Visual generation was one of the first AI capabilities to become familiar to a broad audience, so another impressive image no longer feels like a major conceptual change. The more interesting development now is how quickly AI is moving into 3D creation, rigging, animation, motion capture, world generation, novel view synthesis and other parts of immersive media.

This is important because 3D production was one of the largest barriers during the previous metaverse cycle. The vision often arrived before the production economics made sense. Creating convincing interactive worlds required expensive modeling, specialized animation skills, game development knowledge and a large amount of manual work.

The old 3D bottleneck is being attacked from several directions

Text-to-3D and image-to-3D tools can create starting geometry from ordinary prompts or reference images. AI-assisted rigging and animation reduce the amount of specialist work required to make a model move. Motion capture can increasingly be derived from ordinary video, while generated environments and novel view techniques make it easier to create spaces that can be explored from multiple angles.

Each capability is still imperfect, especially when a project demands production-grade geometry, consistency, physics, optimization or precise artistic control. The direction is still powerful because several expensive parts of the workflow are becoming easier at the same time.

The cost of creating a rough but convincing 3D idea is dropping quickly. That allows more people to create prototypes, scenes, interactive environments and visual experiences before they have access to a large production budget.

The metaverse may return through production capability

The word metaverse collected a lot of baggage during the 2021 cycle. Facebook changed its name, speculative capital poured into virtual land and tokens, and many projects promised experiences that were still difficult and expensive to build. When the financial enthusiasm faded, the word itself became unattractive in many professional conversations.

The underlying direction never required the word. Persistent 3D worlds, interactive environments, spatial computing, games, digital twins, virtual production, avatars and immersive media all continue to develop. Many of those areas are now receiving the production technology that was missing during the earlier cycle.

A second wave of immersive computing may therefore arrive through practical creation tools rather than through a single grand metaverse narrative. If individuals and small companies can produce high-quality 3D content cheaply, the amount of available spatial content can grow far faster than it did in the first cycle.

Digital identity is becoming programmable media

The same technology that creates worlds also changes how people can represent themselves inside them. Faces, bodies, voices, movement, facial animation, clothing and persistent characters can increasingly be generated, edited and reused across different forms of media.

This goes beyond the familiar deepfake discussion. Synthetic identity can become a production tool controlled by the person whose identity is being represented. A creator can establish a visual likeness, voice and set of mannerisms, then use those ingredients to produce content without performing every version manually.

That creates a difficult cultural boundary because the output can be synthetic and still represent a real person's approved identity, opinions and instructions. The legal and transparency rules around this will continue to develop, especially in Europe, where disclosure and identity rights are already important areas of regulation.

Your digital twin can become a production asset

A practical version of the digital twin is easier to imagine than a science-fiction copy of a person. You record enough reference material to establish your appearance, voice, movement and mannerisms, then use that representation as an asset when creating videos, presentations, localized content or immersive experiences.

Instead of recording the same message repeatedly, the person can define the script, setting, language and format, then generate approved versions from the established identity. This could become especially useful for routine communication, education, product demonstrations and multilingual content.

The goal will probably be consistency rather than perfection. A useful digital version should feel recognizably like the person, stay within the boundaries they approve and remain controllable enough that the owner can decide where it appears and what it says.

Immersive media changes what a camera means

There is also a more speculative direction that becomes easier to imagine as 3D reconstruction improves. Traditional video gives the viewer the camera position chosen during production. A spatial version of the same scene could eventually allow the viewer to choose where to experience it from, moving around the action instead of accepting a fixed viewpoint.

A movie scene could be experienced from beside the car, above the action or from another position inside a reconstructed environment. This is far from an ordinary movie format today, although the ingredients behind it, including volumetric capture, novel view synthesis, dynamic scene reconstruction and generated 3D environments, are developing quickly.

The broader change is that visual media can become less dependent on a single recorded frame. Once a scene exists as a spatial representation, it can support viewpoints and interactions that were impossible in ordinary video.

Worlds, avatars and digital twins begin to connect

Cheap 3D creation becomes much more interesting when combined with programmable identity. A person can have an avatar or realistic digital representation, enter a generated environment, speak through a cloned or synthesized voice and interact with content that can be created or modified through AI.

This creates a continuum between games, virtual production, social worlds, product visualization, digital twins and traditional media. The categories may remain separate for business and cultural reasons, while the underlying production tools increasingly overlap.

A virtual showroom, a game level, an architectural twin and an online social space can all use related techniques for geometry, animation, physics, identity and interaction. As those techniques become more accessible, the distinction between a professional 3D studio and a small creator's toolkit can shrink substantially for many everyday use cases.

The missing ingredient was creation cost

The first metaverse cycle asked people to imagine enormous digital worlds before the industry had cheap ways to fill them with enough good content. AI is now reducing the cost of several parts of that process at once, which changes the economics of immersive creation.

This does not guarantee that people will suddenly spend their lives in virtual worlds, because adoption depends on culture, hardware, social behavior, product quality and many other factors. It does mean that one of the largest production barriers is weakening quickly.

That is why I expect immersive computing to return in substance even if the metaverse label remains unpopular. The technology is moving toward a world where creating a space, a character and a convincing digital identity becomes normal creative work rather than a highly specialized production project.