Build to Last: VR Sales Figures You Can Count on!

By December 16, 2015 Editorial
It’s with special thanks to Dr. Jon Peddie and Jon Peddie Research that they gave us the fundamental sales data needed to make this analysis possible.  More than this, even though this article was originally only intended to be published in JPR’s Tech Watch publication, they agreed to allow this article to be public as everyone could discuss, learn, and hopefully benefit from it.  We’ve also made the article available in Immersed Access so industry professionals can openly (though privately) discuss it and its ramifications if any to build for the future.

HTC Vive will be released WHEN?!?

HTC Vive will be released WHEN?!?

A big stink is being made about HTC delaying the commercial release of their HTC Vive until April next year, and others are worried that Oculus has already missed the mark by delaying until 2016 altogether. It’s unbelievable how so few understand the VR market and the core metrics that make it tick. Oculus and HTC could effectively wait a whole year, and it wouldn’t harm their bottom line or lay a dent in any content or business relationship they are working to build.

How could this be, you ask? What bit of data makes it possible for all the VR HMD makers to sit back and relax and get things right? You’ll see.

Oculus Rift DK1

Oculus Rift DK1

When the Oculus Rift first made the scene, they overcame a major hurdle by using dirt cheap smartphone panels as the core of high field of view high resolution VR displays. The technique effectively revalued VR devices that were formerly tens of thousands of dollars to just hundreds of dollars per unit in retail value. As major as this was, it’s a smaller piece of a much bigger challenge that is going to take years to overcome.

Oculus has been shying away from public sales commitments, though I think they are expecting big numbers given the hoopla they’ve received. HTC has been bolder in that they went on record that they expect non-smartphone revenue to account for 10% of their business next year (http://www.taipeitimes.com/News/biz/archives/2015/12/12/2003634579).

I’ve seen all kinds of media study blurbs and claims about VR. Everything from millions to tens of millions of units sold within the first year; most recently a study promoting the industry valued at $70 billion by 2020. With numbers like these, no wonder people think that VR’s future is going to be determined within the next six months.

The good news is everyone can take their time and do things right.

First, we have to rule out the sales figures for SDKs and developer kits; those Ocul-i and Valve-i mind-tricks won’t work here. Unless otherwise stipulated, figure that the first 200,000 in declared sales for all platforms are in fact SDKs sold or sent to hopeful content makers and press. In many cases, they could be purchased in multiples per location so developers have it in their hands. Whether they buy it or not, it’s a valueless part of the equation. We are only interested in actual customers that just want to enjoy VR and immersive tech and are not bound to the technology in any other way. The following numbers are based on this mindset.

Nvidia GTX980 AMD HD290 Series

To achieve what they consider the best VR experience on their HMDs, Oculus and HTC are each recommending an AMD 290 series or Nvidia GTX970 series graphics card or better. In fact, there are already signs that the Oculus SDK will detect if your graphics card isn’t up to spec and their software won’t work for you. So everything that performs less than these GPUs is out of the VR universe – settled.

According to Jon Peddie Research, 33 to 34 million add-in boards (AIB / GPU cards) will be sold in 2015. Three to five million of these are high-end / enthusiast graphics cards. The Nvidia GTX 970 / 980 was only announced September, 2014 and AMD’s R9 290 series was announced a year earlier, though AMD’s card wasn’t framed as a required-for-VR card until 2015. For arguments sake, let’s aggressively predict that by the end of 2016, 7 million high-end GPUs are sold, and this includes the qualified GPUs sold in 2013 and 2014. Factoring in that many gamer PCs use SLI or Crossfire, I estimate there will be three million gaming PCs powerful enough to play in VR. Whether it’s SLI / Crossfire or single GPU doesn’t matter; three million PCs is the potential universe for modern VR. If an optimistic 10% of these PCs own a dedicated VR device, the industry is looking at 300,000 HMD units sold within 12 months from launch. Not 300K for Oculus and 300K for HTC and 300K for other players coming to market…300K for everyone.

The SimpsonsIt takes anywhere from three to five years for today’s top grade graphics chips to be affordably mass marketable, and ten years for them to be inexpensive enough to work in a smartphone. This means that year over year, the VR market potential will have sequential growth, but it’s going to take time. I’d estimate that a potential market size would be about 20 million gaming PCs give or take within the next five years or so if the GPU flexibility is there.

The industry will likely have to come up with a recommended resolution limit on VR display hardware. Unless this happens, the GPU and display markets will never meet up at a mass market scale. The only way around this is if the GPU makers can arbitrarily surpass Moore’s Law (that circuitry doubles on a processor every two years), and that’s unlikely because Moore’s Law is struggling to keep pace as it is. This means that there is little to no room for major bumps in resolution from the HMD makers; too many pixels, and they are trapped in that 300K universe all over again.

There are cheats that can help. Combined with eye tracking, foveated rendering makes it possible for the GPU to only render the pixels you physically see, but this won’t be enough to turn the tide. It will also be tempting for content makers to use horrible 2D+Depth techniques instead of true stereoscopic 3D rendering to compensate for missing horsepower. 2D+Depth killed stereoscopic 3D gaming in the past, so the market should run for the hills if this makes an appearance.

DirectX OpenGL

We also have to consider that DirectX and OpenGL get overhauled every two to three years with new features. Future content makers will have to support legacy APIs or they are going to quickly alienate the mass market consumers by limiting themselves to the latest generation graphics options.

Console and mobile which are practically available now face different metrics and realities entirely.

Samsung Gear VR

Samsung Gear VR

Estimate that with nearly 22% of the market share, Samsung sells 300 million units a year. 20% of these are high-end, so that means Samsung’s Gear VR has a potential reach of 60 million phones. If they can achieve an aggressive 5% early adoption rate within 12 months of Gear VR launch, that brings it to about three million units sold. I think that’s very aggressive, but it’s not crazy.

ImmersiON-VRelia GO Google Cardboard

It’s actually more and I know the numbers won’t match up, but let’s calculate that there are around one billion Android phones sold each year. Estimate that 20% of these are mid to high-end phones that are VR qualified. If Google takes their Cardboard platform seriously, designs their own chips as pseudo-publicized a few weeks ago, and effectively flips the VR switch in all their compliant phones, they have a potential reach of 200 million Android phones each year including Samsung’s. Combined with their dirt cheap cardboard option and more premium clam-shell VR vendors like ImmersiON-VRelia, they could achieve a 10% adoption. So that brings their take to about 20 million Android phones within a twelve month period. When I say adoption, I mean people using the stuff and not just getting it in the mail or attached to their newspaper.

The challenge with mobile is that even though they have the sheer numbers potential, it’s not the same quality level of virtual reality as that of PC or console. So the VR industry is going to be a constant battle between the marketing forces, the artistic forces, and the limited forces of available processing power.

Sony PSVRConsole in every way, shape, and form has it made. I estimate that Sony PlayStation 4 will have 35 to 40 forty million units sold by the end of 2015. Thanks to their console’s fixed specification and their HMD’s committed resolution limit, it’s an easy sell. If there are 35 million PS4’s in the world, and if they could sell the PSVR HMD to 10% of them – that’s 3.5 million units sold in the first year with sequential market growth to follow. The genius of it is that PlayStation is suddenly in the display business as well as the console business. Even if they have a modest success, their hardware consistency makes for a promising future.

So what’s the moral of the story?

First, while we have a very exciting market to look forward to, it’s to no one’s advantage to rush products to store shelves to capture market share because…there is nothing to capture. The graphics market is not yet ready to handle the numbers of units that analysts are boasting about. The scalable technology doesn’t exist to make this happen.

At these early stages, I think it’s worthwhile for the PC hardware makers to consider selling their products for $1,000 a piece. They’ll make money on the mark-up, and the market segment that owns the required GPUs spend this kind of money anyway. This isn’t going to be a lower to mid-tier industry at the very beginning and this is the normal way of things. As the compliant GPU market grows, the display prices can drop.

Content makers are definitely going to have to get behind truly open standards and discussions because every HMD sale is important. It’s the only way their product distribution will add up in the early days. The exception to the rule is if they can somehow garner a lucrative exclusivity pre-paid sponsorship deal, but there are only so many cheques to go around.

PC products and product mind-sets need to be designed to work in universes sized at hundreds of thousands to a few million later on. The peripheral business has been doing this for years, and it’s very possible to flourish if the players have the right strategy. Content makers can succeed too provided they can thrive with a dedicated audience of a few hundred thousand people at the high end, thousands to tens of thousands at the earliest stages.

The professional, public exhibition, and academic markets will do very well. Anything that doesn’t require millions of units right away to make their business work will benefit from VR and immersive tech. I’m confident they will do even better than they were previously.

Don't Panic
Finally, for those concerned about when the Oculus CV1 or HTC Vive or PSVR or other promising option is released, or even developers who feel compelled to pledge their undying allegiance to a specific brand…relax. There is no meaningful advantage or disadvantage to launching first or getting tied down. The real market growth will happen further down the road – not this year or next. Just learn all you can, collaborate with industry friend and foe alike, and build.

Immersive Technology Alliance
These challenges exemplify why The Immersive Technology Alliance is so important. The industry is going to have to seriously get together in a safe and open forum so the market can figure these things out together and build towards a future that works. It’s the only way.

As Executive Director of The Immersive Technology Alliance (http://www.ita3D.com), I’m very committed to helping grow this market because it’s something I’m very passionate and excited about it. This is my career, and I want to be working in this field for a very long time. Oculus got their launch on Meant to be Seen (http://www.mtbs3D.com) after all. As with any industry, the first years will be the most challenging, and we have to proactively plan for this to succeed. I’m hopeful this analysis will help put the industry in perspective and financial strategies can be designed for the long term. With some planning, I feel confident that everyone’s hopes and dreams will become a reality; not just a virtual one.

Join the discussion 13 Comments


  • Avatar Neil says:

    If you want to critique the numbers or add considerations, feel free to share.

    Regards,
    Neil

  • Avatar Dom says:

    This all sounds good, both software and hardware sales will go up. Pc sales were declining I heard. This is nice that a these hmd should really cost a 10’s of thousands of dollars for the same thing.

    If these numbers are worldwide or just North America then I would be concerned. I would try and reach out to people on the apps store within windows 10 android and apple I phone ipod. Some kind of Candy Crush 3d game in VR might bring in a lot of dough.

    They still need to work on 3d chips for gfx cards to really push in the features like they are doing with recording and streaming.

    You also need to know that as the future rolls on I would think hmd designs will compact and be more ergonomic.

    And the cost of Gloves, walk mat sphere, controllers, ect..

    Once someone gets the hmd in place they will start buying everything else to accommodate it.

    I want to see these hmd in stores not just on the internet for pre order. I want to witness someone buying one in the same store as I’m in, I think that would be very cool to see how excited that person would be in that transaction.

  • Avatar Neil says:

    Hi Guys,

    I’ve been seeing external responses to this article, and I figured I’d just summarize some key observations and respond to them as best I can.

    1. This analysis doesn’t factor in people buying GPUs just for VR. It pretends that VR doesn’t exist.

    I actually raised this question with Dr. Jon Peddie early on. Isn’t VR going to stimulate sales of graphics cards?

    It’s possible there will be some of this. The sales numbers are purposely aggressive as far as how many units will sell out of the qualified user base, and it’s very aggressive to assume that nearly 50% of the PCs owning a new enthusiast class GPU are going to be VR Ready PCs.

    Remember these facts:

    – 3 to 5 million enthusiast GPUs are sold each year.
    – Only a portion of these GPUs are 290, 300, GTX970, and GTX980 GPUs. Not the whole lot.
    – This analysis arbitrarily says that a full 7 million enthusiast GPUs are collectively sold in 2015 and 2016, plus the HD290s and 970/980s sold in 2013,2014.
    – Factoring nearly 50% of PCs are in this unique VR Ready premium group is high. There is buffer room here I’m sure.
    – A 10% attach rate…one in ten gamers that have the right hardware buying an HMD…is very high. Most entrepreneurs like to say “If we get just 1.5% of the market, we’ll be rich”. 10% is proactively aggressive.

    EDIT: I pinged Dr. Jon Peddie this morning to get a quote on this:

    >

    2. What about people who buy an HMD before they own proper hardware?

    The reason the GTX970 and GTX980 GPUs are so important is because they now feature architecture that Oculus and other HMD makers needed. Performance is an important part of the equation, but basic architecture is a cut-off as well. AMD is able to get away with an older GPU because their tech already included this functionality. The same goes for PS4 and PSVR because they are using an AMD GPU which didn’t need extra features added.

    http://www.tomshardware.com/news/oculus-oculus-connect-vr-amd-nvidia,27729.html

    3. Are SDK sales an indicator for market success?

    I already ruled that out in the analysis. SDKs are business purchases for developers or press. They ended up being enthusiast purchases as well, though it wasn’t planned that way. SDKs are often purchased in multiples, so it’s not a one to one relationship between the hardware and the customer.

    This analysis is strictly interested in consumer purchases of consumer products. It’s a different world now, and we have to look outside our bubble – and it’s a tiny bubble.

    4. Steam Surveys say there are lots of qualified AMD / Nvidia GPUs in the market.

    The Steam Survey is a survey and is not representative of actual product sales. It’s a temperature check for a portion of their larger audience. Jon Peddie Research, which is arguably one of the most influential GPU analysts in the world, does quarterly reports on who has what market share based on fixed sales numbers given by AMD and Nvidia. Google him. There is very little guesswork in the data that Jon shares.

    5. Nvidia says there are 5 million Nvidia based PCs, etc. etc.

    Here are the actual slides:

    http://www.slideshare.net/NVIDIA/vr-base-camp-scaling-the-next-major-platform

    Jon and I saw this before this analysis was finalized. When it comes to marketing, Nvidia and AMD can position themselves any way they want. Our analysis isn’t a comparable representation of the number of VR Ready GPUs sold today – at least according to Jon Peddie Research.

    6. Reddit shows that the VR market is huge by having 60,000 members.

    No, it doesn’t. Between the countless account duplicates and its very limited growth day to day and month to month, the community has peaked out – at least for now. This is a very small community. Case in point, why haven’t the Oculus forums and the Reddit community multiplied since the Samsung Gear VR went on sale and sold out? Look at the number of Gear VR inquiries on the discussion forums; they haven’t climbed like crazy. It doesn’t add up. It will one day, but not yet.

    7. The GPU sales price will be lower by the time the Rift and HTC units ship.

    Not by nearly enough, and by then, people will want to save for the shinier GPU anyway. The circuitry on a GPU is double that of a CPU. Premium graphics cards cost as much as they do for good reason. They aren’t going to be mass market products until several years have passed in their life cycle.

    8. Yeah, but I’m going to show all my friends and we are going to sell a million within a year! Just you watch!

    This remark says it all far me:

    “…I just disagree some some of the premises. People who would invest as much money as either 970 or 290 and are that interested in PC gaming aren’t as ignorant as the general populous. I somewhat agree that people aren’t going to upgrade just for VR, but I also don’t think that only 10% of people who have the hardware to run VR will buy into it. If you can drop the money on a 970, you can probably afford one of the VR headsets”

    By George, I think he’s got it. The PC gamer is but a tiny fraction of the populous. The smart phone business is a mass market business with over a billion Android phones sold a year. Just one successful console is 30+ million and growing through a life cycle that could be as long as ten years – maybe more. A few hundred thousand gamers in either direction is meaningless against the industry expectations.

    If anyone thinks that investors are dumping tens of millions – perhaps billions – into this market with the hopes and dreams of reaching a million gamers, they are delusional. They want tens of millions of people using these devices, and unfortunately, some think this will happen next year. It won’t.

    Given the dependence on processing power and how long it takes to get from point A to B to C…it will take three to five years at the soonest – and that’s only if the display market sticks with resolutions compliant for a mass market GPU. It will be a worthwhile and prosperous road for those that design their businesses around realistically growing audience sizes, just not an overnight thing for mass market explosions. Don’t panic: it takes three years to develop a AAA game title from scratch. It’s ok to bide our time and learn from each other.

    That said, even though the market segment is relatively small to start, the gamers are valuable. According to Jon Peddie, enthusiast GPUs account for Nvidia’s growth and 50% of AMD’s. 300,000 for a new product launch with such regimented requirements is very reasonable. Provided the VR spec doesn’t change year to year, it will be a sequential growth.

    9. You’re crazy! The numbers are wayyyyyyy off!

    I’ve been biting my nails with this piece for weeks because I knew it wasn’t going to be popular from the get-go. It’s dramatically different from other predictions I’ve seen to date – no question. However, I’ve had it reviewed by several people who know their stuff, including top advocates that sell this stuff, and the feedback has been consistently positive. It’s ok to have challenges as long as they are identified and we work to overcome them. Better now while there is money in the industry and people can schedule their spending than to wait until it is too late and everything has run dry.

    Sure, things can turn out better than what I’ve shown here; nothing is perfect and this would be a great thing. For me, I’d rather everyone have the data so they could form their own opinions and make the choices that are right for them. I had to share this; if only to know that others had the same tools I did. Good luck!

    Regards,
    Neil

  • boomji says:

    Not grim nor wildly optimistic, so good on you Neil :anaglyph
    I’d imagine this giving GearVR & PS devs a warm fuzzy feeling.

    Totally with you on the “working through this together,Friend or Foe” bit.
    Congrats and have a great year ahead.

    behram

  • Avatar NumberSix says:

    To the rebuttal point 5:

    Does Nvidia really think the baseline gaming experience is 1920x1030x30fps? Even the consoles tout that they’re pushing 60fps at that resolution, and FPS fans and sim racers at least are fond of 120fps (and the sim racing.group in particular is fond of triple screens). IMO the gap between currently commonly-owned GPU hardware and the target GPU isn’t vast.

    Anyway, my take is the barrier isn’t the hardware. The key is the killer app. People will throw money at their machines if the must-have experience is there. Any number of good but not great titles will only cause a repetition of the “3d fad” meme from influential tech reviewers.

  • Avatar Neil says:

    >

    Nvidia’s hardware requirements go beyond just the performance. See point number 2. Of course there are different variables that will impact early sales – I agree. This is just some core data we didn’t have before that will help with the analysis.

    Regards,
    Neil

  • Avatar Neil says:

    >

    Thank you! Maybe that’s the crazy thing about all this. If the industry actually sells 300K give or take plus another 200K or some reasonable number for SDKs (for content making developers)…where’s the fire? The current expectation of what a success is relative to market dynamics in the PC world is way off.

    Regards,
    Neil

  • Avatar benplace says:

    I disagree with most of this. VR is going to drive sames of everything else.
    I have let probably 25 people try my DK2 and all but maybe 3 people asked how they can get one. My Dad asked me specifically after trying sightline the chair if he went out and bought a computer like mine could he get the head set. I told him in March he could.

    More sales of 980/970 cards will lead to less costs to manufacture which will lead to cheaper cards and more “Oculus Ready” shields on computers.

    You are completely ignoring the non gamer sales as well.. There are plenty of people that will buy it just for VR porn and adults that will want experiences, classrooms at high schools and colleges for education.

    Once they release a PC for 999$ with that Oculus ready logo.. Well let’s just say there are going to be a lot of people next year with one present under the tree…

  • Avatar Neil says:

    >

    Thank you for your feedback. This analysis isn’t about video games or video game sales. The GPU counts are for all the enthusiast level GPUs in the world. It’s not content or industry specific.

    I would of course prefer that you are right; this would make my year. $1,000 US for a computer plus $500 or so for the HMD is $1,500 (nearly $2,000 Canadian plus tax). If my sales count is a few hundred thousand out of the ballpark, it doesn’t matter. If I’m a few million out of the ballpark (and I really don’t think I am), it still doesn’t matter:

    http://gamerant.com/oculus-rift-sales-estimate-success-zuckerberg/

    The expectation has been set so unrealistically high – intentional or not – it’s madness. GPU manufacturers can’t just pump out cheap graphics cards because they want to; it takes time for the manufacturing technology to develop and the justified demand to build. I’ve wrestled with this too; it just doesn’t work the way we’d like it to work.

    If this industry thinks it will fall apart because it sells a few hundred thousand units to start (which is actually pretty good for an emerging tech as specific as it is), we all need to do some soul searching and figure out what the realistic expectations are. No real industry was based on a one year launch and wild profit plan. HDTV which we take for granted took 10 years. 3D television which everyone calls a failure was selling upwards of 50 million units a year and it didn’t need a powerful computer attached to it. This market probably needs a three to five year plan post-launch; that’s very reasonable.

    I think I said this to someone else: do we schedule our investments and build for the long term, or do we put all our multi-million dollar risk and personal livelihoods on the line and expect to drive our hover-cars by Dec., 2016?

    I’m planning to be around in five years, so I’m stepping off the hype train and driving my hybrid, thank you very much. It’s always strategically wise to plan for less and hopefully (hopefully) benefit from more.

    I’m sorry if I seem brash. I feel ethically concerned that the market has unrealistic and harmful expectations.

    Regards,
    Neil

  • Avatar Dilip says:

    My extract of article

    1) PlayStation VR has more chances of success initially
    2) Google can change the picture if they do develop their own SoC
    3) No company should panic about they are first to launch or not.
    4) Its still considerable years before VR can match popularity and
    abundance of smart phone.

    “Once they release a PC for 999$ with that Oculus ready logo.. Well let’s just say there are going to be a lot of people next year with one present under the tree…”

    Totally disagree $1000 is no where cheap by this price enthusiast pc was always available

    See $1000 Rig i pointed out Back as MAY 2015
    https://choosemypc.net/build/?budget=1000&oc=false&options=od

    here
    http://www.mtbs3d.com/phpBB/viewtopic.php?f=140&t=21117

    Trouble is to get it off totally we need much cheaper Pc and HMD that should be below $350 + Game like DragonAge/Assassins Creed/Battlefield/Mass effect totally supporting it at full glory.

    So its far …. agree with Neil around 4/5 years away for real boom.

    We may see drastic changes in Console and Mobile Scenario.

10 of 13 comments displayed


Log in to leave a Comment