With an estimated 2.95 billion gamers in the world in 2022, gaming is certainly the most common use case for 3D avatars. No matter the genre, any 3rd-person game requires the player to have some kind of virtual representation to carry out in-game action.

Possibilities vary greatly. In story-heavy titles, visual customization is often constrained by the developer’s control over its IP. In Sony’s God of War or Nintendo’s games, customization applies only to skills and power, not to the character’s signature look. Things are very different in mechanic-focused games like Epic’s Fortnite that don’t rely on a specific character and offer no skill tree but instead let players go crazy with cosmetics.

Overall, our in-game identities remain highly fragmented. Each game comes with its own avatar system, cosmetic offering (or lack thereof!), range of motion, and more. The same player may single-handedly play as Mario, Kratos, or Lara Croft, yet these characters’ appearances, skills, and achievements are kept completely separate. Despite some early attempts, still no single avatar system can claim to serve as a universal identity layer in gaming today. The integration of Snap’s Bitmoji system was put on hold indefinitely when the company put an end to its gaming efforts; meanwhile, Epic’s building block for “Identity Provider Management” focuses on a user’s underlying data rather than their avatar’s.

Social platforms are another key destination for avatars. Although the trend debuted many years ago in the context of web forums and internet messaging (IM), it was given new life as various players started playing with 3D. For example, Snap’s Bitmojis can now be used to generate in-chat reactions and augmented reality (AR) “mini movies”, while Apple’s Memojis span in-chat stickers, 3D voice memos, and AR lenses.

The proliferation of new social apps has only heightened the prominence of avatars. Data.ai found that avatar-based social media apps such as IMVU (a BITKRAFT portfolio company), FlickPlay, and ZEPETO “have seen the largest YoY growth in terms of downloads compared to any other social media category at 38 million,” “up 60% YoY and 215% from two years prior.” As is true in gaming, each new destination typically introduces its own avatar system for a unique look-and-feel, keeping the social ecosystem fragmented from the user’s standpoint.

Gaming and social platforms may be front and center, but we’d be remiss to overlook the multitude of other potential use cases.

Today, avatars are enabling a wave of virtual-first artists and bands and empowering tech-savvy livestreamers like CodeMiko with new creative capabilities. They’re even showcasing signs of empathy and intelligence as virtual pets or companions. Together, these examples speak to consumers’ growing comfort with and appetite for digital identities.

On the B2B front, too, avatars are entering virtually every industry. In sales, Synthesia helps customers scale their campaigns by turning a single text script into potentially hundreds of videos covering different ethnicities, ages, styles, and even languages. Hour One uses avatars for corporate training, while Soul Machines mentions healthcare, customer support, and education. Here, the focus is on scalability and efficient integration of an avatar with the customer’s proprietary data streams.

If there’s one thing that the current landscape of avatar-focused companies makes obvious, it’s that there is no single view of what avatars should look like. Depending on its particular ambitions and strengths, each company has to position itself differently along a variety of axes impacting both an avatar’s design and its owner’s experience. Here, we’ll focus on two of them: realism, and centralization.

If there’s one thing that the current landscape of avatar-focused companies makes obvious, it’s that there is no single view of what avatars should look like. Depending on its particular ambitions and strengths, each company has to position itself differently along a variety of axes impacting both an avatar’s design and its owner’s experience. Here, we’ll focus on two of them: realism, and centralization.

On the one hand, the success of blocky platforms such as Minecraft and Roblox is proof enough that fun isn’t dependent on a game’s graphics. “Low-poly” content not only allows developers to create more, faster but also tends to age well since it doesn’t risk falling behind any particular visual trend. It also makes for less technically demanding games, a prerequisite for mass adoption as it bypasses the need for high-end devices and enables larger numbers of concurrent users inside a virtual space.

On the other hand, expectations are noticeably shifting. First, continuous hardware advancements across mobile, PC, and console have led players to see detailed textures and movements as table stakes. Second, specific titles are acting as a forcing function: Fortnite, with its cartoonish yet lifelike light rendering and textures, undeniably raised industry standards. Even Roblox is now pushing toward avatar realism, notably through “layered clothing” and “Dynamic Heads,” a feature that will let avatar heads support real-time facial animation.

The reason why realism matters is because it impacts not just how your avatar looks, but also what you can do with it. The more movable parts an avatar has, the greater the range of animated motions, enabling more natural and granular ways to communicate, show off, and celebrate. As we spend increasingly more time in virtual worlds and start looking for ever more ways to express ourselves, realism is likely to become even more central.

If Meta’s “Codec avatars” and Epic’s MetaHumans are any indication, much can be achieved using today’s state of the art tech. But a few enabling technologies could bring realism to the next level.

The most widespread yet, photogrammetry, triangulates the geometric properties of an object “by using multiple overlapping pictures from different positions and angles.” Among other things, this enables a user to generate a 3D model using only their smartphone’s camera. The technology has attracted significant M&A activity lately, as 3D capture is increasingly considered a cornerstone for the Metaverse.

Another contender is volumetric video, which captures live video from multiple angles and converts it into a 3D model that can be manipulated at will. Today, the technology is being used for everything from music videos and performances, to short and feature films, to sports content. But the need for high-tech set-ups limits its use to projects with high production value — and adequately high budgets.

Last but not least, light field technology aims to represent environments and objects as light fields, the flow of light through an area or volume of space. By capturing the intensity, direction, and reflectance of each ray of light, one can extrapolate “the correct perspective of a scene for any arbitrary virtual camera position.” The result is a navigable photorealistic rendering almost indistinguishable from real life, a promising feature for holographic and AR/VR content. Despite this potential, the technology is currently constrained by specialized hardware requirements and a tremendous data footprint that makes both storage and delivery impractical at scale.

Looking ahead, low poly will likely endure, if only as a way for challengers to enter the lower-end hardware market — or to conjure up a nostalgic feel. But the overall trend certainly points toward higher fidelity.

Beside realism, avatar-focused companies are having to position themselves on another spectrum: that of centralization.

At the most centralized end of the spectrum is a company like Meta. As a platform itself, Meta’s hope is that developers will adopt its proprietary avatar system and look-and-feel as their core identity layer instead of building one from scratch. Via a dedicated SDK, third-parties on the Oculus App Lab and Store can access the company’s technology and future updates, the direct benefits of Meta’s considerable R&D power.

Somewhere in the middle of the spectrum, startups are following more of a “pick and shovel” play and trying to become not full-fledged platforms (though that may come!) but interoperable plug-and-play solutions across every app and virtual world. The most representative and arguably most successful example of this is Ready Player Me, which “today handles about 5 million avatars from across some 3,000 partners.” In contrast with Meta’s obviously branded tools, Ready Player Me makes itself available as a neutral provider, blending with its partners’ brand identities and acting as a bridge between them.

One last alternative lies in fully open solutions built upon web technology. For example, Webaverse, an immersive OS for virtual worlds, is leveraging VRM, an open-source file format for humanoid 3D avatars, and experimenting with “interoperable file formats, wearables, and animation systems.” Tapping natively open formats and software in this manner seems like the best way to maximize the creator audience and enable User-Generated Content (UGC).

Each approach comes with its own challenges. The platform play is notoriously hard to build from scratch, as it requires significant reach to attract developers in the first place. The “pick and shovel” strategy essentially makes a company B2B, limiting the potential for future IP developments that are typically pursued by consumer media companies. Finally, building on open-source solutions gives a company little defensibility against its competitors.

Until concentration inevitably happens — a successful “pick and shovel” player would be a dream target for any large virtual platform — we’ll likely see many more avatar systems pop up on the market.

Up to this point, we’ve covered the potential of avatars at large. We’d now like to take a closer look at how they fit in the context of Web3.

As a reminder, Web3 is a version — some say, the next version — of the internet built on open, decentralized technologies, chief among which, the blockchain. Although Web3 for now isn’t anywhere close to fully replacing the internet as we know it, as a new technological paradigm, it has the potential to impact digital identity in a number of ways.

While the term “avatar” typically refers to a person’s look as a whole, that look can be composed of a multitude of different items: in Fortnite, even the most extravagant outfit leaves room for your preferred “back bling,” glider, and harvesting tool. The more customization options, the more granularly and accurately an avatar can represent you.

Non-Fungible Tokens (NFTs) seem uniquely suited for individual cosmetics. As verifiably unique assets, they can appropriately reflect the uniqueness of an accessory. With ownership permanently recorded, and the safety of self-custody, the risk of seeing one’s assets confiscated is also greatly mitigated. As for monetization, blockchain-powered royalties could enable individual creators and developers to capture financial upside over the long term, rather than limit their audience due to high sale prices.

Multiple companies are leveraging this potential across categories. Among them is the now Nike-owned RTFKT, which started out with virtual jackets, pants, and sneakers. In the watch world, legacy brands including Hublot, Bulgari and Panerai have all experimented with NFTs, joining more digital-native ventures like the jewelry-focused Chains. Jadu stood out early on with AR-ready jetpacks and hoverboards.

We expect developers to increasingly start turning part or the whole of their in-game inventory into NFTs. Just recently, Roblox at its Roblox Developer Conference revealed that all items on the platform will have limited supply enforced through “scarcity contract” — something that sounds strangely similar to NFTs, even though the company has done its best to not use the term once. A few months ago, Reddit airdropped free Polygon avatars to its most active users. We’re excited to see many more companies across the consumer media landscape wake up to the blockchain’s potential for enabling unique avatars.

As discussed, the current fragmentation of avatar technology across a multitude of distinct IP, avatar systems, degrees of fidelity, and in-game capabilities not only represents considerable friction for users, but also forces developers to reinvent the (digital) wheel for their own needs.

The way Web3 aims to solve this problem is through something called interoperability — the ability to seamlessly carry and use your personal assets across applications, protocols, and even chains.

The theory is compelling, with one commonly touted use case being the possibility to use a CS:GO skin inside Fortnite, or to have a single character’s progression recognized inside two different RPG titles. By leveraging assets ubiquitously, interoperability would essentially bypass the “sunk cost” problem and let players exit a title they’ve outgrown, a stark contrast with Web2’s siloed state of affairs.

But things are more complicated in practice. To implement interoperability at the technical level, developers must first agree on a myriad of variables including industry-wide ERC standards and an asset’s visual rendering, metadata, utility, and scarcity. Business-wise, they may have very little incentive in giving competing IP access to their own player base and liquidity at the risk of cannibalizing their own marketplaces.

Still, we’re confident that rigorous coordination and a steady focus on users’ needs can successfully steer Web3 toward greater interoperability. Startups such as CryptoAvatars and Kinetix are already making that bet, respectively with cross-world avatars and emotes.

As we’ve aimed to demonstrate in this piece, the avatar space is experiencing great momentum. But where is it going next?

We believe avatars will likely only grow more prominent, as every digital destination starts using them to enable richer interactions — a social foundation for further engagement and monetization.

The creation stage is showing perhaps the most progress. A combination of hardware and software improvements are dramatically lowering barriers to entry, bringing previously complex 3D capture to smartphones everywhere. In June, Meta revealed that generating an individual Codec Avatar — which once required a specialized capture rig of 171 high-res cameras — can now be done with an iPhone scan. With enough time, and sustained R&D spend, even volumetric video and light field rendering could one day become consumer-ready.

In the meantime, AI tools like DALL-E, Midjourney, and Stable Diffusion are transforming entire creative workflows. Using only text prompts, creatives can already generate thousands of unique 2D avatars in the style that best matches their world’s aesthetic, and refine that output until it’s ready to be turned into 3D. Even the intermediary 2D step may soon be unnecessary, as text-to-3D is on the verge of a breakthrough.