Skip to main content

HoloLens 2 will get a boost from a dedicated custom AI chip

HoloLens and Microsoft Windows logo
HoloLens and Microsoft Windows logo (Image credit: Windows Central)

Work on the second iteration of the HoloLens is well underway at Microsoft, and the company has just teased a small part of what we can expect. Speaking at the Computer Vision and Pattern Recognition conference, Executive Vice President of Microsoft's AI Research Group, Harry Shum, revealed that HoloLens 2 will sport a dedicated AI coprocessor alongside its main silicon, the holographic processing unit (HPU).

The goal, Microsoft explains in a separate blog post (opens in new tab), is to give HoloLens 2 added horsepower dedicated specifically to implementing deep neural networks (DNNs). That will ultimately allow HoloLens to better tackle intensive tasks such as image and voice recognition without relying on sending data to the cloud to be processed.

The AI coprocessor is designed to work in the next version of HoloLens, running continuously, off the HoloLens battery. This is just one example of the new capabilities we are developing for HoloLens, and is the kind of thing you can do when you have the willingness and capacity to invest for the long term, as Microsoft has done throughout its history. And this is the kind of thinking you need if you're going to develop mixed reality devices that are themselves intelligent. Mixed reality and artificial intelligence represent the future of computing, and we're excited to be advancing this frontier.

There's no word on when Microsoft plans for its HoloLens successor might land, but reports from earlier in 2017 indicate that it won't be soon. Following a reported product roadmap shakeup that saw Microsoft skip what would have been its second-generation HoloLens in favor of what was to be the third-generation release, it's expected that the next version of HoloLens will arrive as late as 2019. If true, that's a long wait — particularly because the first version of HoloLens was initially shown off more than two years ago. However, such a wait would likely be justified by considerable bumps in technical capability, some of which will be derived from the new AI coprocessor.

A critical look at HoloLens and the future of Windows Mixed Reality

Dan Thorp-Lancaster is the Editor in Chief for Windows Central. He began working with Windows Central as a news writer in 2014 and is obsessed with tech of all sorts. You can follow Dan on Twitter @DthorpL and Instagram @heyitsdtl. Got a hot tip? Send it to daniel.thorp-lancaster@futurenet.com.

30 Comments
  • and will buying one be as hard and expensive as the first one?
  • You won't be able to buy one, they aren't releasing the 2nd generation
  • Instead of skipping the second generation HoloLens they should've given the design to a partner OEM and allowed them to produce a Windows Mixed Reality device. They should've done something similar to the Google Pixel/Nexus devices. I'm really excited about HoloLens, I just think Microsoft is moving way too slowly and risk losing the critical first mover advantage. If Microsoft doesn't think they can sell HoloLens to consumers then they should find an OEM partner who can market the tech to consumers. Microsoft can focus on enterprise "halo" devices and the MixedReality UI. Sitting back and waiting years for Apple/Google to catch-up and grab the consumer market seems like a defeatist plan. I assume they don't want to repeat mistakes like Tablet PC and SPOT watches (too little too soon), but they also run the risk of repeating a Windows Phone or Zune mistake (too little too late).
  • But since HoloLens is the only device, unique, in its kind so far, it still gives Microsoft an advantage. Yes mixed reality is a think and google almost monopolizes the market with gear VR mostly I think, but even Google doesn't have a augment reality device like HoloLens. If they make a device as unique and adorable for consumers then that's it, the game is theirs.
  • They don't have a device....that we know about at least
  • It's an advantage they are squandering by not putting something on the market for the masses to buy. At the very least, a gaming focused version for Xbox should be in the cards THIS YEAR. 
  • Google has the Google Glass which provides AR 
  • and don't forget Tom,  it's the size of a pair of glasses,  not a football helmet sized gadget.  I know what I will be purchasing.
  • Though Google glass's capabilities will likely be more limited at that size
  • It's not the only device. There are others like it that came out before the HoloLens. Check out Daqri, for example.
  • What if selling it to consumers is not part of their business strategy at all?  I hope not but, it's not out of question that MS continues to do what has always been their core...sell to the enterprise and whatever happens in the consumer market as a radiating effect is just gravy.
  • Microsoft has pushed back the consumer release of HoloLens until 2019 simply because the ONLY competitor on the horizon - Magic Leap - is failing all over the place.  No need to rush this one.  Take the time and get it right.  The Holographic Processing Unit (HPU) is unique and I can't see anyone even coming close to competing with this by 2019. An AI support chip might be interesting too.  I don't see MS making any mistakes here. Rushing a half-baked product to market would be far worse.
  • While I agree that ON PAPER, 'such a wait would likely be justified by considerable bumps in technical capability', that extra capability will be irrelevant when Microsoft is once again late to the party and has to fight Apple, Google etc. for market share. Smh, I guess they'll never learn.
  • Unless they dramatically improve the FOV and launch it in the market at a decent price, it's not going to matter.
  • I am confident the FOV will be noticeably improved. It will be interesting to see how things play out. Apple are going down the route of getting software developed first with ARKit that can run on dedicate hardware / glasses down the line. Where as MS have built advanced hardware that's not as accessible as writing code for iOS (by that I mean everybody has an iPhone, not everybody can justify a HoloLens). Google are down the route of heavily phone based VR that you wear on your face. I think it will all come down to how quickly Apple can get some AR Hardware out there. If they can do it in 2018 they'll have a years worth of AR Apps ready to run on their new hardware
  • "Apple are going down the route of getting software developed first with ARKit that can run on dedicate hardware / glasses down the line." Well, Microsoft is doing the same thing with Windows Mixed Reality VR glasses. Per Microsoft Store's website Developer versions of HP's and Acer's are expected to ship in August. https://www.microsoft.com/en-us/store/d/hp-windows-mixed-reality-headset... https://www.microsoft.com/en-us/store/d/acer-windows-mixed-reality-heads... All apps developed for those will be able to be used by HoloLens as well.
  • Hololens too have years worth of app already! They are just not as showy as snap chat;")
  • Google is actually working on Glass for enterprise. A few articles came out last week about it. Essentially google realized it wasn't a consumer product and have been developing it for the enterprise market over the past 2 years. Sound familiar? Already being used by Boeing, GE, Dignity Health, etc.
  • There isn't even a HoloLens 1 for normal consumers, so who gives a flying F about HL2? It will never go anywhere like everything promissing that Microsoft touches. Whats the point of inflated tech demos, when it leads absolutely nowhere. At least when Google or Apple announce something you can actually buy the finished product and use it.    
  • BINGO! Nobody is believing their smokes and mirrors or even believing some products that people want will materialize. This product started under Ballmer lets see what the Nutella man does !!!
  • Yep.  Besides a few kool-aid drinkers here now,  most people who were fans of everything microsoft have mobed on to bigger and better things....that get results instead of smoke, mirrors and coulda/woulda/shoulda's!
  • Because neither one have announced products that never came to light right?
  • Look at track records there axman...Action speak louder than words....and Microsoft's actions when it comes to consumer products is .....to put it politely........Shhhhhhiiiiitttttt!  SIMPLE
  • What track record does MS have when it comes to annoucing products that never come to light? If you're talking product support(which has nothing at all to do with my comment) what products did they stop supporting? My memory of consumer products that are no longer supported or discontinued are Zune, Band, and arguably Windows Mobile(phone), yet their list of supported products is decently long, so I'm not sure what track record you're talking about. My comment was about Apple and Google both annoucing consumer products that never reach the consumer. I'm not entirely sure there are many tech companies that never have that happen. If you want to go back to discontinued or no longer supported products, both Apple and Google have their share as well(not counting EOL software since every company that makes software has that).
  • You accidentally included a TradeDoubler link (to the blog post) in this article. Here's a clean link: https://www.microsoft.com/en-us/research/blog/second-version-hololens-hp...
  • I think the HoloLens dev kit includes some type of simulator that let's you test apps on a normal tablet. They should make that simulator part of the windows 10 os to let people run AR apps now, and then they might have a chance in 2019. 2019, some of us are going to be dead by then. 
  • By 2019, HoloLens 2 will have likely been beaten to market and end up another example of Microsoft's inability to push a market forward.
  • Hololens is gearing towards enterprise. Consumer, good luck about that!
  • Hopefully it's below $1000 this time.
  • Yes, yes, but when do we get to just plug this into our brains?