They're using Voxels to render everything instead of polygons. The problem is that you can't animate them. If they can use a hybrid system it can actually be a very good system. But as is it's very useless.

yeah, saw that 60 minutes video, they should release that tech demo for the public, i want to see it

The future is now.

nice i saw this before but not the interview. can't wait for this to come!

Indeed, it's very old news. They've been plugging this "unlimited detail" stuff for over a year now, but nothing concrete has been produced.

Of course, they've been accepting generous "investments" during that time.

Oh, it's porbably real. And it probably can be applied to games.

But only visual novels.

Anything that involves movement seems to be out of the question.

Over a year? These guys have been displaying voxel rendering for nearly a decade now. Their intents might be genuine, but they are so shady and dodgy it's borderline scam. Besides, there are obvious memory and computing limitations to voxel engines you are extremely unlikely to see it long, short or medium term, if it can be done at all. Maybe,MAYBE something with static backgrounds a la REmake, but that's it.

Is it that long?

I thought they'd been up to this for longer as well, but the company's wikipedia article says that they were founded in 2010.

Did they accept "investments" under a different name before that?

I have seen this demo in various forms at various times. It is nice that you think this is an "engine" - but it is not. It is a technique that most experts think will simply not work with animated stuff (or would need impossible amounts of ram/horse power). I do not know enough about this technique to judge its merits, but if most experts say it simply won't work in the "real world", then I go along with that opinion. Maybe all experts are wrong, or are paid by the competitors <insert ypur private paranoia here>, but at this time I go with all the experts in the field. (it's not like nobody ever tried voxels).

Two thirds of your text is just wishful thinking, though. There is no "optimization" possible - again, this is not "an engine", it is a technique.

Forgive my ignorance, but how exactly can these 'atoms' not be used for animations? If you were to have a rendered amount of atoms on screen all moving in the same direction/rotation, could that not form an animation?

I can understand physics calculations and things of the like being complicated to think of, but it's my impression that they render this by the pixel. To me, it means that they're not rendering trillions of atoms at once, only what can be shown by the pixels on the screen. It might explain why they can render such static detail at incredibly long distances without LOD.

I asked myself the same question, so I went digging into this.  The whole problem with this isn't that you can't animate it, it's that the data structure they're using to hold this information that's the problem.  The data is stored in a tree, if you have to animate something you have to either regenerate the tree in real time (Which is quite taxing on a cpu) or you hold multiple trees for multiple animations in memory (Which there isn't enough memory for).  They might be able to make hardware for the tree generation to speed things up, but that might not be cost effective and the hardware vendors have to take a chance with it.  Though hardware vendors went with tesselation instead, which one ends up being better used, we'll have to wait see.  Though I think gaming will ultimately switch to voxels, the cpus and memory just isn't here and may not be here for a while.