A big area of interest for me going into the show was Virtual, Augmented and Mixed Reality, something I have taken to calling the Digital Multi-verse. As I mentioned in that post, this area seems to finally have some tangible commercial reality. It is an area that really gets people excited.
Much of the press and blog coverage of CES centered on the official availability of Oculus VR headsets. Many other companies were demo’ing their own headsets. The big vendors in this space – Oculus, Samsung, HTC and Sony – all had a lot to say about VR at the show with demos of their own.
For me, the interesting takeaway from the show was that with real hardware starting to reach the market, the practical realities are coming to the fore. If you want to buy a VR headset today, you are going to need some pretty significant computing power to drive it. The vendors have downplayed this, but if you look at the demos you can this is a real issue.
For instance, once company, Ant VR, had a very open demo with their rig. (Most others were doing demos in dark, controlled rooms behind closed doors.) In the picture below you can see Ant’s demo rig. Note first, that there is a huge cable connecting the user to his computer. I talked to one investor who thinks that the cabling market is going to have some good growth as the result of the very serious requirements for VR.
Second, note the white box in the background, in the middle of the photo. That is a desktop PC, with a clear case. You can see two green lines in the photo, which are LED labels for that’s systems Graphics cards. Here is a slightly blurry close-up.
The labels read GeForce GTX, those run about $500 a piece. Not every VR gear needs two cards, and the vendors say users can make do with (slightly) less expensive cards. Nonetheless, it should be clear that VR headsets are not yet ready for the average user to run on their old PC. Of course, this will change with Moore’s Law, but VR is always going to put pressure on the user’s graphics cards.
Still, VR has tremendous momentum. Enough, that I think the real focus for the industry is going to turn to content this year. Almost everyone I speak with agrees that the main driver for VR will be media consumption. In particular, this will mean games. However, I think it is also important to remember that this is effectively a new medium and that there is scope for considerable expansion of what ‘gaming’ is. One person who works for one of the large headset makers repeatedly corrected me when I talked about VR for gaming. He said VR is for ‘immersive experiences’. That includes gaming but will also include several new categories of digital interaction. You can wear a headset and find yourself sitting in the front row of a concert, for example. This is a lot of room for creativity here, and I think that we will be surprised by what content makes will create.
AR still has a ways to go
By contrast, there was very little to be said about Augmented Reality (AR). The people I spoke with repeated variations on my earlier theme, that AR actually has some very real technical challenges to overcome before it reaches its potential. First, the hardware is still too big and power hungry. The industry needs a lot of time to miniaturize everything. Moreover, the software challenges are still significant. In particular, the issue is that AR requires a lot of computation to match real world visual cues with digital data. This is not a task that can be easily performed in the cloud. If a set of AR glasses needs to wait for some remote data center to provide visual cues, the delays will create a lag between head movement and the alignment of data on the screen. In order for AR to not give users nausea and a headache, the system will have to compute locally. This, of course, leads back to the problem of having processors powerful enough and small enough to wear on the head. I am starting to think that it will take ten years to get True AR working. I recognize that is about the most pessimistic thing I have said on the subject, but we still have a long way to go.
That being said, the market is not going to wait for “True AR”, we are going to have real VR and “Mixed” Reality (MR) products coming to the market soon. And that means it is time to start thinking about content. Here, I saw a lot of promise. There were dozens of companies at the show displaying VR solutions for a host of applications – medical, industrial, energy, field force automation, and many more. The Digital Multi-verse is going to need an ecosystem. Everyone is so excited about the science fiction appeal of VR and AR, that companies are eagerly signing up.
I find this encouraging. So many people I speak to want to try VR or AR. During the week I spoke to many hardened electronics and gaming execs extolling big picture ideas. Even the most cynical think the potential for these systems is so big that they are getting excited to get started and at least dip their toe in the waters.
I think the next few years are going to see a slew of acquisitions of VR-related software. The chip makers are already adding new, interesting software features to their chips to facilitate their customers. This is true for VR and for drones and robotics. This is the industry at its best. The chip makers will take common software features, say drone automation, and put it in their chips freeing their customers from having to design their own automation software. In turn the module and device makers can then focus on other higher layer features like user interface and experience. It is a good time to be a machine learning company. Their expertise is in demand.