HTC announced a new product out of left field here at Mobile World Congress (MWC) this week, the 5G Hub mobile hotspot that includes a media streaming device, a digital assistant, and Internet access point for up to 20 concurrent devices. HTC’s growing focus on virtual reality has also led it to include what the company says will be one of the first devices to deliver on the promise of cloud-based app rendering for VR headsets.

Update (3:35 PM ET): In a previous version of this article it was stated that the demo used a 60Ghz 5G transponder, when in reality it was actually a full sub-6Ghz 5G network. More precise information was also added regarding expected latency.

Edge computing is an important part of the cloud-based gaming of the near future: ideally apps should be rendered as close to the person as physically possible to minimize latency across the board, including user input over the realistically assumed 100 Mbps 5G data transfer speed and near instant delivery of the compressed image to the headset. I couldn’t get much closer to ‘the edge’ today than HTC’s booth setup, which had a fully functional miniaturized mock-up of cloud-based rendering farm, sub-6Ghz 5G network, and a 5G Hub to wirelessly serve up VR content.

The promise is that—at some point during the 5G Hub’s product cycle—users will be able to stream SteamVR apps to VR devices that simply don’t have the compute power at their disposal. Prominently featured in the product reveal video was Vive Cosmos, HTC’s upcoming VR headset which connects to standard VR-ready computers via a traditional tether. When it was revealed at CES in January, Cosmos was teased to also be compatible with smartphones, and I think that particular piece of the puzzle is mostly answered by the inclusion of 5G Hub.

The headset’s local compute power doesn’t need to be much more involved than handing standard controller/room tracking and decoding the compressed image though, which hypothetically opens the door to low-power mobile headsets like Vive Focus Plus to play games designed for dedicated at-home, VR-ready gaming computers.

For the purposes of the demo space, which was absolutely teeming with crosstalk from a sea of 5Ghz WiFi signals, the company opted to tack on an external WiFi antenna with greater band ranged to the outside of the Vive Focus Plus for signal clarity, a little white box the size of a 5-pack of gum. I was assured the headset’s on-board 5Ghz WiFi antenna was up to task under normal conditions though, performing similar to the headset’s ability to locally stream SteamVR apps from an at-home gaming computer to the headset using HTC’s optimized version of the Riftcat VRidge software. In this case however HTC actually worked with AMD to use a cloud-optimized streaming solution in the 5G demo.

To that tune, I was treated to a demo of the cloud-based rendering setup using Vive Focus Plus running the immensely successful shooter Superhot VR (2017), a PC VR game that doesn’t natively support mobile headsets.

Image courtesy HTC

Under the pressure of the throng of onlookers, I can’t say I was exactly good at replicating my awesome red-guy beatdown skills that I had learned playing the game off and on since it was first released, although having the analogue of a game that I’ve played before made for an excellent point of comparison. That said, I wasn’t so much concerned with performing well, as I was pushing the limits of the biggest factor involved in the cloud-based VR rendering: latency.

Moving my head around the scene at a medium speed, black bars crept into the periphery of my field of view, the telltale sign that my head was moving faster than the hardware could serve up frames. It wasn’t like I was used to with less powerful setups though, as it’s clear the rendering hardware wasn’t skipping a beat. While the black bars crept in, the visible portion of the image was still fairly smooth, likely thanks to some aggressive time warping technique to generate synthetic frames to keep things moving along.

SEE ALSO
Hands-on: HoloLens 2 is a More Than Just a Larger Field of View

The setup was more of a conceptual demo, and isn’t close enough to a final product to talk latency specs, HTC China president Alvin Wang Graylin told me. Graylin fully expects to get an effective 20ms latency with predictive techniques, which in practice would be very close to the range the VR industry has mostly agreed upon to be an acceptably low enough latency to be unnoticeable. Graylin also told me that latencies might be longer depending on network conditions however.

Despite being an imperfect implementation, HTC is one of the first to show something of this caliber actually working. And while VR users will definitely need to hold off on buying a 5G Hub in hopes of playing VR games through the cloud soon—two entirely separate industries need to materialize before that can happen—it’s incredibly interesting to see the early steps in action.

Newsletter graphic

This article may contain affiliate links. If you click an affiliate link and buy a product we may receive a small commission which helps support the publication. More information.


Well before the first modern XR products hit the market, Scott recognized the potential of the technology and set out to understand and document its growth. He has been professionally reporting on the space for nearly a decade as Editor at Road to VR, authoring more than 3,500 articles on the topic. Scott brings that seasoned insight to his reporting from major industry events across the globe.
  • Proof XR Lab

    Very interesting to see an early deployment of this forthcoming technology. There is an edge cloud computing revolution coming as infrastructure is rolled out. Although as an early adopter of full fibre broadband (which seems a lifetime ago), i have no illusion about the time frame for functional edge cloud.

    My thoughts from EGX2018, in the image below:-

    https://uploads.disquscdn.com/images/fe155184ce4b0761bd428301fc1348e85ed7bfeaaae83c719b4f91b516ebd92a.jpg

  • kontis

    When all consumer computation is done remotely then consumers no longer own the computation. A quite dystopian idea.

    So far cloud gaming is not very usable even for traditional gaming (depends on genre) and the cost of hosting is high. How can they solve it for a medium that requires 3x-5x lower latency to be decent?
    I tried Witcher via Steam streaming on android phone (5 GHz AC router in the same room), the latency stats were quite good, but actual experience was mediocre with noticeable degradation of camera look responsiveness.

    • jj

      well my tpcast streams all of my hmds data and all of motion through, so that right there proves its possible… next.

      • Baldrickk

        There is a massive difference between a point to point high bandwidth dedicated link, as per your tpcast, and even Steam in-home streaming, which is local wifi (unless you can go wired).

        In my experience, in-home-streaming with a gigabit wired connection is good enough to play most 2D games without issue – the latency is low enough.
        Replacing one of those links with 5G Wifi, and you get increased latency – to the point where you have to anticipate your actions and perform them ahead of time.
        Examples –
        Mirror’s Edge, falling into a roll. Pulling the trigger to initiate the roll has to be done noticeably earlier. Usual timing leads to a hard fall with no roll – it’s too late
        NFS-Shift, + other driving games. General increase in latency causes rampant over-correction to be performed. In order to re-assume control, you have to force yourself to use very minor actions, and pre-empt corners.

        This is with non-VR games.

        Replace the wired link from the router to the PC with another wireless link back to a base-station, back to a server *somewhere* on the Internet, and increase the bandwidth requirements massively for VR….

        It’ll work, I’m not saying it won’t, but it won’t be the same experience you can get with local processing. And yes, I count your tplink connection as local.

  • Very interesting early work. But as you say, the world is not ready for this :)

  • fuyou2

    5G STILL NOT FAST ENOUGH!!!

    • jj

      its fast enough

    • It’s fast enough for an ultra-compressed and lackluster Wireless Streamed VR experience, and undoubtedly and unfortunate will drive a Revolution towards that sort of anemic experience because “convince” always seems to trump “quality” It’s bad enough that first adopters to VR think “The-Go” is a “True” VR experience with only 3DOF, and what not. Whilst I do agree that it’s necessary to get more peeps interested and adopted into VR, at least Quest is a good step towards showing off a decent mobile VR solution for those not wanting to spend an Arm and a Leg on the Current High End VR Spectrum! It’s just I sure hope that doesn’t become the Norm! Who’d want to live in a “Compressed” Virtual world?! And if that takes off and becomes the Standard, with exclusives and experiences that the masses adopt into that you can’t be a part of unless you also follow suit into that “Lackluster” experience, then well, we’re in trouble! HAHA! I mean I could easily see that becoming the case, but let’s sure hope that High End VR solutions don’t get overshadowed! Otherwise this will be like Crummy VHS Tapes being the Standard all over again!

  • Trenix

    VR is dead, just like the comment section.

    • jj

      then why are you coming to a vr specific site to comment in a dead comment section? Your actions prove that you’re wrong…. but thats obvious because youre just trolling

      • Trenix

        Because I’m waiting for any significant information of the new generation of VR and there is literally nothing. The only hope left is Oculus Quest, but that seems unlikely since it wont be as powerful as being tethered to a PC and will also be significantly heavier. Besides, who’s going to want to charge their headset every few hours or so?

        Seems like the last resort is PlayStation 5 being released with the support of a newer and better VR, since they have the highest selling VR headset currently. I don’t get what’s so hard with having a headset that is attached to a PC with slightly better resolution, eye lenses, eye tracking, and comfortable controllers. That and having a wireless add-on which can be sold separately. I’d easily pay $600-800 for such an upgrade.

        Eye tracking also can already drastically reduce the need of a high-end graphics card. I hope it’s just time that it’s taking to put such a device together, but honestly I doubt it. Mark Cuck rather focus on stupid projects like Oculus Go and Oculus Quests, rather than releasing quality products which people would be more likely to purchase. This is exactly why you don’t sell your business to mainstream companies, the outcome is always the same.

        HTC Vive may also randomly come out with something that will wash away the competition, but they seem too greedy for that to be possible.

        • jj

          yeah you’re off ur rocker….

          “Mark Cuck rather focus on stupid projects like Oculus Go and Oculus Quests, rather than releasing quality products which people would be more likely to purchase”

          this statement proves ur insanity.
          you cant force tech progression to go light speed. the quest is a HUGE step forward

          • Trenix

            How is it even a step forward? It doesn’t offer anything better than what is already possible. This product will still be worse than the Oculus rift which was released 3 YEARS AGO! It’s a slightly better Oculus Go, which was a failure because Mark Cuck is an idiot and is going for the wrong target market with this product, as if sales didn’t statistically prove it.

            Those who want the best mainstream VR headsets, have a computer capable of running it. I mean really, nowadays you can build a computer costing $900-$1,000 to be VR ready. Those who couldn’t do that, went with the Playstation VR where the vast majority of VR users are, based on sales. Playstation VR sold more headsets than all other headsets combined. This is already the majority of the target market that will not fork over more cash for an inferior product, unless they have money to waste.

            Now the only people left to target are those who probably aren’t gamers or simply can’t afford it. They’re most likely go with the Oculus Go or use their own phone as their VR headset, which is the minority of VR users. In fact, I can support VR with my phone but refuse to do so because it’s so trash when compared to a tethered PC VR headset. This is the market that Mark is targeting, he is a complete idiot. Clearly he didn’t learn with the Go and he’s trying again.

            Again, VR is dead and I can bet that Quest wont change a damn thing.

          • zeef

            Hmm, I think during this lull in VR technology, since you have nothing to do, you should consider taking some college courses, maybe earn enough money to move out of your parents’ place and grown out of your sense of entitlement. Heck, when you actually pay for your own mobile phone and service, you might actually appreciate that it runs VR at all.

  • That’sright

    I’ll say in 8 years we could have a decent virtual revolution but first headsets would have to come down in price, in size, in weight, should be wireless with better batteries, higher resolution, no screen door, higher fov than the binocular 100°, easy set up inside out tracking, eye tracking, and we need a real great full game made from the ground up for vr not just little short experiences, a good game can drive a system far.
    Should also combined with 5g to take off because the average consumer doesn’t have expensive gaming computers, 5g is the only way the average consumer will be able to play full vr experiences in this decade. Question is, is 5G fast enough to stream PC type vr?

    • It’s worth Fantasizing about! I mean, who know? It could be just so! Nobody knows for sure what the Future will hold! It’s more promising now for that to be the realistic reality forecast of Tomorrow, rather than some Sci-Fi based depiction of a Future with technology that just wasn’t viable for many many years! I mean Back To The Future trying to suggest we’d all have flying cars by 2015, lol. Well at least this “Run” of Virtual Reality is here to stay, and is being implemented in such a way that’s actually “Good” and not the crap that say the mid to late 90’s was promising, that degree of VR that died a quick and horrid death! I suspect that if things with this whole VR Revolution continue it’s course as is has, and if it’s truly like the Mobile Smart Phone Revolution that began with the I-phone ((well at least how the I-Phone brought such into the mainstream, granted they were not the first to release such products)) Then I can say the Future does look promising, since in just 8 years time, Smart-Phones have gone through such a HUGE change! I was skeptical thinking “Glasses-Like” VR would only be a thing way off in 10-15 years time, or even 20, but I’m quickly being disproven in my predictions as things like those “Hole-Punch” AR Glasses Demo being shown off capable of 4K level Resolution at 100 Degrees FOV is actually a working prototype! It’s only a matter of time that shinks down from “Concept” to “Product” If History is any indication of how things do tend to realistically transpire, well then all those “CES” prototyped things I had witnessed way back in 2012, actually take shape here now in 2019.. then yes, it’s very likely a lot of these technologies demonstrated as current protyped things are around an 8 year or less cycle for actual commercial release! That excites me, because it might just be only 5-6 years when these “Light-Field” based solutions being demo’d at tech Conventions now, are perhaps gonna be placed into real ultra light portable HMD’s in that said 5-6 years time window! Woot!! That’s something to certainly cheer about! Here’s to hoping! Like I said, I used to think that “Glasses-Style” Ready Player One like, Ultra-Light-Weight, Virtual Reality HMD was at least 10+ years away, I’m thinking, I’ve changed my mind, it’s highly likely it’s only 5-6 years away! I wouldn’t be surprised at all if we see anything very close to that in just 3-4 years from now! I’m mostly saying, it’s a hell a lot close than we all realized! It’s also improving at an accelerated rate! With so much competition and of course “Greed” driving production, that is everyone’s wanting to be the “VR-HMD-King” since it’s potentially the Largest Untapped Market in the near Future akin to Automobiles and Cell Phones! ((Billions-of-Users)) The Arms Race will greatly shorten the progress in bettering the quality of these HMD’s! Think like the Space Race! We managed to go from 0 rockets to the Moon in just 1 Decade ((The U.S.A. that is)) just because we were motivated to beat the Russians! It’s gonna be just that like! Forget Moore’s Law, forget all that, things for VR are improving at a higher than doubling in power every 2 years due to that fact, that folks are desperate to be the one’s to provide the HMD that everyone wants to buy / Covets!

  • Mateusz Pawluczuk

    People are already cloud gaming in VR by connecting Oculus Go to a cloud powerhouse via Paperspace and then installing ALVR on it :D So I am quite hopeful that this will be a feasible way to experience VR in near future – you can already play Hellblade: Senua Sacrifice in VR on a Snapdragon 821 which sounds pretty crazy tbh

    • jj

      good share!!!

      • Phyllis Buchanan

        Internet work opportunities have become a trend all over the over world these days. A newly released research tells around 75% of people are working in internet jobs from home with no issues. The Internet is thriving daily basis and therefore everyone is having an ocean of money making opportunities from the internet to make more money. Every person wants to hang out with his/her friends by getting out for any wonderful place in the country or any other country. So web-based income enables you to complete the work at any time you want and enjoy your life. But finding the right strategy and also setting up a right target is our milestone toward achieving success. Already lots of people are gaining such a decent revenue of $32000 every week by using recommended as well as powerful ways of making money on the internet. You can start to earn from the 1st day as soon as you visit our web-site. DISCOVER >>>>> http://ragv.im/cdl

      • Lashawnda

        On line work opportunities are becoming a trend in all over world nowadays. A current research tells beyond 75% of people are doing work in on line work at their home without any complications. The Affiliate marketing world is booming each day and therefore we are now having an ocean of work at home opportunities through the internet to get paid off. Everyone really want to spend more time with his/her best friends by getting out for any specific lovely place in the country or any other country. So on-line income enables you to carry out the work at any time you want and enjoy your life. Though selecting the proper direction and also setting up the proper objective is our end goal towards financial success. Already most people are making such a fantastic pay check of $16000 every single week with the help of recommended as well as successful ways of generating income online. You will start to get paid from the 1st day at the time you have a look at our web site. START TODAY >>>>> http://cummins.pt/ey

  • thewebdood

    I have been talking about the advent of “Cloud GPU” for some time. This has always been dependent on 5G+ speeds. When we minimize communication latency, we we reduce the necessity to maintain the compute horsepower near to the user. Ultimately, this will manifest as a hugely powerful, cloud-based, supercomputer with (most likely) dedicated groups of cores assigned to each user consuming its resources. An easy way to think of it is that when you establish a connection to the “GPU Cloud” an imaginary 1080ti spawns up in the air above your head. Your application sends telemetry data from your position and actions within the application and render jobs take place “in the cloud.” This is clearly some years out but it is good to see that progress in being made in that direction.

    • In retrospect… If a “Cloud-Computing” Hub thing ever manifested itself… Let’s pretend it has already for more current generation GPU’s for comparisons sake.. The “Feasible” result would be more or less a GTX 1050 level of graphical power for each person subscribed to whatever “Cloud-Based” VR Game / application.. It’d be like WOW, where there’s a Monthly Subscription that helps cover the cost for this large data farm of hardware / servers via the Company / Studio providing access to such and giving each subscribed player a GTX 1050ish level of power at $30ish bucks a month, possibly even $14.99 Subscription ((For-Basic-Sub’s)) Of course they’d then offer “Premium” Cloud VR Computing at say $24.99 – $29.99 / month with a dedicated GTX 1070, and finally your “GTX 1080ti” edition would unfortunately be the “Ultra-Premium” package for $39.99 / month! Let’s face it, no company serving up Dedicated Cloud Based GPU Farms, one for each “Client” on it’s server’s ends, is gonna shoot that high, the costs would be way too much for them to reasonably do, and of course they’d know the limits and minimum specs these HMD’s run at, and will undoubtedly provide the lowest tier of those for cost effective solutions..and since the Oculus Rift now only needs to target 45 fps in games with the addition of their “Asynchronous-Space-Warp” Interpolating half the frames to reach that sweet 90fps much like how Motion-Flow works in Modern High End Smart T.V.’s adding new frames to smooth out Video Judder / input lag. With that in mind, peeps putting such advances and Networks into place, they’ll opt for a lower tier GTX 980 – GTX – 1050 side of things, so don’t be expecting Super High End Perfomance, it’ll be a LOT better than the GPU capacity of say The Oculus Go’s Qualcomm, Snapdragon 821, Processor in mobile VR solutions, but it won’t be anything near what a person with a Powerful Desktop or Labtop VR Capable Rig could do with a Powerful GTX 1080 – RTX 2080 level card! Not for many many years at least!

  • DonMac

    Steaming from a high powered PC in your own house to your HMD and Rendering in the cloud are two very different animals.
    Rendering games in the cloud isn’t even new, companies have been trying to get us to ditch our top end rigs and play non vr games from ‘thin clients’ for several years.

    One big hitch is these services are almost universally ‘pay as you play’ and many consumers (and some developers) are just not comfortable with this business model or see the benefits to it.

    Viveport Infinity’s new access to all content model (like Netflix) may be a clever plan to move us all to that way of thinking, possibly combining renting us not just access to the games but the online rendering to power them on the go.

    At the moment the weight of the PC player base is of the mindset that “I bought it, I run it, it’s mine”, steering existing power users away from their investment and mindset may prove as difficult as getting past the technical barrier of getting over 50fps in VR.

  • JesuSaveSouls
  • Lol, the Article clearly states “Stream The Compressed Image.” UGh! So here we’re at it again!! If the Death of Physical Media, A.K.A. Blu-Ray Discs and Blockbuster Rentals of such over the Convenience of Streaming via Netflix and what not, wasn’t one step forward, three steps back, enough! Now we’re heading into a “Compressed” VR Reality!! I’m already feeling a tad “Short-Sided” from the Lackluster quality of always in this “Compressed” window into the Interwebs… It’s tolerable for Youtube and what not, because at least some content creators do a much better job with their Camera setup’s than others, and of course it’s just Youtube, where I’m not entirely expecting A Micheal Bay / James Cameron level Film Experience! It’s just they quit putting Disc Drives in most even High End Gaming Labtops, computers and what not, at least when you “Stream” games off of Steam, most the large 50GB+ level of Detail in a game is saved directly to your Hardrive, and thus textures, eviornments, and what not in Game all look Fresh and Sharp, without being “Compressed” to death! Whereas Steam’s mostly thee to just throw you a Digital Copy of a Game, but they do so in Full, unlike Netflix, where when you downlaod a Movie for Offline Viewing or even purchased Films on Amazon or Google Play Videos, are all Small Compressed Lossy Files! A whole Movie that Should have been around 25-50GB in size, is 1-2 GB! I get that compression has come a long way!! Sometimes with HEVC 2.65 Video Codec Standard of late, you can get really impressive Compression results! It’s wonderful for say me throwing a Full Movie onto an SD Micro Card for viewing on the go in my Gear VR! HOWEVER, I refuse to live in a Future, where all Virtual Reality Experiences are mostly this Dull-Downed, ultra Compressed thing, all for the sake of Wireless freedom, cheaper devices / device components in HMD’s, and overall convenience over quality factor! But we all know, in most the format wars of Yesteryears, Cheap, low quality usually wins the day, and most will follow suit, jumping on that “VHS-Tape” bandwagon giving us a 4:3 aspect Ratio Image that’s been “Reformatted to Fit your screen” unlike the original Theatrical Video versions for Decades on end, before finally 20-30ish years Widescreen Full HD Video at 16:9 Aspect Ratio finally becomes a thing! The same will happen, unfortunately for VR! Where 20+ down the line, folks will finally say “What’s with this awful compression?!” And finally go back to a “Raw” lossless pictured Format as it should have been to begin with!