The office building on Facebook Way is in the unfinished style that honors materials like plywood, concrete, and steel. The I-beams supporting its soaring walls still have the builders’ chalk placement instructions on them. It takes a business making billions of high-margin dollars to make plywood and concrete seem so appealing. The merely ordinary have to put up drywall.

Facebook’s spokeswoman calls its headquarters the largest single room in the world. Maybe. It feels like it, anyway. The space isn’t square, so it doesn’t seem pointedly vast; it’s long and narrow. Heading to meet Mark Zuckerberg, the wizard of this open-plan office, you wind through it like an Ikea, following a painted path. The desks are orderly and clean with minimalist Macs. From time to time, there’s a map with a “you are here” helpfully posted. Then, at the center, standing at his desk announcing something to a colleague, there’s Zuckerberg. He’s a great stander; he has terrific posture. Sheryl Sandberg, chief operating officer, author of Lean In and the de facto leader of all corporate women, wraps something up and heads down the painted path.

Featured in Bloomberg Businessweek, August 1-7, 2016. Subscribe now.

Cover Illustration: Steph Davidson for Bloomberg Businessweek

If you spray-painted Zuckerberg a high-gloss white and told him to gaze off into the distance, he’d look exactly like a 1st century A.D. bust of Tiberius at the Capitoline Museum in Rome. Zuckerberg would get the reference. A scholar of the classics, he named his daughter Maxima, after the Roman, not the Nissan, and once declared at an anti-Google Plus all-hands meeting, “Carthago delenda est.” (This was Cato the Elder’s call to destroy Carthage, which posed a threat to Rome’s active user base.) Zuckerberg doesn’t wear a toga, unfortunately, but like any icon, he has a signature look—gray T-shirt, jeans, and sneakers.

He joins the interview immediately, projecting sunshine. This won’t be a grind, like his recent emergency meeting with U.S. conservatives, convincing them that his army of overachieving twentysomethings isn’t totally biased in what it promotes on the news feed. Our subject today is … The Future. In particular, that of Oculus, the maker of virtual reality goggles and software, which Zuckerberg bought in 2014.

The interview takes place in a fishbowl-like room in the middle of the middle of the world’s largest room. There’s an L-shaped gray midcentury-modern couch, a square coffee table, two enormous black flat-panel screens. Zuckerberg has light green eyes that fix on an interviewer like security cameras. You can’t avoid them. You can’t figure out exactly what’s on the other side of them, either.

In one way or another, he says, he’s thought about the rendering of reality for decades. “It’s something that I’ve dreamed about since I was a kid,” he says. “I remember in middle school I would just sit in my math class with my notebook and write code. I didn’t even have a computer in middle school. I’d just, like, go home and write it. And I sketched out how I thought that eventually the operating system and the experience should be 3D, and basically more of a VR thing.” Zuckerberg is 32 and was in middle school in about 1995, a few years after Neal Stephenson outlined his dark vision for the “metaverse,” a computer-generated alternate reality, in his novel Snow Crash.

About 20 years later, Zuckerberg offered the founders of Oculus $2 billion to join him. That was hard to turn down. First of all, it was $2 billion. More important, it implied the long-term backing of Zuckerberg, who controls Facebook with a special class of shares. There are other shareholders, many of them, but they don’t have the same rights or power over the cash flow, which amounted to about $1.8 billion in the first quarter of 2016.

Oculus is Latin for “eye,” and the Oculus Rift, which went on sale earlier this year and lists for $599, is an incredible device. Strapped to the head, it offers 360 degrees of vision and sound, potentially opening new possibilities in playing games—the gateway drug for VR, Zuckerberg says. He also wants it to be used for watching sports, making movies, joining conversations around the world, or things no one’s imagined yet. But it’s still limited—in resolution, how it tracks movement, and how the body responds to what it projects, among many other things. The problems are enormous and require a deeper understanding of human sensory mechanisms than currently exists. (For example, how should a pair of goggles follow the movement of the eye to allow the processor to manipulate the plane of focus?) It’s going to take billions to make it work.

Zuckerberg, asked about this directly, doesn’t flinch at the thought of building a NASA-like research park for VR. “This is early, and it’s going to be a long-term thing,” he says. “This is a good candidate to be the next major computing platform. It’s worthy of a lot of investment over a long period.”

He often talks of connecting the world. But with virtual reality, the terms of that connection have been upped exponentially. “We’ve connected 1.65 billion people through Facebook,” Zuckerberg says. “But if you want to help get all 7 billion people connected and make a step function in the fidelity of how people can share and consume content, you need to make significant investments in some of these longer-term things where you actually don’t know what the time horizon is. … I don’t know who said this first, but it’s not hard to predict what the world will be like in 20 years. The hard thing is actually predicting or figuring out how to get there.”

A decade ago, people online mostly shared text. “Then we all got good cameras that were attached to our phones, and it got a bit richer,” Zuckerberg says. “And now we’re at the beginning of this—we call it the golden age of online video, and that’s just richer. Photos are richer than text; video, much richer than photos. But that’s not the end, right? I mean, it’s like this indefinite continuum of getting closer and closer to being able to capture what a person’s natural experience and thought is, and just being able to immediately capture that and design it however you want and share it with whomever you want.”

The spread of video has taken thrilling and sometimes shocking turns, and VR will likely build on that in ways we don’t yet understand: Imagine Facebook Live at a riot, but in fully immersive form. Talking about the future, even Zuckerberg can get stumped and slide into the mystical. Some of the problems don’t even have names yet. VR, at true fidelity, entails creating another reality, the presence and automation of everything that exists. Then there are the deep problems of connecting to the brain, which leads to telepathy, something he isn’t opposed to discussing. “I think you get—I mean, there’s something that’s just—that’s deeper, that I don’t even think we scientifically understand about just how you actually experience the world,” he says. “I think there’s outside, and there’s different fidelities of capturing that. And then there’s the human experience of it, which I think is—I mean, we don’t even have enough of a scientific understanding to even have—I think I don’t, at least, have the vocabulary to even fully describe this.”

Apart from cracking the workings of the mind—pretty cool!—Oculus will give Zuckerberg the chance to actually make an object, as opposed to the intangible millions of lines of code that constitute Facebook. That code is an achievement, but a physical object that people love might be talked about a generation later. People hate throwing things away, because they remember how they loved them. Things get tied up with one’s youth and years later get found in an attic and caressed, or maybe traded on EBay.

There’s a less romantic reason for getting into the hardware business, too: Facebook wants to own VR the way Apple and Google own mobile. That means taking control of the technology, from the software to the hardware.

But making things at a mass scale is ruthless and brutal in a way that coding isn’t. Getting a billion or more pairs of goggles into people’s hands will be an immense undertaking. Apple makes things, but its 110,000 or so workers don’t put them together. Hundreds of thousands more, at places like contract manufacturer Foxconn Technology, do that. Samsung has factories, too, and they employ almost 500,000. Facebook now has about 13,000 workers. The Oculus division had about 400 the last time Facebook announced the number. The company won’t say why, but it no longer announces the number. It may be trying to build something bigger than anyone thinks, and even if you’re a benign autocrat, as Zuckerberg essentially is at Facebook, shareholders get skittish when you start getting into the land war that is hardware. John Carmack, Oculus’s chief technology officer, says he’s well aware of what they’re up against: “I had an aerospace company for 10 years. I understand the cussedness of physical things.”

The hardware war will be waged on the grounds of the old Facebook headquarters, across the way from the new digs. When employees want to go to the older part, they hop on brand-new blue bikes and pedal through a tunnel. Offices surround an alley that looks like Disneyland’s Main Street. There’s a coffee shop, a burrito joint, and an arcade, among other shops, all free to employees. Next to a candy store there’s a printmaking shop, where employees make beautiful posters with team-building slogans that get put up on the few walls in the office. (Carthago delenda est, for instance.) Only an ascetic misanthrope wouldn’t want to work at Facebook. It’s usually sunny, too.

At the end of this snack food allée, in a repurposed warren of offices, Oculus is spreading its messy-reality tentacles over expanding rows of desks. The rooms are tangled, forbidding, a jumble suggesting workers too busy to tidy up. Surfaces are piled with headsets, lenses, knots of wire, chips, boxes. There are oscilloscopes, mass spectrometers, power supplies, soldering irons, naked circuit boards with chips on them and USB cables sticking out, isolated testing platforms on special wide legs that damp vibration, and giant microscopes. On one table, three cameras focus on a tiny sphere balanced on a needle. On another, a vacuum pump leads to a chamber that’s like a pressure cooker. The Rift uses two lenses, and these are everywhere, balanced on little stands in front of testing equipment.

Here in this gadget maze are the offices of Palmer Luckey and Nate Mitchell. Luckey, now 23, has been the boy-inventor face of Oculus since it first blew the media hive mind in 2012. But Facebook would like to rewrite this founding story and make it less boy-in-a-garage and more nerd-supergroup, so it’s made him available only in a group interview, and not for photographs. Luckey betrays no anxiety about such revisions. Sitting in Mitchell’s office—which has its own name, Nothing to Announce at this Time—Luckey wears a Hawaiian shirt, cargo shorts, and sandals. He has all the poise and confidence one would expect from a home-schooled millennial with an awesome name, on whom everyone wants to hang the invention of virtual reality.

Mitchell, 29, affable, put-together, and of the super-positive variety, wears a gray hoodie. The vice president for product, he’s been working with Luckey since the beginning. Mitchell spends most of his days figuring out how to improve the technical aspects of the Rift, while Luckey woos developers, who are also being asked to build games for rival headsets. HTC’s Vive allows greater range of motion in a room than the Rift, and hand controllers are already available. Sony’s PlayStation VR, launching later this year, will try to turn the audience for existing video games, such as Resident Evil, into customers. Magic Leap, an ostentatiously secretive startup in Florida, won’t explain exactly what it’s building, saying only that it’s working on augmented reality, or AR. A sort of partway VR, AR involves digital images that appear to interact with the real world. Magic Leap has already signed deals with movie studios, including Lucasfilm, which makes Star Wars, and attracted gobs of investor cash—$1.4 billion and counting. And, of course, there are the simple, no-goggles-required smartphone AR games, such as the suddenly ubiquitous Pokémon Go.

Asked if he’s scared Oculus will be beat, Luckey answers, “I’m never scared. I know too much.” He goes on: “Some companies are figuring out their long-term vision as they go in terms of where is this technology going to be in 10, 20, 30 years. But for a lot of the people here, we’ve all read science fiction. We all know what virtual reality is in sci-fi. Even though the product we have today is not the one that we want to have 10 or 20 years from now, everyone wants to get there. The goal is clear: It’s to make VR technology that’s as real as real life with none of the limitations.”

The science fiction he’s referring to is not only Snow Crash, but also Ready Player One, by Ernest Cline, which imagines a Facebook-like “Oasis,” where all social networking happens in VR. Cline talks with the Oculus people frequently. New employees are issued a copy of the book.

Mitchell and Luckey were brought together by Oculus Chief Executive Officer Brendan Iribe, whose office, named Inception, is just down from Mitchell’s. Above the desk and couches and armchairs is an odd steampunk-style bust by Jeremy Mayer, an artist in Oakland, Calif. Made entirely of parts from Oliver manual typewriters, the figure is dubbed Papa Oliver. He’s a nod to the past but also a wink at a bionic future, when we may live in haptic suits that sense and translate every movement and reflect them in our avatars, ocular implants show information on our retinas, and so on through the visions of Cline and others. Iribe bought one for himself and one for Zuckerberg.

Iribe, 36, has dark turned-up hair and an easy, confident manner; his mother says he never had trouble walking right up to the counter at the computer store when he was little, and sometimes right past it to the backroom. He has many roles, but the main one is making sure the final product Oculus is building today is elegant, intuitive, and comfortable—and doesn’t make people throw up. Nausea is the “Um, sir, there may be a problem” problem of VR.

He thinks bigger thoughts, too, about where VR might be years from now. He wonders about things like what would happen if Oculus one-upped Google Earth and mapped the whole world in 3D, or what it might take to get eye tracking right. “I sleep on and off—a couple hours—and then wake up and think about it, and then go back to sleep and think about it more,” he says.

Iribe grew up in Maryland and spent a year at the University of Maryland before leaving to work in software development. In April he went to his alma mater to inaugurate the construction of a computer science center named for him after he pledged $31 million. He stood under a tent with a senator, the governor, and Oculus executive Michael Antonov, also a former student. They all put on VR goggles and pretended to break ground. (With Antonov, Iribe founded Scaleform in 2004, which sold software tools to video game producers. It was bought by Autodesk in 2011 for $36 million.)

As Iribe remembers it, a friend called him from a product show in 2012 and told him, “You need to go meet Palmer. He has this really cool prototype, and I think VR is, like, finally ready to work, maybe.” So Iribe brought a group together at the STK steakhouse in Los Angeles, including Mitchell and others Iribe had worked with. At the dinner, he says, “Palmer walks in with his shorts and flip-flops and Atari T-shirt. I’d spoken to him on the phone, but I didn’t know how young he was.” Luckey was 19.

Iribe pitched turning Oculus, which had been a message board-inspired Kickstarter project to help people build their own VR headsets, into a company that could sell a fully designed headset direct to consumers. If this was the dawn of a new computing age, here was Steve Jobs telling Steve Wozniak that the Apple computer should be for regular folks, not just engineers.

But Luckey already had another job offer. “I was considering a lot of different options,” he says. “We met, and I talked with Brendan and his friends. He really helped convince me.”

Says Mitchell: “I do remember you said there was this one job offer, and you said, ‘Think about this salary. I’m going to walk away from that.’ Someone was like, ‘Palmer, if you start your own company, you can pick your own salary.’ ”

Iribe: “The big pitch to Palmer was, Look, we have done this before, and we’re all still working together. We’ll do a partnership, the four of us. … We’ll get equal share. I’ll be the CEO, and we’ll put the company and the product together, build a team, raise the money”—the company was almost entirely funded by Iribe, his family, Antonov, and Mitchell, in addition to the Kickstarter money—“and you’ll be the founder and evangelist, and you can go out and be spokesman. You have this incredible story, developing this in your garage for years.”

Luckey, remembering it all, laughs. “I wasn’t convinced for a while,” he says. “It took a few weeks. It was pretty clear that Brendan was the right guy to work with. I never wanted to be CEO. That’s just not my skill set. Some people can be that founder-CEO. I knew from the very start I never wanted to be a CEO. It’s not my type of role.”

At Sanzaru Games, they joke a lot about getting the bucket ready. Sanzaru makes one of the 70 games and experiences Oculus has seeded, hoping they’ll come up with a winner. The bucket is for vomit. Tin Guerrero, Sanzaru’s creative director, is working on a sports game, and his team uses the bucket to answer questions like, “What will happen if we have the user running down a field with a football?” There’s a guy named Flemming Wahl who gets sick easily. When Guerrero wants to know if a feature works, he has Wahl test it. If he doesn’t get sick, it works. Wahl isn’t some kind of weakling—his hobby is racing dragon boats—but for some reason he’s sensitive to the vestibular distortions of VR.

Michael Abrash, chief scientist at Oculus, spends a lot of his time studying perception. For Abrash, the great dress fight of 2015—where the world couldn’t agree on whether a dress was blue and black or white and gold—wasn’t a time-wasting meme, but a fundamental question about what the brain processes and why. For Abrash, Oculus can’t stop at games and gimmicky immersive experiences. It’s supposed to be as good as reality. It’s supposed to be reality.

“When you wake up in the morning, you don’t say, Ah, reality, what an interesting platform,” Abrash says, lifting his arms to show off the slice of life that is this Facebook conference room. “You don’t think, What’s the killer app? You think, Anything is possible.” He motions to his half-rimmed glasses and his phone. They’re examples of ways we already augment our perception of the world, he says. But before VR, “we haven’t ever had it so tightly coupled to our perceptual system and our environment.”

The problem for a research team benchmarking against actual reality is that Oculus falls short in so many ways. The way lenses are designed now, field of view is 90 degrees, not the 110 degrees your eyes have. And there’s no way to adjust depth perception so you can focus on a strand of hair and then something in the distance without highly precise eye tracking. “The only way to figure that out is to build it,” Abrash says. “This is only a perceptual psychology problem. The key is that what you experience is constructed in your brain.”

Tracking eyes, for example, isn’t as simple as tracking pupils—which change size and may not be symmetric. The eyes wiggle and the iris travels during each blink. “If you took a video of your eye and watched it in slow motion, it would be very disturbing,” Abrash says. Eventually, Oculus will need to track mouth movement and hand movement, which are potentially even harder, but necessary for people to hang out and have conversations in some VR chat app of the future.

Oculus early on committed 20 percent of its budget and hiring to the research division, and Abrash spends most of his time trying to find people who’ve actually studied the things the company is trying to solve. In fields such as nanofabrication, nanolithography, and waveguide technology, he says there are only a few people in the world to ask. Finding them has been complicated by how completely VR stalled after its failure in the ’90s. There was nowhere for a specialist in VR to continue to grow. “I talked to one of those people to see if he wanted to work with us—it turns out he’s a doctor now,” says Abrash. “Things just imploded so completely that people walked away from it.”

In the early ’90s, Abrash met CTO Carmack on an online bulletin board for 3D graphics. They got together and built Quake, a hugely popular video game, then went their separate ways, Carmack to ID Software and Abrash to Valve. About 15 years later, Carmack introduced him to the Oculus team, before Facebook bought it. Abrash tried a demo where he looked over the edge of a tall building and felt his knees lock up. That’s when he knew the future described in his favorite sci-fi novels was possible.

He quit his job, anticipating a long vacation before Oculus could raise enough money to get serious. Five days later he got a text telling him the company had been acquired for $2 billion. “I thought, Well, the train’s leaving the station,” Abrash says. He met Zuckerberg and grilled him about whether VR was a serious part of Facebook’s strategic plan, then he hopped on.

Carmack is at a Microsoft event the night before E3, the annual gaming conference in Los Angeles. VR is such a new, open field that competitors are more collaborative than they might be otherwise, and Carmack is at E3 to introduce the VR version of Microsoft’s Minecraft. The game runs on the Gear VR, a stripped-down $99 set of goggles that holds a Samsung smartphone inside. (The gear is a Samsung-Oculus collaboration based largely on Carmack’s work.) VR on mobile phones is the technology Carmack thinks will spread first and fastest. “The phone,” he says, “is the golden path to how we get to a billion users.”

Carmack, 45, has spiky blond hair and small rimless rectangular glasses. He answers questions with precise, unusually inventive language. He joined Oculus in the early days, after meeting Luckey through a message board devoted to VR. He describes Luckey’s first prototype as “a warpy, distorted mess.” Carmack worked with Luckey to refine it and brought it to E3 in 2012. “This little shoe box that Palmer had taped together with two plastic lenses and a surplus screen was better than these super high-end displays that cost a hundred times as much,” he says. “But this was the thing that made people go ‘Whoa.’ They saw it. They felt it.”

Minecraft, says Carmack, is “the biggest game in existence.” He spent months persuading Markus Persson, its inventor, to work with him, and then Microsoft, after it bought Minecraft, to let him turn it into a VR game. And then Facebook bought Oculus. “I was a big backer of the Facebook acquisition,” he says, explaining that he understood enormous resources would be required for true VR to come to fruition. He was so all-in, he even got a Facebook account.

He says he has no one reporting to him, which frees him to study problems like sensor fusion, the process of getting different locating technologies to work together. Carmack ran Armadillo, an aerospace company he founded, for about a decade, and he’s applying some of the same positioning technology to VR. “Sometimes you have an opportunity to build something from the future, before there’s a top-100 list for the genre,” he says. “I am more excited about this than anything that has come before.” Like Abrash, he met with Zuckerberg before joining Facebook. He wanted to gauge Zuckerberg’s commitment and came away convinced.

In a way, it all hangs on Zuckerberg, who tracks his personal and professional goals in an almost aggressively pedantic manner. He counts miles for a year of running and logs books for a year of reading. Asked at the interview in his office if he’s ready for the long term, he leans in, looks around the room, and fairly yells, “I don’t think we’ve met before!” Then he seamlessly slides into, “We’re a very mission-focused and long-term-oriented company.”