Augmented reality feels like a technology of the moment. Samsung just revealed its Galaxy Glasses. Snap is days away from a launch event. Apple has confirmed its first glasses for 2027. The industry has never moved faster.
But augmented reality did not arrive in 2026. Its roots go back more than half a century — to a laboratory at Harvard, a Boeing factory floor, a Californian research park, and a summer in 2016 when millions of people walked into parks at midnight to catch virtual creatures with their phones.
This is the full story.
1968 — The Sword of Damocles
The history of augmented reality begins with a machine so heavy it had to be suspended from the ceiling.
In 1968, Ivan Sutherland — a computer scientist at Harvard who had already invented the first computer graphics program — built what is widely considered the first head-mounted display system. The device was so unwieldy that it couldn’t be worn without mechanical support, which is how it earned its name: the Sword of Damocles.
The display showed simple wireframe graphics — basic geometric shapes rendered in real time and overlaid on the wearer’s view of the room. By today’s standards it was laughably primitive. In 1968, it was science fiction made physical. Sutherland had demonstrated that a computer could add things to your view of reality, in real time, that weren’t physically there.
That idea — unchanged in its essence — is what every AR product in 2026 is still built on.
1974-1990 — Myron Krueger and the Research Years
Through the 1970s and 1980s, AR developed largely in academic laboratories with little commercial interest. Myron Krueger coined the term “artificial reality” in 1974 and built immersive interactive environments he called “videoplace” — rooms where a camera tracked participants and projected their silhouettes into a shared virtual space. The work was conceptually ahead of its time and commercially invisible.
In 1980, Steve Mann — later dubbed the “father of wearable computing” — built an early wearable computer with a head-mounted display while still a teenager in Canada. Over the following decade he refined the device through multiple iterations, wearing successive versions almost continuously, prefiguring the quantified self movement and the wearable computing category by thirty years.
At NASA’s Ames Research Center in the late 1980s, researchers developed the Virtual Visual Environment Display — a helmet-mounted display for flight simulation that combined real-world imagery with computer-generated overlays. The project was practical rather than philosophical: pilots needed to see instrument data without looking away from the world in front of them.
1990 — The Term Is Coined
The phrase “augmented reality” appears to have been coined in 1990 by Tom Caudell, a researcher at Boeing. Caudell was working on a system that overlaid wiring diagrams onto physical boards to assist aircraft technicians — a heads-up display that showed workers exactly where each wire needed to go, in the correct position, without them needing to consult a printed manual.
The problem Caudell was solving — how to deliver the right information to a person in the right place at the right time, without interrupting what their hands were doing — remains the core problem AR is trying to solve today.
1992 — Louis Rosenberg and Virtual Fixtures
In 1992, Louis Rosenberg at the US Air Force Research Laboratory developed Virtual Fixtures — an AR system that overlaid virtual sensory information onto a real-world environment to enhance human performance of motor tasks. In plain English: the system helped users perform physical tasks more accurately by showing them where to move and what to do, overlaid on the real world.
Rosenberg’s work was among the first to demonstrate that AR could improve human performance in measurable, practical ways — not just impress researchers in a laboratory.
1994-1998 — Theatre, Sport and the First AR Definition
In 1994, Julie Martin created “Dancing in Cyberspace,” the first augmented reality theatre production, in which dancers and acrobats performed alongside virtual objects projected into the physical stage. The show demonstrated that AR had artistic and entertainment potential beyond industrial applications.
In 1998, American broadcaster ESPN used a virtual first-down line in NFL television coverage for the first time. Viewers could now see exactly where the offense needed to reach for a first down — information that existed in the broadcast but not in the stadium. It was AR for television, invisible as technology and immediately indispensable as a viewing aid. The yellow line has been in every NFL broadcast since.
In 1997, Ronald Azuma published what became the defining academic paper on augmented reality, establishing a framework with three characteristics that still define how the industry understands AR today: it combines real and virtual elements, it is interactive in real time, and it is registered in three dimensions.
1999 — ARToolKit
In 1999, Hirokazu Kato at the Nara Institute of Science and Technology released ARToolKit — an open-source library that made it possible for developers to create AR applications without building the underlying tracking systems from scratch. For the first time, a developer could take a standard camera, print a marker on paper, and have the camera recognise the marker and overlay a 3D object on top of it.
ARToolKit democratised AR development in the same way that WordPress would later democratise web publishing. Hundreds of researchers and developers used it to build early AR applications. Many of the ideas those applications explored — marker tracking, 3D object overlay, mobile AR — are now standard features in consumer devices.
2000s — AR Goes Mobile
The proliferation of smartphones in the late 2000s changed AR from a laboratory technology into something that could reach consumers. The combination of a camera, a GPS receiver, an accelerometer and a connected processor in a pocket-sized device created the preconditions for mobile AR.
In 2008, Wikitude launched the world’s first mobile AR application — a camera overlay that showed information about points of interest as a user pointed their phone at the world around them. The information appeared on screen, anchored to the physical location the camera was pointing at.
In 2009, the MIT Media Lab’s Sixth Sense project — developed by Pranav Mistry — demonstrated a wearable gesture interface that projected information from the internet onto physical surfaces and objects. A user could point at a product in a shop and see its price. Point at a newspaper and see a video clip. Hold up their hands to form a rectangle and take a photo. The demo went viral and showed millions of people, perhaps for the first time, what ubiquitous AR could look like in everyday life.
2012 — Google Glass
On June 27, 2012, Google announced Project Glass at Google I/O. The following day, Sergey Brin appeared on stage wearing a pair — and a demonstration involving skydivers rappelling down the outside of the building, live-streaming from Glass as they descended, gave a global audience its first clear picture of what AR eyewear might look like in practice.
Google Glass was available to developers for $1,500 from 2013 and released more broadly in 2014. It generated extraordinary public interest and almost immediate backlash. Concerns about privacy — the glasses had a camera that could record without a visible indicator — led to bans from venues, public ridicule, and the coinage of the term “Glasshole” for early adopters who wore them inconsiderately.
Google suspended consumer sales in January 2015. The device never recaptured the excitement of that initial reveal. But its commercial failure masked a genuine achievement: Google Glass had introduced the concept of head-worn computing to a mass audience and established most of the questions the industry is still trying to answer. Privacy. Form factor. Social acceptance. Battery life. What the killer app actually is.
Glass lived on as Google Glass Enterprise Edition, finding a quieter but commercially sustainable home in manufacturing, logistics and healthcare — where workers genuinely needed hands-free information access and nobody cared about the design.
2013-2015 — The Research Years, Again
While Google Glass consumed public attention, more consequential work was happening out of sight. In 2013, Meta (not the Facebook company — a separate AR startup) raised funding for binocular AR glasses with a 15-degree field of view. Osterhout Design Group was building modular AR glasses for enterprise. Microsoft was working on something it hadn’t yet announced.
In 2014, Magic Leap — a Florida company with extraordinary ambitions and minimal public profile — raised $542 million in a Series B round. The investors included Google. The product was described in vague, extraordinary terms: a “mixed reality” device that could project digital light fields into the real world so convincingly that the brain couldn’t distinguish them from physical objects. Nobody had seen it. The funding round was the largest in the company’s history. Everyone in the industry wanted to know what they were building.
2015 — Microsoft HoloLens
In January 2015, Satya Nadella took the stage at a Microsoft event and introduced HoloLens — a self-contained holographic computing device that placed digital objects in the physical world and let users interact with them using gestures and voice.
The HoloLens demonstration was unlike anything that had come before. Developers placed holographic apps around a room, manipulated 3D models with their hands, and played a version of Minecraft projected onto a physical table. The technology was real, working and shipping to developers. Microsoft had built what Google Glass had implied was possible.
HoloLens found its strongest market in enterprise — surgeons using it to prepare for operations, engineers using it to inspect components, architects using it to walk through buildings before they were built. It was expensive, heavy and uncomfortable for extended wear. It was also the best spatial AR computing device available for years, and established enterprise mixed reality as a viable commercial category.
2016 — The Year Everything Changed
Two things happened in 2016 that defined everything that followed.
On March 26, Nintendo and Niantic announced Pokémon Go. On July 6, the game launched in North America, Australia and New Zealand. Within days it had spread across the planet, generating two of the most surreal sights of the decade: parks full of people staring at their phones in the dark, and news reports attempting to explain what was happening.
Pokémon Go reached 100 million downloads in its first month. It reached half a billion downloads by the end of the year. It did something no AR product had managed before: it made the technology invisible by making the experience irresistible. Players weren’t thinking about AR. They were thinking about Pokémon. The technology earned its place by serving something people already cared about.
The lesson — that AR succeeds when it’s a vehicle for something with existing emotional resonance, not a product in its own right — is one the industry is still learning to apply.
The second defining event of 2016 was Magic Leap finally showing something. After years of funding, secrecy and extraordinary promises, the company released a video of a virtual whale leaping out of a gymnasium floor in front of a crowd of people. The video, it emerged, was a fabrication — a visual representation of what Magic Leap hoped its technology would eventually do, not what it could actually do. The backlash was severe and lasting.
2017-2018 — ARKit, ARCore and the Platform Moment
In 2017, Apple released ARKit — a framework that allowed developers to build AR applications for iPhone using the camera, motion sensors and machine learning built into every device. In 2018, Google released ARCore, the equivalent for Android.
The release of ARKit and ARCore made AR a mass-market capability overnight. Every modern smartphone was now an AR platform. Developers could build applications that placed objects in the real world, tracked surfaces, estimated lighting and recognised images — all using hardware that hundreds of millions of people already owned.
The apps that followed ranged from genuinely useful to genuinely frivolous. IKEA Place let users place virtual furniture in their home before buying. Snapchat’s AR filters — face lenses that overlaid virtual objects on a live camera feed — reached tens of millions of daily users. Measure let users measure the dimensions of physical objects with their phone camera. None of these applications required special hardware. The platform was already in everyone’s pocket.
2019 — Magic Leap 1 and Microsoft HoloLens 2
Magic Leap finally shipped its first consumer-facing device in 2018 — the Magic Leap 1 — after years of promises. The device was impressive in some respects, disappointing in others, and priced at $2,295 in a configuration that included the headset, a handheld controller and a separate compute puck worn on the body. Sales were modest. The company subsequently laid off roughly half its staff and pivoted to enterprise.
In 2019, Microsoft announced HoloLens 2 — a significant hardware improvement over its predecessor with a wider field of view, more comfortable fit, and deeper integration with enterprise software. The device became the standard for enterprise AR deployments over the following years.
2021-2023 — The Metaverse Moment and Its Aftermath
In October 2021, Mark Zuckerberg renamed Facebook to Meta and announced a strategic pivot to building the metaverse — a shared virtual world that would blur the line between physical and digital reality. The announcement injected extraordinary amounts of capital and attention into the XR sector. It also raised expectations to a level that was structurally impossible to meet in the short term.
Meta’s Reality Labs division spent more than $36 billion over the following three years. The Quest 2 headset sold tens of millions of units and established consumer VR as a real, if modest, market. The metaverse, as a shared social experience, did not arrive. By 2023, the word had become something of a punchline.
What the metaverse era did produce, less visibly, was an extraordinary acceleration in the underlying technology. Display quality improved. Tracking improved. Battery life extended. Processing power increased. The infrastructure for consumer AR — the chips, the displays, the operating systems, the AI models — was built during this period, even as the headlines were largely negative.
2023 — Apple Vision Pro
In June 2023, Apple announced Vision Pro — a spatial computing headset priced at $3,499 that it described not as VR but as something new: a device that let users work, watch content and communicate while remaining visually present in the physical room around them.
Vision Pro shipped in February 2024 to extraordinary initial interest and a more muted long-term reception. The device was heavy, tethered to an external battery, and priced beyond the reach of most consumers. But the operating system — visionOS — was polished in a way that established a new benchmark for spatial computing interfaces, and the micro-OLED displays were the sharpest anyone had put in a head-worn device.
Apple didn’t define the category with Vision Pro. It defined what the category should eventually become.
2024-2025 — The Smart Glasses Era Begins
In September 2024, Meta launched the second generation of its Ray-Ban smart glasses — a camera, open-ear speakers and Meta AI in a form factor that looked exactly like a pair of Ray-Ban sunglasses. The response surprised almost everyone, including Meta. The glasses sold in large numbers. People wore them. The camera worked. The AI was useful. The design was acceptable.
For the first time, a mass-market AR-adjacent device had achieved genuine consumer adoption at scale. Meta had found the formula: take a product people already buy for non-technology reasons, add technology that earns its keep without changing the design, and price it accessibly.
By the end of 2025, smart glasses shipments had grown 167% year over year. The category had crossed the line from interesting to commercially meaningful.
2026 — The Year of the Glasses
The spring and summer of 2026 have brought more smart glasses announcements than any equivalent period in the technology’s history.
Samsung revealed its Galaxy Glasses at Unpacked in London on July 22 — Android XR powered, Google Gemini built-in, available with frames from Gentle Monster and Warby Parker, compatible with iPhone. Snap confirmed a September 16 launch event for its $2,195 Specs — the first mass-market full AR glasses with a standalone display and dual Qualcomm processors. Apple confirmed its first glasses for a WWDC 2027 reveal, with privacy as “priority No. 1.” Amazon is building AR glasses codenamed Jayhawk. Even Realities reached unicorn status with its camera-free G2. ENGO raised €5.1 million for its sports display glasses. XREAL launched its most affordable display glasses at $299.
Courts across the United States have begun banning smart glasses from their buildings. The Texas Attorney General has opened an investigation into Meta’s camera practices. Apps that detect nearby smart glasses over Bluetooth have gone viral.
The technology Ivan Sutherland hung from a laboratory ceiling in 1968 is now the subject of legislation, legal action, acquisition speculation and a global consumer market that didn’t exist two years ago.
What comes next
The history of augmented reality is, more than anything, a story about the gap between what the technology promised and what it could deliver — and the slow, uneven, occasionally spectacular process of closing that gap.
The gap is closing faster now than at any previous point. The displays are bright enough. The processors are powerful enough. The AI models are capable enough. The price is coming down. The form factor is approaching normal.
What remains is the hardest part: finding the thing people want to do badly enough that they’ll wear a computer on their face to do it. Pokémon Go found it, briefly, in 2016. Meta found it, partially, in 2024. The full answer — the application that makes the hardware disappear because the experience is too good to put down — hasn’t arrived yet.
When it does, everything that came before it will look like prologue.