Augmented Reality vs Virtual Reality: Differences, Uses and What Comes Next

Augmented reality (AR) puts digital information on top of the world you already see. Virtual reality (VR) throws that world away and hands you one built by a computer. That is the whole game in one line. In AR the real world stays in front of you and the device hangs data, labels or 3D models onto what is there. Walk down a street and an arrow appears on the tarmac. Point a phone at a machine and the repair steps float over the broken part. In VR nothing you look at exists physically. The system builds all of it.

Add or replace. That single design choice decides everything that follows: which technology fits your daily work, which one belongs in training, and why one lives in your pocket while the other needs a dedicated device on your head.

Picture a maintenance technician at a water plant. She walks up to a pump, raises a tablet, and the pipe layout, the last inspection date and a numbered repair sequence sit right on top of the metal. She never leaves the room, never opens a manual, never guesses which valve is which. Now picture a surgical resident minutes before his first attempt at a rare procedure. There is no pump in front of him and there should not be. He puts on a headset and rehearses the whole thing in a synthetic operating room, fails, resets and repeats until his hands stop shaking. Same family of technology, opposite jobs. One adds. The other replaces.

The border between the two is soft. Picture a spectrum. Paul Milgram and Fumio Kishino mapped it in their 1994 taxonomy, the reality–virtuality continuum, and it is still the reference for ordering ideas such as mixed reality. This piece walks through what each technology is, how it works under the hood, how they differ, what they are used for by sector, what hardware you can buy today, which limits and health effects matter and where both are heading.

What is augmented reality

Augmented reality is the technology that integrates computer-generated elements into a real scene, in real time and in a way that matches the physical space. The world stays in front of you. The device enriches it. It captures its surroundings with cameras and sensors, interprets what is there (walls, tables, floors, objects) and places content that seems to occupy a specific spot. A navigation arrow on the tarmac. A product label hovering over a shelf. Maintenance instructions pinned to the faulty machine.

The classic definition belongs to Ronald T. Azuma. In his paper A Survey of Augmented Reality (1997) he set three conditions that separate augmented reality from every other system. It combines real and virtual objects, it is interactive in real time and it is registered in 3D, which means the virtual content stays aligned with the physical world. Miss any of the three and it is not augmented reality in the strict sense. A graphic pasted over a camera feed with no spatial anchor is a simple overlay. A 3D object that ignores what you do is not interactive.

Azuma’s three traits, explained

• It combines the real and the virtual. The scene you perceive contains physical things and computer-generated things at the same time. You are not in a separate world. You are in your own, with one added layer.
• Real-time interaction. When you move the device or your head, the overlay is recalculated instantly. Delay the update by a fraction of a second and the virtual object appears to float, peeling away from reality.
• Three-dimensional registration. Content is anchored to coordinates in the real world. It behaves like an object parked in a specific spot. Change position and it stays there.

The definition has survived almost three decades because it never depended on a particular technology. It depends on the relationship between the physical and the digital, and that relationship is the same whether you are holding a phone or wearing a headset. The gap between a simple overlay and true spatial anchoring is the gap between a gimmick and a tool, and it runs through every augmented reality application you will meet.

You have almost certainly used it without noticing the label. Point Google Lens at a sign to translate it. Hold your phone up to see how a sofa would look in your living room. Follow the live arrow in a maps app that paints the route onto the street through the camera. All of that is augmented reality working exactly as Azuma described. What those examples share is that the phone stays a window onto the real world. The moment the screen stops showing your surroundings and shows only synthetic content, you have crossed the line into virtual reality. We go deeper into the overlay case in this guide to augmented reality.

What is virtual reality

Virtual reality is the technology that creates a fully synthetic environment and puts the user inside it, replacing the stimuli of the physical world with those a computer produces. The physical world gets swapped out entirely. To pull that off, stereoscopic displays sit very close to the eyes, head and hand movement is tracked precisely, sound is placed in space and, more and more, the system adds touch feedback.

Seeing something in 3D is only the beginning. The real target is a sense of presence, the feeling that the brain reacts as if you were really there. Presence is what makes a simulation of heights produce vertigo or an emergency drill generate genuine tension. It depends on many factors (field of view, latency, resolution, sensory consistency) and it is a central concept in academic research on virtual environments.

Immersion, presence and the role of hardware

Immersion is an objective property of the system, meaning how much it isolates you and how many senses it stimulates. Presence is the user’s subjective experience. A headset with a wide field of view, good tracking and low delay produces more immersion, and more immersion raises the odds of presence. The content matters just as much as the device, though. Without a well-designed virtual environment, the best hardware on the market is a technical demo.

In virtual reality, in fact, immersive content is as important as the device. A headset is a door. What lies on the other side is decided by software. That is why the field brings together hardware makers, content studios, rendering engines and interoperability standards.

The centre of gravity for VR is training and simulation, where the value is repetition without consequence. Airline and military pilot training moved into simulators decades ago for the same reason a surgical team rehearses a rare procedure in virtual space. A mistake in the real world is unacceptable. A consumer may buy a headset for games, but the organisations that adopt VR at scale usually do so to practise something dangerous, expensive or impossible to stage on demand.

What mixed reality is and the reality–virtuality continuum

Mixed reality (MR) is the point on the spectrum where the digital merges with the geometry of the environment. Virtual objects hide behind real ones, rest on surfaces, cast shadows and respond to the space. To organise these shades, Milgram and Kishino proposed the reality–virtuality continuum in 1994, a line running from one fully real extreme to a fully virtual one.

On that line, every intermediate point is mixed reality. Augmented reality sits near the real end, where the physical world dominates and the virtual is added. Augmented virtuality sits near the virtual end, a synthetic environment into which real elements are inserted, for example a video call inside a virtual room. The border between “augmented” and “mixed” is fuzzy, and companies use the words differently. Microsoft, for instance, groups under “mixed reality” much of what others call augmented reality.

A 2021 academic review in Frontiers in Virtual Reality revisited the Milgram and Kishino model to clarify how the term “mixed reality” is used today and why it is better not to tie it to one specific technology. The practical conclusion is simple. Arguing over labels is a waste of time. What matters is how much of the real world remains and how much is replaced.

You will also run into the umbrella term extended reality (XR), which covers AR, VR and MR together and is the word most used in the industry and in standards work. It is a useful shorthand precisely because the borders are blurred. A modern headset can switch from a fully virtual scene to a passthrough view of your room in a single tap, and the same application may be tagged AR, MR or XR depending on who is selling it.

Comparison table: AR, VR and MR

The quickest way to see the differences is to line them up side by side. These are the key points:

Criterion	Augmented reality (AR)	Virtual reality (VR)	Mixed reality (MR)
Real world	Kept as the base	Replaced completely	Kept and merged with the virtual
What is added	Layers of information and virtual objects	The whole environment is synthetic	Virtual objects that interact with real geometry
Immersion	Partial and open	Total and closed	Variable, depending on the degree of integration
Typical device	Phone, tablet or see-through glasses	Opaque headset (HMD)	Passthrough headset or see-through glasses
Tracking	SLAM, markers, GPS	6DoF, controllers and hands	Spatial SLAM with occlusion and mapping
Dominant use	Assisted work, navigation, commerce	Training, gaming, simulation	Design, collaboration, industry
Isolation from surroundings	None or low	High	Medium, controllable
Space requirement	Low	Medium or high	Medium

Read the table as a spectrum. Nothing in it is a sealed compartment. A virtual reality headset that shows the outside on video, known as passthrough, moves toward mixed reality. So do augmented reality glasses with occlusion and advanced spatial anchoring. What counts is the experience you design, not the label on the box.

Key differences between augmented reality and virtual reality

Beyond the table there are deeper differences that decide which technology makes sense in each case. These are the ones that weigh most when you have to choose.

The real world: foundation or replaced

In augmented reality the physical world is the main stage and the digital is an addition. In virtual reality the physical world disappears from your perception and the digital is the entire stage. That difference shapes everything else. It shapes safety, because in one case you can see what surrounds you. It shapes mobility, because you can walk through a workshop or you are confined to a clear space. And it shapes the kind of task, whether you assist in the real world or simulate one that is not there.

Immersion and presence

Virtual reality aims to isolate and submerge. Augmented reality aims to inform without disconnecting. That is why VR needs headsets that cover the eyes, while AR gets by, in its most basic form, with the camera on a phone. Intense presence is the goal in VR. In AR it is a welcome side effect.

Hardware and cost

Accessible AR runs today on phones and tablets that most people already own. VR requires, at minimum, a headset. That entry barrier shapes adoption. Millions of people have used augmented reality without realising it, through filters, navigation and virtual try-ons. VR is still tied to buying a specific device. The cost gap narrows as standalone headsets get cheaper, but the habit of carrying a phone in your pocket is hard to beat.

Tracking and spatial anchoring

AR has to recognise the environment and keep virtual objects anchored to it as you move. VR has to track your body inside a reference space. Both share techniques such as computer vision and inertial sensors. The difference is what they fight against. AR copes with the variability of the real world, from changing light to uneven surfaces. VR controls a known environment. AR’s problem is uncertainty. VR’s problem is precision at speed.

Interaction

In AR the natural interaction is touching the screen, pointing or using voice commands, because your hands stay free for the real work. In VR, interaction is designed inside the virtual world: tracked controllers, bare hands, gaze and gestures. Designing for AR means designing for someone who is also doing something else. Designing for VR means designing for someone who is entirely inside.

Comfort, hygiene and session length

A headset is worn on the face, which turns comfort into a specification. Weight and how it is distributed decide how long you can keep it on before your neck complains. The materials of the face gasket decide how much heat builds up. The lenses decide how quickly your eyes dry out. In shared or professional settings, hygiene matters too. Foam gaskets touch the skin and need cleaning between users, which is why many deployments stock disposable covers. AR on a phone avoids most of this at the price of holding your arm up. AR on glasses brings the hygiene question back but leaves your hands free.

Network, latency and bandwidth

Both technologies are sensitive to delay, but in different ways. In VR, too much latency between your movement and what you see causes nausea. In AR, a delay breaks the anchor and makes the virtual content drift away from the real. Cloud-based AR and streaming VR depend on fast, stable networks, a mobility challenge that connects to the quality of network infrastructures. You cannot hide a bad connection behind good graphics.

Where people get them confused

Confusion is common. Many products advertised as “virtual reality” use cameras to show the outside and edge toward mixed reality. Many “augmented reality glasses” simply project a screen in front of the eyes without anchoring anything to the space. The practical test is to ask yourself one question. Is the real world still the base and the digital anchored to it, or has the digital replaced it? Your answer tells you which part of the continuum you are standing on.

A short history: from early headsets to light glasses

Knowing where these technologies came from helps explain why AR and VR look so alike and yet answer to different ideas. Virtual reality was born as a simulator. Augmented reality was born as an assistant for real tasks. That origin still shows in how each one is used, and it makes more sense inside the wider history of computing.

The roots of virtual reality

Long before digital displays existed, people were already experimenting with the idea of immersing a person in an artificial environment. In 1962 Morton Heilig presented the Sensorama, a mechanical device that reproduced images, sound, vibration and even smells to simulate a motorcycle ride. It was a multisensory experience built before digital electronics. A little later Ivan Sutherland laid the foundations of a computer display system and, in 1968, built one of the first virtual reality headsets, the so-called Sword of Damocles, so heavy it had to be hung from the ceiling. That device showed simple computer-generated scenes and already posed the three keys of VR: isolation, freedom of perspective and immersive experience.

Through the 1980s and 1990s the video game industry explored the ground. Sega and Nintendo tested virtual reality systems. In 2009 Palmer Luckey’s Oculus Rift project revived interest. In the 2010s, Oculus and HTC Vive brought quality headsets with powerful graphics and motion tracking to market, and the sector professionalised. More recently, Meta with its Quest family and Apple with Vision Pro have put virtual and mixed reality back at the centre of the technology conversation.

The origins of augmented reality

Augmented reality has its own genealogy. In 1974 Myron Krueger created the Videoplace laboratory at the University of Connecticut, devoted to “artificial reality”, where he combined cameras and projection so users could interact with projected digital silhouettes. In 1990 Boeing researcher Tom Caudell coined the term “augmented reality” in the context of industrial assembly. In 1992 Louis Rosenberg developed “Virtual Fixtures” at the United States Air Force laboratory, one of the first fully working systems, which let military personnel guide machinery virtually.

In 1994 augmented reality reached the stage with the play Dancing in Cyberspace. In 1998 it arrived in sports broadcasts with the NFL’s yellow first-down line, one of the most recognisable uses of graphic overlay. Also in 1999 NASA used augmented synthetic vision for navigation of its X-38 vehicle, and the open-source ARToolKit gave anyone the tools to build marker-based experiences without a research lab. In the 21st century, Google Glass brought connected glasses closer to the general public, Microsoft presented HoloLens as a wearable computer, and apps such as IKEA Place popularised previewing furniture on a phone. Each milestone added a piece: capturing the environment, tracking, 3D anchoring, mobility and, finally, comfort of use.

That double history explains the underlying difference. VR has always wanted to build worlds. AR has always wanted to improve the one that already exists.

How each one works under the hood

Beneath the labels there is a set of shared technologies. Understanding how they work helps you read specifications, compare devices and anticipate limits before you spend a cent.

Tracking and positioning

Tracking is the heart of both. In VR the position and orientation of the head are measured with six degrees of freedom (6DoF), three of rotation and three of translation, so the virtual world responds to your movements. In AR you also have to place the device relative to the environment. That is where SLAM comes in, from Simultaneous Localization and Mapping. The system builds a map of the space while working out where it is inside that map.

To achieve this, cameras are combined with inertial sensors such as accelerometers and gyroscopes, in what is called visual-inertial odometry. A simple case of VR without translation is 3DoF, rotation only, typical of basic mobile headsets. You can look around, but if you lean or step forward, the scene does not change. The difference between 3DoF and 6DoF is one of the factors that most affects the sense of realism and the risk of motion sickness.

Degrees of freedom and why they affect comfort

The number of degrees of freedom is a shorthand for how much the system can follow you. 3DoF tracks only rotation (pitch, yaw and roll), so it knows where you are looking but not where you are. 6DoF adds the three translations along the axes, so leaning forward actually moves the viewpoint. The practical consequence is comfort. With 3DoF your inner ear feels movement the image denies, and that mismatch is a common source of nausea. With 6DoF and low latency, the image follows the body closely enough that the brain stops noticing the trick. That is why 6DoF has become the baseline for serious VR.

Anchoring: markers, environment and geolocation

Augmented reality places content in several ways:

• Marker-based. A visual pattern, a code or an image, acts as a reference and the content is placed on top of it. It is precise and cheap, but it depends on the marker being visible.
• Markerless. The system recognises the environment or surfaces by their features and anchors objects without needing a printed pattern. This is the basis of ARKit and ARCore.
• Location-based. GPS and other sensors place the user in the world and put content at geographic coordinates. Useful outdoors, with limited precision.
• Projection-based. Light is projected onto real surfaces and interaction with that projected light is detected.

Types of sensors

Tracking leans on several kinds of sensor worth telling apart. Positioning and tracking sensors locate the real world and the device: computer-vision cameras, marks or references in the environment, GPS outdoors, and gyroscopes, accelerometers and compasses for orientation, direction and acceleration. Many of them are already built into phones and tablets. Environment sensors measure humidity, temperature or other variables, useful in industrial and scientific applications. User input sensors capture direct interaction: buttons, touchscreens, controllers or gesture sensors. The combination of all of them is what lets the real and virtual worlds stay in sync.

Rendering and perception

Virtual reality generates two slightly different images, one per eye, to produce stereoscopic depth. Because the displays sit very close, lenses are used to allow focusing. The interpupillary distance (IPD) adjustment matches the separation of the pupils to each person to avoid fatigue and double vision.

Here one of the classic problems appears, the vergence–accommodation conflict. The eyes adjust focus according to distance, which is accommodation, and turn to point at the same spot, which is vergence. In a stereoscopic headset, vergence indicates a depth but focus stays fixed on the display, and that mismatch can cause eye strain. Research on vergence–accommodation conflicts in head-mounted displays describes this phenomenon and its effects on visual fatigue. Some advanced solutions, such as light-field displays and variable focus, aim to reduce the mismatch.

To lower perceived latency, engineers use techniques such as reprojection, which reuses the last frame corrected for the most recent head movement. Increasingly they also use foveated rendering, drawing the area the eye is looking at in more detail and the periphery in less, taking advantage of eye tracking.

How the lenses shape what you see

The screen sits a few centimetres from your eyes, far closer than they can focus on, so lenses bend the light to make the image appear at a comfortable distance. Older designs used heavy glass. Modern headsets use Fresnel lenses, concentric rings that bend light with less material, or pancake lenses, folded optics that are thinner and sharper but waste more light. Each choice trades off brightness, bulk, clarity and cost. That is why two headsets with identical displays can look very different through the eyepieces.

Interaction: hands, gaze, voice and haptics

Modern systems combine several forms of interaction. Hand tracking lets you manipulate objects without controllers. Eye tracking uses gaze to select and improves foveated rendering. Controllers remain the most precise option for many tasks. Voice is useful for quick commands. Touch feedback, or haptics, adds a sense of contact through vibration or resistance, with dedicated gloves and vests.

Spatial audio matters as much as the image. Sounds are placed in space with binaural techniques, the head-related transfer functions known as HRTF, so you can hear where they come from. Good spatial audio increases the sense of immersion and helps you orient yourself, often at lower cost than improving the graphics.

Where the scene is computed

There are three main computing models. A standalone headset carries its own processor and needs no PC. It is convenient and the most popular for consumers. The PC- or console-tethered model offloads the graphics work to a powerful machine and achieves higher quality at the cost of cables. Cloud XR renders on servers and sends the image over the network, which reduces local hardware but demands very low-latency connections. This is where network infrastructure becomes decisive.

Devices and hardware

Hardware is where the distance between the two technologies is clearest and, at the same time, where they converge.

Virtual reality headsets

VR headsets cover the eyes and rely on high-resolution displays, lenses and tracking. They come as standalone units, such as the Meta Quest family, as PC-tethered models such as Valve Index and HTC headsets, and as console units such as PlayStation VR2. Standalone headsets have been the adoption lever because they do not require a powerful computer. Meta publishes its platform and developer tools in its Horizon OS documentation, and Apple entered the segment with Vision Pro and visionOS, which bets on passthrough mixing and control by eyes and hands.

Augmented reality glasses and smart glasses

The goal of AR is a pair of light glasses, resembling ordinary spectacles, that show information without isolating you. The usual optical technologies are waveguides, which guide the image toward the eye through the lens, and birdbath systems, bulkier but with better quality. The big challenge is the balance between field of view, brightness, weight and battery life. Manufacturers such as Microsoft (HoloLens), Magic Leap, Vuzix, Xreal and Rokid explore this ground. Meta showed its Orion prototype in 2024 as a step toward “real” augmented reality glasses. Apple has brought passthrough to a headset and is building out its ecosystem around visionOS.

Phones and tablets

The phone remains the most widespread augmented reality device. Apple offers ARKit on iOS and iPadOS, and Google offers ARCore on Android. Both let you recognise surfaces, estimate the device’s position and place virtual objects without extra hardware. The LiDAR sensor, present on some iPhone and iPad Pro models, improves depth and environmental mapping. Apps for measuring distances, previewing furniture or translating signs support daily use.

For AR to work on a phone, three things are needed. A system to capture the real image, which is the camera. Software that interprets that image in real time and generates the virtual elements. And a way to place them correctly. Geolocation provides the context of where you are, while features based on environment recognition let objects be anchored without marks. On top of that, many apps use the cloud to download models or sync content between devices.

The advantage of the phone is twofold. It reaches far more people than a headset, and it accompanies the user in the real world, which is exactly where AR adds value. Its limits are just as clear. The small screen forces you to hold the phone in front of you, which is tiring, and the experience is not hands-free as it is with glasses. That is why the phone has been the gateway and glasses the intended destination.

Compared with the phone, VR needs a headset. There is no equivalent “pocket virtual reality”, apart from basic cardboard solutions that are now outdated. That asymmetry explains why AR has slipped into so many applications without the user noticing.

Accessories

Around headsets an ecosystem of accessories has grown. Haptic gloves and vests simulate touch. Multidirectional treadmills let you “walk” without moving. Full-body sensors capture movement. There are specialised controllers and, of course, the connection cables, where USB-C has become the norm on modern gear. These peripherals broaden the experience, though they tend to be expensive and niche.

Specs worth checking

• Field of view (FOV). The larger it is, the greater the sense of immersion. It determines how much of the virtual world your vision takes in.
• Resolution and pixels per degree (PPD). A sharp image avoids the “screen-door effect” and reduces visual effort.
• Refresh rate. Higher rates improve smoothness and reduce motion sickness.
• Motion-to-photon latency. The time between moving your head and seeing the result. The lower, the better.
• IPD adjustment. Key for comfort and image quality.
• Weight and balance. They determine how long you can wear the device without fatigue.
• Battery life. The Achilles heel of standalone headsets.
• Tracking. 6DoF, hands, eyes and full body depending on the use.

Two of those specs are worth interpreting rather than memorising. Pixels per degree matters more than raw resolution, because what your eye notices is angular sharpness. A wide field of view with the same panel spread thinner looks less crisp. Motion-to-photon latency is the number that hides behind “refresh rate” marketing. The shorter it is, the less the world seems to lag behind your head and the fewer people feel discomfort. The rest, from IPD range to weight and battery, decides whether the device is still pleasant to use after the first ten minutes.

Applications by sector

The question “AR or VR?” almost never has a single answer. It depends on the task. Here are the sectors where both are already used in a realistic way.

Education and training

Augmented reality lets you overlay information on real materials. A biology student sees labels and data on what they observe under the microscope or during a dissection. An apprentice mechanic follows instructions on the part in front of them. Virtual reality shines in simulating situations that are hard to reproduce or dangerous: lab practice, emergency procedures or trade training without risk. VR’s advantage is safe repetition. AR’s advantage is support at the real time and place.

Industry and maintenance

On the factory floor, AR gives the technician the status of a machine, repair steps or the location of faults overlaid on the equipment. VR is used to design products, simulate production lines and train operators before they touch real equipment. Both reduce errors and downtime, but at different points in the process. VR comes before. AR comes during.

Healthcare

Virtual reality is used in medical training, surgical planning, rehabilitation and therapy for phobias, anxiety or post-traumatic stress, as well as controlled exposure to difficult situations. Augmented reality helps in image-guided surgery, where anatomical data is overlaid on the surgical field, and in remote assistance. The trend is to combine both with cloud resources and clinical data, always with the regulatory and privacy cautions the health field demands.

Retail and commerce

AR lets you preview a piece of furniture at home, try on makeup or clothes virtually, or see how a product fits a real space before buying it. VR is used to walk through stores or configure products immersively. In both cases the goal is to reduce the uncertainty of the purchase and the number of returns. AR has the edge because it fits into the phone and into the customer’s real context.

Entertainment and media

VR has found its quarry in video games, with experiences designed for movement and immersion. It is also used in concerts, sports events and interactive storytelling. AR has been built into filters, mobile games, advertising and broadcasts, where it overlays graphics on the real image, from sports markers to live effects. Every major sporting event now carries some layer of virtual graphics that viewers take for granted.

Architecture, engineering and construction

Here the two technologies complement each other. VR lets you walk through a building before it is built and spot design problems. AR takes the drawings to the construction site, overlaying them on the real structure to check that everything fits. BIM modelling and digital twins rely on this kind of visualisation.

Defence, security and emergencies

Flight and driving simulators are a historic use of VR, because they reproduce dangerous situations without risk. AR feeds real-time information to firefighters, police or medical staff: building plans, the location of colleagues or sensor data overlaid on their vision. In both cases, reliability and latency are critical.

Remote work and collaboration

Mixed reality aims to let several people share the same workspace even when they are in different cities, with avatars and manipulable 3D objects. It is a recurring promise of the sector, improving what video calls fail to capture, such as spatiality and interaction with objects. Its adoption will depend on the hardware being comfortable and cheap.

Public sector, tourism and culture

Augmented reality has found a natural home in tourism and museums, where a phone can label a monument, rebuild a ruin on the spot or translate a plaque into your language. Virtual reality lets a museum offer a walk through a site that no longer exists or is too fragile to visit, from a Roman villa to a wreck on the seabed. Public services use similar tools for training, urban planning and public consultation, where showing a project in its real context removes misunderstanding long before the first brick is laid.

The influence of cinema and popular culture

Much of what the public expects from these technologies comes from cinema and fiction, and almost never from technical manuals. Films have shown floating interfaces, headsets that expand perception and virtual worlds that are hard to leave. That imaginary has shaped expectations, often beyond what real hardware allows. Separate the spectacle from what lies underneath.

Augmented reality appears in fictions where someone sees data overlaid on the real: layout and diagnostics interfaces in Iron Man, screens in the air in Minority Report, helmets that add layers of information in Robocop or glasses that reveal a truth hidden from others. Virtual reality stars in stories of total immersion, worlds where people live, work and play, such as the Oasis of Ready Player One, or simulations that blur the border with the real, as in The Matrix or Tron. War films also popularised elements that are real today, such as the information displays in a pilot’s helmet.

That imaginary has two effects. It attracts interest and funds research. It also generates unrealistic expectations, because people expect perfect holograms and total lightness when the technology still wrestles with field of view, weight and battery life. Cinema is a poor technical source, but a good clue as to where the sector wants to go and to the ethical debates (privacy, identity, dependence) that accompany every emerging technology.

Limitations and challenges

Neither AR nor VR is solved. These are the limitations that most condition their use today, and none of them is worth minimising.

Technical limitations

• Field of view. Headsets do not cover the whole human visual field, which reduces immersion.
• Resolution and sharpness. Even with high-density displays, the closeness to the eye demands many pixels, and optical limits still show.
• Weight and bulk. They feed directly into how long the device stays comfortable.
• Battery life. The battery limits the sessions of standalone headsets.
• Latency. The delay between movement and image remains a challenge, especially in streaming.
• Imperfect anchoring. In AR, content can float or drift if tracking fails because of poor light or uniform surfaces.
• Occlusion and physics. Without well-estimated environmental geometry, the virtual does not hide correctly behind the real.
• GPS indoors. Geolocation loses precision under a roof, which limits part of location-based AR.

Health and ergonomics

Prolonged use of these devices raises questions of wellbeing. The best known is cybersickness, with symptoms such as nausea, disorientation, sweating or fatigue, caused by discrepancies between what the eyes see and what the body and inner ear feel. Research on presence and cybersickness indicates that the two phenomena are related and that this is a central problem for VR adoption. AR, because it does not isolate you completely, tends to cause less of it, but it is not immune.

There is also visual fatigue, made worse by the vergence–accommodation conflict and by keeping the eyes forced. As with any screen, take regular breaks, adjust the IPD and the distance, and respect the manufacturer’s age recommendations, especially for children, whose visual system is still developing. In the workplace, ergonomics (posture, session length, hygiene of shared headsets) matters as much as the technology.

Privacy and data

An augmented reality device carries cameras and sensors that constantly capture what is around it. A VR headset can record gaze, movement and even biometric data. This raises privacy questions of its own: who is recording, what is processed on the device and what is sent to the cloud (which connects to the shared responsibility model of cloud computing), how the data of third parties who appear without giving consent is protected and what happens in workplace settings where the device can measure performance. In the European Union, the processing of this data is subject to the GDPR, and data protection authorities have stressed the importance of minimisation and transparency.

Accessibility

Accessibility is both a challenge and an opportunity. Many people with disabilities benefit from adaptable virtual environments, but design must take in captions, audio description, seated modes, alternatives to gestures and compatibility with assistive technology. The WCAG guidelines are a useful starting point even in immersive experiences, although they do not cover every nuance of 3D. A headset or an app that is not accessible leaves part of the population out.

Cost, content and the digital divide

VR hardware is still an investment, while AR leans on the phone. Even so, the digital divide, between those with access to devices, networks and training and those without, conditions adoption. Add to that the availability of quality content and the need for organisations to develop specific skills to make use of these technologies. A headset that nobody knows how to use is an expensive paperweight.

Safety when you are not fully present

A user inside a headset cannot see the room, and a user wearing AR glasses sees the room with their attention elsewhere. That raises practical safety questions. Where is the play area? Do cables run across the floor? Can the content be used seated? What happens if tracking fails mid-movement? On building sites, in workshops and in hospitals, the device has to be treated like any other piece of protective equipment, with clear rules, a marked area and a way to step out instantly. Comfort and safety are the same design question seen twice.

Standards and ecosystem

For these technologies to become widespread, software must be able to run on many devices without being rewritten. That is where standards come in:

• OpenXR (Khronos) defines a common API for accessing virtual and augmented reality hardware, so that one application can run on different headsets. It is the interoperability backbone of the sector.
• WebXR (W3C) lets you build immersive experiences that run in the browser, without installing apps. It includes modules for augmented reality and for inputs such as hands.
• ARKit and ARCore are Apple’s and Google’s frameworks for developing augmented reality on phones.
• Platform systems: visionOS at Apple, Horizon OS at Meta and Google’s work on Android XR. Microsoft offers its mixed reality platform with HoloLens.
• 3D content formats such as glTF and USDZ make it easier to move models between tools and devices.

Standardisation is what lets developers’ effort multiply. Write once and run in many places. Without it, each manufacturer would be trapped in its own ecosystem and adoption would stall.

How to decide: AR, VR or MR?

Start with the problem in front of you. The fashionable technology of the year has nothing to do with it. As a quick guide:

• If the user has to keep operating in the real world, whether repairing, serving, inspecting or buying, you probably want augmented reality.
• If you need to simulate something that does not exist, is dangerous or is expensive to reproduce, such as training, design testing or therapy, you probably want virtual reality.
• If the virtual must coexist with real geometry, in collaboration, design or assembly, you are looking for mixed reality.
• If you want the lowest cost and the largest user base, start with the phone and AR.
• If immersion and total control are essential, accept the cost of a headset.

Weigh each case on adaptability, productivity, time and return. And remember that sometimes the simplest answer wins. Good augmented reality on a phone can solve more than an expensive headset that nobody uses.

The future of augmented and virtual reality

The direction is clear: lighter, faster, more affordable and, above all, convergent. These are the trends that will define the next stretch.

Convergence toward a single pair of glasses

The desired horizon is a single device, looking like ordinary glasses, able to show augmented reality and switch to full immersion when needed. Today’s passthrough headsets already blend both worlds. Light AR glasses do not yet match their field of view. The evolution of waveguides, micro-OLED displays and low-power processors will determine how far off that is.

Artificial intelligence and computer vision

AI improves tracking, object and scene recognition, occlusion and content generation. It also enables more autonomous tools: identifying what you see, translating in real time, summarising or assisting. In virtual reality, it helps create more realistic avatars, voices and environments. AI is probably the biggest accelerator of these technologies in the coming years, which is why it pays to understand what an algorithm really is before expecting magic from it.

Networks, cloud and distributed computing

Cloud XR, split rendering (part in the headset, part on the server) and next-generation networks aim to reduce local hardware and improve quality. Fifth-generation mobile networks, defined by the ITU under the IMT-2020 framework and developed by initiatives such as 3GPP, point to more speed and lower latency, the necessary conditions for immersive experiences on the move. Without a good network, the cloud does not solve the problem.

Optics, rendering and sensors

Improvements in optics (waveguides, light-field displays), in rendering (foveated, neural, perceptual reconciliation) and in sensors (LiDAR, eye and body tracking) bring the image closer to what the human eye expects. The aim is to reduce the vergence–accommodation conflict and fatigue, an indispensable condition for prolonged use.

Eye tracking and gaze as an input

Eye tracking started as a way to save rendering power, but it is turning into a primary input. Look at a target to select it, then confirm with a pinch or a word. That feels closer to how we point in the real world than a laser-pointer controller, and it is faster in menus. It also allows attention-aware interfaces, content that reacts to what you look at, which raise the same privacy questions as any other gaze-logging system. Expect it to become standard in headsets before it becomes common in AR glasses, where the optics leave less room for the sensors.

New uses and maturity

As hardware improves, uses that are incipient today will open up: industrial digital twins, more realistic medical training, remote collaboration with presence, configurable commerce and remote technical assistance. Health, industry and education are the sectors with the most room to grow, because in them the value comes from measuring better, training without risk and cutting costs.

At the same time, questions of privacy, accessibility and technological dependence will have to be resolved. A technology that captures its surroundings and your gaze must come with clear guarantees about what data is collected and why. Real adoption will depend as much on usefulness as on trust.

The bottom line

Augmented reality and virtual reality occupy different spaces. One expands the world and the other replaces it. Once you understand the reality–virtuality continuum, from Azuma and Milgram to today’s headsets, the confusion disappears and the choice gets easier. AR fits where the user keeps working in the real world and needs information on top. VR fits where you need to simulate, train or experiment without risk. Mixed reality blurs the border when the virtual has to coexist with real geometry.

Making these technologies light, comfortable, affordable and respectful of health and privacy is the work that remains. When hardware stops getting in the way, the question stops being “AR or VR?” and becomes “what problem are we solving with them?”

Here is your move. Before you buy anything, write down the task and the environment where it happens. If the task happens in the real world and the person still has to keep their hands free, look at augmented reality first, because a phone you already own may be enough. If the task happens somewhere dangerous, expensive or impossible to stage on demand, look at virtual reality and budget for the headset. Then run a small pilot with one team, one task and one clear number to measure. The hardware will keep getting better every year. Your decision does not have to wait for it.