictate every private message on a train. Waving both arms in the air like a science-fiction conductor is even less practical.

This is where surface electromyography, or sEMG, becomes interesting. A 2025 Nature study led by researchers at Meta Reality Labs demonstrated a wrist-worn interface that learns the electrical signals associated with hand and finger movements. The system was trained across large numbers of people and could support gesture control and even handwriting-like input. Nature Electronics highlighted results including continuous navigation, discrete gestures and handwriting at roughly 20.9 words per minute in closed-loop tests.

The interesting part is that this research path has already crossed into consumer hardware. Meta’s Ray-Ban Display and Neural Band use wrist-based EMG for subtle gesture input. The commercial interaction vocabulary is still much simpler than a research handwriting demo, but the connection matters: the same family of signals that can move a cursor or recognize a pinch could eventually provide a quiet text and control channel for glasses.

This is not mind reading. The band does not decode thoughts from the brain. It detects neuromuscular signals that appear when you intend to move your hand. But that distinction may not matter to everyday computing. If a tiny pinch, finger movement or “writing” gesture on your leg can scroll, select or enter text, then the touchscreen stops being the only convenient private input method.

In that sense, Minority Report may have predicted the wrong gesture interface. The future is unlikely to require large theatrical arm movements in front of a transparent wall. It may use movements so small that people around you barely notice them.

Smart Contact Lenses: The Perfect Form Factor With Brutal Engineering Problems

If glasses are the obvious replacement for a phone screen, contact lenses are the seductive version. They promise a display with no visible device at all. Science fiction loves this idea because the interface can become indistinguishable from vision itself.

Research on smart contact lenses is real, but the field is much closer to sensors and medical monitoring than to a hidden iPhone on the eye. A 2025 review in Microsystems & Nanoengineering describes progress in microfluidic contact lenses for diagnostics and drug delivery, while also tracing the broader evolution of lens-based wearable systems. Experimental designs have explored miniature electronics and displays, but a consumer AR lens has to solve power delivery, heat, wireless communication, optical alignment, oxygen permeability, durability and eye safety at the same time.

AR makes the problem harder than sensing. A diagnostic lens may only need to measure a signal and transmit small amounts of data. A visual interface must generate precisely aligned light, refresh fast enough to feel stable as the eye moves, communicate with other devices and do all of this without blocking oxygen, disturbing the tear film or creating unsafe heat. The engineering budget is measured in millimetres, milliwatts and biological tolerance rather than in the much more forgiving dimensions of a phone.

The eye is also a terrible place to hide sloppy engineering. A phone can become warm. A lens cannot. A phone can carry a large battery. A lens cannot. A phone can be taken off the charger and replaced after four years. A lens sits directly on sensitive tissue and has to remain comfortable, sterile and optically correct.

So the “Black Mirror contact lens” is not impossible in principle, but the display is only one part of the problem. The more realistic path is that smart lenses first become medical or sensing devices, while high-function AR stays in glasses for much longer.

That suggests a likely sequence: lenses become useful sensors before they become general-purpose computers. They may monitor the eye or body, provide limited cues or pair with external hardware long before they can reproduce the rich, bright, wide-field interface that science fiction places directly on the cornea.

Brain–Computer Interfaces: Eventually, But Probably Not for Your Group Chat

The most radical smartphone replacement is a direct brain–computer interface. Instead of typing, tapping or speaking, the user would simply intend a command. That is the point where computing stops being wearable and starts becoming integrated with the nervous system.

The science is advancing rapidly in medicine. In 2025, researchers reported an instantaneous voice-synthesis neuroprosthesis in Nature that decoded activity from 256 implanted microelectrodes in a man with ALS and generated expressive synthetic speech with real-time feedback. Non-invasive work is advancing too, but with very different limits. A November 2025 Nature Communications study trained models on EEG and MEG data from 723 participants reading or listening to millions of words across three languages. The authors reported progress in identifying individual words, while explicitly describing natural-language decoding from non-invasive recordings as an open challenge.

The distinction is important because “brain decoding” is easy to oversell. These experiments do not amount to a device that can listen to arbitrary private thoughts. They work with structured tasks, known experimental conditions, large training datasets and signals that are much cleaner in some recording technologies than others. In the 2025 non-invasive study, performance also improved when multiple responses could be averaged — useful scientifically, but very different from silently composing a message in real time while walking down a street.

That difference between a clinical breakthrough and a mass-market interface is crucial. Medical users may accept surgery because the benefit can be transformative. A healthy consumer is unlikely to accept brain surgery merely to answer messages faster. Non-invasive systems avoid that barrier, but the signals are noisier and lower-bandwidth.

So direct neural control belongs in the long-term future of personal computing, but it is unlikely to be the first technology that replaces ordinary smartphone interaction. The wrist, voice, gaze and cameras are much easier ways to capture intent.

Engineer wearing transparent AR glasses viewing a detailed three-dimensional virtual spacecraft component anchored above a physical worktable.
Future “holograms” may not actually float in the room. Near-eye displays can make virtual objects appear convincingly anchored in physical space while the light itself is generated inside the wearer’s glasses.

What About Holograms? Science Fiction Is Half Right

Iron Man and Minority Report trained audiences to expect free-floating screens that can be grabbed and rearranged in mid-air. The real research is moving toward something subtler: holographic and light-field images delivered through near-eye optics.

A 2024 Nature paper demonstrated full-colour 3D holographic augmented-reality displays using metasurface waveguides, while a 2025 Nature Photonics study explored synthetic-aperture waveguide holography for more compact mixed-reality displays. These techniques address a real problem in AR: current near-eye displays can create a mismatch between where the eyes converge and where they focus, contributing to visual discomfort.

The important word is prototype. The 2024 Nature system demonstrated a compact and sophisticated optical architecture, but its reported field of view was 11.7 degrees. The 2025 Nature Photonics work pushed other constraints such as the usable eyebox and realistic 3D cues. That is genuine progress, not a finished pair of everyday glasses. Research can solve one optical bottleneck while weight, brightness, manufacturing, eye tracking and power remain separate problems.

That research makes a JARVIS-like visual layer much more plausible than a room full of bright free-space holograms. The user may see a three-dimensional object floating above a table, but the light is likely being generated by the glasses and directed into the wearer’s eyes. People without the glasses may see nothing at all.

Science fiction may therefore get the experience right while getting the optics wrong: the wearer can perceive a convincing object occupying space, even if no luminous object is actually floating in the room.

Five Science-Fiction Futures, Tested Against Reality

Her — an AI companion in your ear

Technically plausible in pieces and probably one of the closest fictional interfaces to the next stage of computing. Speech, earbuds, contextual AI and cloud tools already exist. The missing ingredient is dependable agency: software that can act across your digital life without making costly mistakes or asking for constant confirmation.

Minority Report — spatial interfaces controlled by gestures

The display part is increasingly realistic through AR glasses and holographic waveguides. The famous large gestures are less likely to survive. Quiet microgestures, gaze, voice and wrist-based EMG are more efficient and socially acceptable.

Iron Man — JARVIS plus holographic overlays

Multimodal AI that hears, sees and speaks is already real. AR can place contextual graphics into a user’s field of view. What remains fictional is the seamless reliability, deep world model, unrestricted autonomy and effortless integration with every machine around the user.

Black Mirror — an interface inside the eye or brain

There are genuine research paths toward smart lenses and neural interfaces, but today’s medical, optical and power constraints are severe. The dystopian questions about consent, surveillance and memory may arrive before the hardware becomes invisible.

Star Trek — talk to the computer and let the environment respond

This may be the most important prediction of all. The “device” matters less than the fact that intelligence is ambient. A room, car, pair of glasses or earbud can all be front ends to the same assistant. The successor to the smartphone may therefore be an account and an AI identity that follows you across hardware rather than a single object you own.

The Technologies That Still Stand in the Way

The smartphone has had nearly two decades to become astonishingly optimized. Any replacement has to compete with mature cameras, bright displays, secure biometrics, app ecosystems, global payment systems, all-day batteries and social habits that are now deeply embedded in everyday life.

Battery and heat remain especially unforgiving in glasses. Displays need brightness outdoors, cameras and AI consume power, and processors generate heat next to the face. Some of the most advanced current systems still shift computation into a phone or pocket device because physics makes full independence expensive.

Visual comfort is another constraint. Research on next-generation AR optics repeatedly focuses on the vergence–accommodation conflict: the eye can be asked to converge on a virtual object at one apparent distance while focusing optically at another. A 2025 Light: Science & Applications paper explored multi-depth varifocal optics specifically to reduce this mismatch and the fatigue it can cause.

Privacy may be harder than optics. A phone camera points where the user chooses. Smart glasses can point wherever the user looks. That creates an obvious benefit for contextual AI and an equally obvious problem for everyone standing nearby. Meta, for example, says its supported glasses use a capture indicator and automatically disable the camera if that indicator is covered — a safeguard described in its 2026 privacy explainer. But social acceptance will depend on more than a single LED. People need to understand when they are being recorded, what is processed locally, what goes to the cloud and how long contextual data is retained.

Security becomes more consequential for the same reason. A compromised phone can expose messages and money; a compromised always-on wearable could also expose what the user sees, hears and asks an agent to do. A credible post-phone platform will need hardware-backed identity, granular permissions, visible recording states, audit trails for agent actions and recovery methods that do not depend on the very device that was lost or stolen.

Then there is attention. A device that can place information directly in front of the eyes can reduce phone distraction — or turn the entire world into a notification surface. Good AR may need to be defined as much by what it refuses to show as by what it can display.

There is also a social interface problem. When someone glances at a phone, other people can usually tell that their attention has shifted. With private overlays and audio prompts, attention becomes invisible. A future etiquette may need to answer surprisingly basic questions: Are you reading a message while looking at me? Is the camera active? Did an AI just suggest your reply? Technology can become physically unobtrusive while becoming socially harder to interpret.

So What Will Actually Replace the Smartphone?

Probably not one product. The more convincing answer is a stack of devices that divides the phone’s jobs between them.

In that stack, glasses are the most plausible visual surface and world-facing camera; earbuds provide private audio; a wrist device supplies quiet input, biometrics and perhaps authentication; a pocket device or local hub supplies battery, radios and compute; cloud systems handle the heaviest models and live services; and an AI agent becomes the layer that makes all of those pieces feel coherent. It is less a new phone than a distributed personal computer.

The first stage is already happening. The phone stays in the pocket while glasses, earbuds, watches and AI assistants take over short interactions. You ask for directions, capture a photo, translate a sentence or hear a message without touching the phone.

The next stage is more important. The phone becomes infrastructure: battery, connectivity, secure identity, storage and compute. You may still own it, but it behaves more like the hidden computer in a car than the object you stare at all day. Glasses become the display; a wristband, voice and gaze become input; an AI agent becomes the software interface.

Only after those pieces mature does a true post-smartphone world become plausible. Some users may carry a small compute puck rather than a phone. Others may rely on glasses with their own cellular connection. Contact lenses could eventually take over portions of the display. Clinical neural interfaces may become far more capable. But the central change is not the shape of the hardware. It is that computing moves from a rectangle we constantly consult to a layer that surrounds us.

The transition will probably be uneven because different jobs have different replacement thresholds. Voice can already handle timers, short replies and simple searches. Earbuds can already translate conversations. Glasses can already capture a first-person view and surface limited information. But editing a spreadsheet, reviewing a legal document, gaming, reading long text or recovering a locked account still favors a large, private and precise screen. The smartphone can lose half of our daily interactions and still remain indispensable for the difficult half.

The Phone May Vanish the Same Way the Camera Did

The camera did not disappear when smartphones arrived. Photography became more common than ever. What disappeared for most people was the need to carry a separate camera.

The same may happen to the smartphone. Messaging, maps, photos, payments and social media will not disappear. They may simply stop requiring a dedicated handheld screen. One day, “checking your phone” could sound as dated as “logging on to the internet.” You will still be connected. You just will not think of connection as a place you visit.

If that happens, the phone’s real successor will not be the device that reproduces every feature of a phone. It will be the system that makes the phone unnecessary often enough that carrying one starts to feel optional.

And that may be the clearest sign that the successor has arrived: not when a company launches a gadget called the phone killer, but when we stop noticing which device is doing the computing at all.

Smartphone lying beside a compact camera, MP3 player and paper map while a person wearing smart glasses leaves home with empty hands.
The camera, MP3 player and paper map did not really disappear — their functions were absorbed by the smartphone. One day the smartphone itself may undergo the same transition.

FAQ

Will smartphones disappear completely?

Probably not quickly. They are extremely efficient general-purpose computers. The more likely transition is that people use them less directly as other interfaces take over common tasks.

What is the most likely smartphone replacement?

AI glasses are currently the strongest candidate for the primary everyday interface because they can combine cameras, audio, contextual AI and glanceable displays while leaving the user’s hands free. They will likely work alongside other wearables rather than replace everything alone.

Will apps disappear too?

Probably not, but their interfaces may become less visible. An AI agent could call an app or service in the background while the user interacts mainly through voice, glasses or a compact visual surface. Complex creative and professional software will still benefit from dedicated interfaces.

Could contact lenses replace phone screens?

In principle, yes. In practice, power, heat, wireless communication, optics, safety and comfort make full AR contact lenses far harder than glasses. Medical sensing is a more mature near-term use case.

Will we control computers with our thoughts?

Brain–computer interfaces are already producing major medical breakthroughs, particularly for people with paralysis or speech loss. Consumer mind-controlled computing is much further away, especially if it must be non-invasive, reliable and safe.

Are holographic phone replacements realistic?

Near-eye holographic AR is a real research field and is becoming more capable. Free-floating, high-resolution holograms visible to everyone in a room are much less practical than images generated by glasses directly for the wearer.

Next Horizon Guides

Explore Topics

Explore More Topics

Go deeper with curated guides that connect our best reporting around a single subject.

Dive into our in-depth guides and discover more stories around the ideas shaping our world.

Comments