The Limits of Gen AI, the Immersion Divide, and The Rise of the Metaverse Generation

Above: InTrance and LAGurlz, creators of Starving Artists, a game approaching half a billion visits

If you accept the common Valley wisdom, the Metaverse is “dead”. As recently as 2022, tech conferences and startup pitch decks were awash with the term, first coined in Snow Crash; now, you scarcely see it anywhere, as if wholly supplanted by generative Al.. The backlash is so profound, it almost has the feel of an intellectual purge. 

So what I’m about to say may seem totally deranged, but I assure you it’s completely defensible:  

Metaverse platforms continue to grow in mass market adoption and revenue, far outpacing even current darling OpenAI in terms of actual user numbers. (More on those numbers down below.) This growth is happening while sustainable, scalable use cases for generative AI still seem far away or niche, and sober analysts are already pointing to the category as likely being an overhyped bubble.

Why is the business world missing this? The answer, I’ve discovered, speaks to deeper cultural and generational biases that can even obscure sound financial analysis.

The Metaverse by the Numbers 

To level set our references, the Metaverse as originally described in Snow Crash is roughly this: A vast immersive virtual world simultaneously accessible by millions of people through highly customizable avatars and powerful user creation tools which are integrated with external technology and the real world economy. Not coincidentally, that also happens to describe the core features of highly popular platforms like Roblox and Fortnite. That’s not a surprise, since both of them (and many others) were directly and avowedly inspired by Snow Crash.

Most of the leading metaverse platforms I tracked in my book Making a Metaverse That Matters in early 2023 have since seen strong growth. According to analytics firm RTrack, Roblox in December 2023 reached 354 million monthly active users; it counted 400 million MAU this July, and currently has 380 million MAU, 

In other words, not only does Roblox already have more users than the entire United States has people, it’s in striking distance of reaching half a billion MAU fairly soon. And that’s just one of many successful metaverse platforms.

Overall, total monthly user growth tracked among the leading metaverse platforms -- including Roblox, Fortnite (100 million MAU), Rec Room (23 million MAU), Avakin Life (14 million MAU), and VRChat (10 million MAU) -- has increased by about 140 million people since 2023, for a total global audience of an estimated 690 million active users.

To be sure, this is incremental growth to existing platforms; but surely generative AI represents a transformational, categorical leap?

Maybe, but that’s still far from proven. And setting aside the rhetorical jazz hands of AI evangelists, the actual numbers are not quite convincing.

Metaverse v. Gen AI: Some Hard Contrasts

The top generative AI platform by far, OpenAI’s ChatGPT, only has some 100 million monthly active users -- far less than Roblox. This is obviously not an apples-to-apples comparison, but in basic terms of “activity that many people regularly do on the Internet”, metaverse platforms still dominate over gen AI.

A revenue comparison is also not an impressive win for the AI narrative. OpenAI is on track to earn $3.4 billion this year… which is almost exactly the figure Roblox is forecast to earn in 2024. And what revenue AI earns comes at tremendous losses: as Sequoia Capital recently reported, AI companies will have to earn about $600 billion per year just to cover the costs of their computing infrastructure. Which might be justified if a compelling scalable killer app had already emerged, but as Goldman Sachs firmly noted, it has not.

So why has attention shifted so thoroughly away from the Metaverse towards such a profoundly unproven technology? Valley flightiness for the latest buzzword is part of the explanation, as is Meta’s uniquely floundering metaverse strategy, desperately centered as it is around its own line of VR headsets. (Despite Neal Stephenson himself frequently insisting VR is not a prerequisite for the Metaverse. ) 

But I’m convinced something deeper is at work.

The Cultural Divide of Immersion

By my estimate, the current potential audience for the Metaverse is roughly 1 in 4 people. That’s approximately the number of people in the world who enjoy immersive digital experience -- i.e., 3D games, for the most part, mainly on video game consoles, PCs, or higher-end smartphones.

The majority of them are Millennials and Gen Z. For them, immersive games are not simply games, but core to their online social experience -- where friendships are created and sustained, virtual communities that transcend distance and real life origin are formed, and new ways of looking and thinking about life are delightfully experimented with. (Roblox began, by the way, as an educational tool; Minecraft’s education edition is extremely popular.)

Very few people who grew up on immersive interactivity, however, are part of the demographic who largely comprise the investor class. Their flight to AI may seem like a rational apprehension of new technology, but the dream of creating thinking machines predates Snow Crash by decades if not centuries. Not having grown up on interactive immersion as an integral thread in their social fabric, it’s difficult for them to see the appeal of metaverse platforms like Roblox and Fortnite, and easy to dismiss it all as “just online games” -- even when the evidence of their continued growing mass adoption is right in front of them. 

I’m convinced this cultural/generational blind spot is causing Silicon Valley’s leadership to miss the next great creative / technological shift of the social Internet -- even when it’s often happening right down the hall in their own kids’ bedroom. 

“Zuckerberg grew up on IRC and on email, and he created a product [Facebook] for him and his peers,” as Matthew Ball put it to me recently, while talking about the new 2024 edition of his essential book, The Metaverse, “What I am excited when it comes to Roblox and others, is that technology really emerged between 2005 and 2015, and then we've had a generation that's grown up using it, and we're just starting to get to the point where the people who grew up with it will start to create products themselves that reflect the usage.”

I’ve interviewed this new cohort for my own book, and the breadth of their creative ambition is scintillating. I’m thinking of people like “LAGurlz”, a girl from Jamaica who began playing around in Roblox in her teens; now in her early 20s, her Roblox experience, “Starving Artists”, attracted over 400 million visits in its first two years. 

Upcoming Metaverse Platforms to Watch

While missed by the many people in tech still fixated on AI, the race to capture this Metaverse Generation is already happening in the form of several startups. To name just four: 

EVERYWHERE: Now in closed beta testing, the platform from veteran developers of the Grand Theft Auto games, Everywhere is a platform enabling users to “unleash your creativity to build, remix, and play with others in a vast connected world.”

HELIX: Hypersonic Laboratories is a new startup which just got funding to build HELIX, an upcoming multiplayer role-playing sandbox platform for user-generated content (UGC). I recently got an early peek at what they're building, and the ambition is pretty impressive. As David Chiu, Hypersonic's head of business, put it to me recently: "Think of it as Garry's Mod meets Grand Theft Auto meets Second Life powered by Unreal Engine 5".

Playable Worlds: Founded and led by venerable game designer Raph Koster, Playable will launch (perhaps as early as this year) with a “games first” approach: The primary goal will be to roll out a successful sci-fi MMO, and once that's out, gradually release ever more UGC tools/features so that players can sustain and build on that activity.

RP1: Unveiled at Augmented World Expo 2024, backed by a bootstrap startup, RP1 is creating a platform that's the closest I've seen to being the canonical Metaverse.  

This is just a small sampling of startups building something like a metaverse platform as defined above; many of them are now loath to use the forbidden word, keeping their Metaverse ambitions close to their heart. Having grown up on immersive experiences, they know their social power, and transformative potential. 

By the time the rest of the tech world finally realizes that too, chances are the Zuckerbergs of the Metaverse Generation will already be ascendant -- and happily referenced in tech conferences and startup decks, as if the current backlash never happened.

Please support posts like these by buying Making a Metaverse That Matters and joining my Patreon!

Become a member!

Comments

[Iggy 1.0]

Hamlet, I get your point, but you miss one thing: work.

GenAI is developing into a workplace co-pilot. That is how my college students regard it. There is no apples-to-apples comparison to virtual words, for now a place to game and socialize.

15 years ago, my students’ elders did not get into SL, partly because it was sold as something they did not need. They were using social networks and gaming for arguably superior experiences, as compared to the laggy, graphics-challenged SL of 2005-2010. SL’s killer app, collaboratively building a simulation, lay beyond nearly all of their abilities.

Today, my careerist Gen Z students want jobs. They know that they will use AI there. It’s boring, but so is MS Office, a little app for doing work.

If had to pick one tech sector for investment it would be AI.

[Wagner James Au]

> collaboratively building a simulation, lay beyond nearly all of their abilities

Yet a huge, huge number of them went on to do that very thing in Minecraft, Roblox, Fortnite, Rec Room, VRChat, etc.

[Iggy 1.0]

Yes indeed, Hamlet. But how many of those platforms are used on the job?

That is the boring venue for GenAI, which gives one reason for rapid uptake by US college students.

[Wagner James Au]

Well like I note up there, ChatGPT has way less active users than Roblox. Students are using it when needed, like any other tech tool, but the lower usage suggests it's not a primary focus. Very, very few people *primarily* use a program like ChatGPT for work the same way that professional 3D content creators use Fortnite etc. to help make a living.