Meta isn’t selling these things. That’s the point — at least for now. This week, the company’s Reality Labs team is bringing a handful of wild, clearly experimental headsets to SIGGRAPH 2025, and the demos are less “preorder page” and more “what we’d love to one day make practical.” If you want a quick shorthand: one prototype chases photographic realism, another chases the field of view of the human eye, and both make you painfully aware of what still stands between impressive lab demos and a pocketable product.

Tiramisu is the “hyperrealistic” prototype — the one Meta basically calls a milestone for realism. On paper, it’s bonkers: Meta says the headset reaches roughly 90 pixels per degree (PPD) in its center (about 3.6× Quest 3), brightness up to 1,400 nits (roughly 14× Quest 3), and around 3× the contrast of Quest 3. Those numbers are the sort that, in the lab, can trick your brain into believing what you’re seeing is real. But there are tradeoffs: Tiramisu is described as bulkier and considerably heavier than consumer hardware, and its usable field of view is narrow — a price Meta appears willing to pay to explore what true visual fidelity feels like.

If that sounds like an 8K TV stuck to your face — well, that’s close to the point. Meta’s researchers loaded the headset with the best real-time graphics they could: Unreal Engine visuals run with neural upscaling tech like NVIDIA’s DLSS to push frames at those crazy pixel counts. That means the demo is as much about software and rendering pipelines as it is about the optics and displays. But cramming all those pixels into a headset today requires heavy duty compute and some engineering sacrifices around size and comfort.

If Tiramisu is about per-pixel realism, the Boba family is about scope. The Boba 3 (and Boba 3 VR) prototypes pursue an ultrawide field of view — Meta claims roughly 180° horizontal × 120° vertical, which is getting close to what humans see when you include peripheral vision. That’s a big jump from the Quest 3’s ~110° × 96°. To make that wide view feel sharp, Boba 3 uses very high-resolution displays — Meta cites 4K × 4K per eye for these designs — and updated optics that let a lot more of the scene fill your vision without distorting it.

Those wider designs are grounded in more “production-ready” display parts than Tiramisu’s exotic glass — the Boba line leans on mass-production displays and lens tech closer to today’s Quests. The promise: an immersive view that better matches how we actually look at the world, which could reduce that “window-in-front-of-your-eyes” feeling VR still suffers from. But again: wide FoV + high resolution = a lot of pixels to render, and that has cascading costs in power, heat, and price.

Meta stresses these are research prototypes — experiments meant to map engineering tradeoffs, not to hit store shelves. But that’s important. Lab prototypes expose what’s possible and, critically, what problems remain unsolved: optics that don’t bulge the headset out of proportion, displays that are bright and contrasty without burning battery or melting your face, and rendering systems that can produce those pixels affordably. The work done on Tiramisu and Boba could feed into thinner, lighter headsets over the years, even if the exact designs never leave the lab.

There’s also an industry effect: when a big player publishes what it’s exploring, suppliers, developers, and competitors take notice. Display makers might prioritize micro-OLED or different silicon processes; GPU teams refine upscaling and foveated rendering strategies; game and app developers start thinking about content that benefits from wider FOVs and higher central fidelity. In short, the demos can reshape the roadmap across the whole stack even if you never strap a Tiramisu to your face.

If you actually put Tiramisu on, people who’ve previewed Meta’s footage say the “wow” is real — but so are the practical issues. Heavy headsets make long sessions uncomfortable, tiny useful FOVs can feel like looking through binoculars, and the render load needed to keep everything crisp can demand PC rigs and server-side help that rule out standalone, wallet-friendly headsets for the near term. And even if the optics get solved, there’s the simple reality that high-end hardware is expensive; not every leap in fidelity will survive the economics of mass market devices.

Don’t expect a Tiramisu on shelves next year. These prototypes are research directions: experiments in what’s worth chasing. That said, bits and pieces — better microdisplays, improved lenses, smarter upscaling, more efficient pipelines — tend to trickle down. So while your next consumer headset may not match these lab numbers, it might borrow ideas that make screens look sharper, colors pop more, and immersion feel a hair closer to reality. If you’re an enthusiast, the takeaway is hopeful: Meta’s labs are pushing hard, and some of those pushes will find their way into products over time.