304 North Cardinal St.
Dorchester Center, MA 02124
304 North Cardinal St.
Dorchester Center, MA 02124
[May 5, 2020 – Added a small section on direct retinal scanning (ex. North Focals)]
In February 2019, when I realized that Microvision’s technology was definitely going to be used in Hololens 2 (HL2), I bought what was for me a tiny position (less than 1% of all my investments) in Microvision stock that I still hold. I think when you read the rest of this article, you will find that it was not enough to influence me favorably.
Additionally, I am working for RAVN as CSO, a company working on Military AR Headsets. While the HL2 is being used in Military AR applications, RAVN is developing a different type of product with a different use case. The opinions here are my own and not those of RAVN.
I also want to be clear that I have no inside information on Microvision. All my conclusions are based on public information and my years of studying Microvision’s and related technology.
Monday, May 4th, 2020, Microvision stock jumped to 4.4X in a week from 25 cents on April 28, and closing at $1.10 today (and closed at $1.46 in after-hours trading). It looks like someone has the hots for Microvision. I should warn, this could still be pure stock speculation.
From the outside, Microvision is looking like a “walking dead company.” Some of their key people were hired away by Microsoft in the last few years. In January at CES, they said they had a customer for a down-shooting projector based their technology only to announce a few weeks later that the customer pulled out. Then they announce a change in CEO. Then on April 6th Microvision seemed to hang out the “For Sale Sign” by hiring Craig-Allum Capital Group to look at licensing and “other strategic alternatives.”
I’m going to my stock- and business-related observations at the top of this article. I will then follow with links to articles in my blog for those that want to know more about the history of Microvision as a business and their technology.
I wrote all this in a hurry to get it out before the morning of Tuesday, May 5th, 2020. There was little time for editing. You have been warned, this was a long rush job there will be lots of typos and maybe at times a bit of a stream of consciousness. I may or may not clean this up later if I have time.
Microsoft is one of the main suspects as their Hololens 2 is using laser beam scanning. It does appear that Microsoft got a license to make the technology and pay Microvision a royalty. Microsoft also hired away several top technology people from Microvision. The contract almost certainly would be written to survive any takeover. It is also not clear that Microsoft is firmly committed to laser beam scanning technology in the long term and are looking at other technologies.
STMicro, which has partnered with Microvision for several years now and who manufactures Microvision (and likely the Hololens 2) mirrors, is the other likely suspect. STMicro is already a major player in MEMs devices for various applications. They bought Btendo, another laser beam scanning company, before agreeing to a joint development deal with Microvision.
Then you have “all the usual companies” including Facebook, which just bought MicroLED Plessey but has published research papers that use laser beam scanning. Just about any big player thinking about getting into AR, including Apple and Google could buy Microvision with petty cash. Bosch was showing an LBS AR set of glasses at CES and might want to bid. In short, just about anybody in AR could bid.
It could also be someone looking at Microvision for some other purpose, say Lidar. I would remind people that Microvision has been pivoting since 1993 and has yet to find a use for their laser beam scanning that made money.
Then there are patent trolls that might just want to see who they might be able to cause trouble with Microvision’s patents.
To be clear, I only bought Microvision stock because I figured that the Hololens 2 involvement would cause people with more money than knowledge of displays to buy into Microvision. Just read the rest of this article and anything this blog has written about laser beam scanning. I think the image quality is always going to be miserable with laser scanning. There are just so many limits with an electromechanical scanning system. There may also be issues, both real and perceive, with eye safety using laser scanning.
Most of the “smart money” in the AR industry believes that MicroLEDs are the future. But as I note in the next section, the pupil replicating waveguides, such as Hololens, is using don’t work well with MicroLEDs (this is surprising people).
I was planning on writing a lengthy article on why several companies, in addition to Hololens 2, are looking at using laser beam scanning. It involves a lot of physic and will take more time than I have right now to explain, but it is a real issue. It has to do with etendue coupling into the waveguide combined with massive light losses associated with pupil replication.
The “shocker” is that MicroLEDs, which many expect to the AR display technology of the future, with several million nits are not bright enough for outdoor use. A saying I often use is that when smart people do something dumb, it is because they found the alternative to be worse.”
LCOS or DLP illuminated by LEDs can be bright enough and will cost much less and have better overall image quality. Some companies are claiming to have a new generation of LCOS that is much better (I haven’t seen it yet). But even with these technologies, it gets hard to be bright enough without using too much power.
The critical point here is that in the last few years, there have been several companies with pupil replication waveguides that are starting to look at laser beam scanning. But still based on my knowledge of laser scanning displays, “if the answer is laser scanning display, then you are asking the wrong question.”
I should add that I am not sure that Laser Beam Scanning will work for outdoor use either due to eye safety (I don’t know).
[Added May 5th, 2020]
There are two main types of laser scanning displays used in AR thus far, pupil replicating waveguides like the Hololens 2 and direct retinal scanning as used by North Focals. Prior to the HL2, most if not all laser scanning near-eye displays used direct retinal scanning.
In October 2018, I wrote North’s Focals Laser Beam Scanning AR Glasses – “Color Intel Vaunt” I wrote about the many drawbacks direct retinal scanning. The most noticeable drawback is the exceptionally small “eye-box,” which simply put, is how much the eye can move and still see the whole image. With direct retinal scanning, if the eye or glasses move even the slightest, then the image disappears. Glasses have to be custom-fitted to each user and even then can be hard to even see an image.
The HL2 (and now some others) does something called “pupil expansion,” which is roughly speaking a tiny rear projection screen, followed by (massive) pupil replication in the waveguide with a bunch of other optics in between (see https://www.kguttag.com/2019/02/22/hololens-2-combining-two-bad-concepts/ for more on the HL2 optical path). This process results in a very wide eye box and supports much larger fields of view, but it looses the “infinite focus” (also known as Maxwellian) effect of direct laser scanning.
Hololens 2 is pretty much an R&D project that escaped the lab. It is costly and the image quality is terrible, and neither of these major problems is going to get better soon. There is some real application for it in industrial settings, but we are probably talking a market of maybe 100,000 per year. Remember, Hololens 1 reportedly only shipped 50,000 units in it first two years or only 25,000/year.
I know a big deal has been made about the Army’s IVAS program and the $900M budgeting for ~40,000 units (about $22,500/headset). But you need to read the fine print. The 40,000 units are the maximum and they were to be delivered over five years (avg. 8,000/year).
It comes down to a problem that there are not enough units to generate a lot of revenue for licensing the laser beam scanning display. If Microvision gets $10/eye or $20 per unit, that is still only $2M/year at 100K units. It is highly doubtful that Microsoft would agree to more than about $100/headset, which would result in $10M/year. You can play with these numbers, but to get the volumes up, the prices and royalties have to come down. Also, there is no way this is going to get anywhere near consumer quality or pricing in the next decade from everything I know.
On top of all this, Microsoft has had massive manufacturing, yield, and quality problems. Most of the rest of the market, including Facebook and likely Apple, are looking at MicroLEDs as the future. I know that Microsoft has to be looking at alternatives given their experience with the HL2.
I probably first saw a Microvision headset display, it was red-only back then, in the late 1990s when I was working on my first head-worn display. From 2007 to 2011, I competing with Microvision working on Pico Projectors as CTO of the LCOS company Syndiant. While at Syndiant, I did some in-depth technical studies of laser beam scanning. I learned that laser beam scanning is one of those concepts that looks and seems simple until you understand it. There must have been over 20 significant companies, including no less than Samsung and Sony, that put serious efforts into laser beam scanning that all failed to make a successful commercial product. Every year or two, Microvision gets an end customer, including Sony and Pioneer, as well as many small companies to build a product using Microvision devices, all of which quickly failed.
This blog has been writing about Microvision’s laser beam scanning (LBS) technology since 2011, and there are about 40 blog articles mentioning Microvision. Over the last 8.5 years, this blog has explained how laser beam scanning works and the many technical issues with it. One of this blog’s earliest articles, Cynic’s Guild to CES — Measuring Resolution, explains many of the core problems with Microvision’s bi-directional interlaced laser beam scanning. While Microvision’s newer design, used in the HL2, uses two mirrors, it still uses bi-direction interlaced scanning with all the problems outlined in the “Cynics Guide” article.
Microvision is a “27-year old startup” that went public back in the dot-com days of the 1990s when a company could go public without any profits or revenue. Microvision bought the patent rights to The University of Washington’s laser beam scanning technology developed in the early 1990s. Early on, Microvision lived on earmarks given to them by congressmen, including Senator Gordon Slade, that conveniently ended up on the Microvision board. BTW, those earmarks were for a Nomad Headset that the Army refused to use (see $4.5 million for a boat that nobody wanted and scroll down to the section “Battle gear in boxes”).
Microvision’s primary source of income was to use its listing on Nasdaq to print money in the form of stock. To date, Microvision has total shareholder losses of almost $600M (OK, chump change compared to Magic Leap private equity of $2.6B). It seemed that every time the company was on death’s door every year to two), Microvision would announce a new business deal. The deal was described cryptically to make it seem like a big deal. These announcements were followed by printing more stock. When the deal when south, it usually just quietly went away, and Microvision stopped talking about it. The pattern repeated so many times that the Microvision stock message boards talked about “Lucy and the Football” from Peanuts. The company routinely lost about $1M or more per month for at least the last ten years, but at least the executives of the company didn’t miss a paycheck, and they kept getting stock options.
Another significant source of Microvision spending was on patents, mostly related to laser beam scanning. The patents Microvision licensed from the University of Washington have long ago expired. Even many of the patents Microvision filed have or are about to expire. Any existing Microvision LBS related patents are on improvements they have made, but the basic LBS concepts and nearly twenty years of developments are now in the public domain. Some of Microvision patents are for product concepts that have failed. Since there has never been a successful volume product using laser beam scanning projectors, there has never been any significant potential revenue from their patents.
One of the newer product areas that Microvision seems to stress today for their laser scanning is Lidar for both gesture recognition and automotive autonomous driving. I have not looked at this aspect of Microvision seriously. I would note that the Lidar field is very crowded with many companies that have been seriously Lidar longer than Microvision. So I wonder if Microvision has something that these other companies would need to license.
As I wrote on February 22, 2019, Hololens 2 is Likely Using Laser Beam Scanning Display: Bad Combined with Worse, it was becoming clear that Microsoft was going to use Microvision’s Laser Beam scanning in the HL2. As the title suggests, I was not expecting the image quality to be very good, and this has since been born out to be accurate. The news of HL2 using Microvision caused Microvision’s stock to jump by about 2x to about $1 before drifting back into the $0.70 to $0.95 range for the next year. The pandemic caused the stock to fall as low as 21 cents and a market of about $30M on April 9th, 2020.
Holelens 2 was announced on February 22, 2019, and I followed up with Hololens 2 First Impressions: Good Ergonomics, But The LBS Resolution Math Fails! In this article, using my knowledge of laser beam scanning, I went on to predict that they were using interlaced scanning, which has since proven to be true. I went on to explain that interlaced scanning would cause problems, which it does.
The big surprise was that despite it being obvious that the HL2 was using Microvision, Microsoft, to this date, has never mentioned Microvision. Furthermore, Microvision will not admit that they are working with Microsoft. This game with Microsoft and Microvision became a farce when sharp-eyed people on Reddit found the Microvision logo in a video showed by Microsoft’s Alex Kipman. Kipman repeatedly said that Microsoft developed the laser scanning technology as I reported in Hololens 2 Video with Microvision “Easter Egg” Plus Some Hololens and Magic Leap Rumors.
It was a long wait from the February announcement to when Hololens 2 was shipping very limited numbers of units in November 2019 which I reported on in Hololens 2 Is Shipping, But Not Really That Many. The reports I had was that Microsoft had terrible problems making the laser beam scanning engine and then aligning it with the rest of the optics. Soon reports were coming out that the image quality was awful on many units (see: Hololens 2, Not a Pretty Picture and Hololens 2 (HL2): “Scan Lines” Making Text Hard to Read and Quality Issues with Waveguides).
In January and February 2020, I got a chance to try on an HL2 and wrote about my experience Hololens 2: How Bad Having Tried It? The image quality is about what I was expecting. The color uniformity was very poor and small text that should have been readable was hard to read. Furthermore, it clearly has a problem with flicker.
It looks to me that the HL2 could be useful for some industrial applications. Still, the image quality is always going to be horrible as it inherent with combination of laser beam scanning and Microsoft’s diffractive waveguide technologies.
In the first 2 weeks of this blog, I started talking about green lasers and participated in some heated discussion on Investor’s Village on Microvision (which predated the Microvision Sub-Reddit). I was saying was no way that direct green lasers capable of supporting laser beam scanning were going to be ready for mass production in 2012. Just two weeks after this blog started, Microvision felt compelled to issue an 8K to the SEC and a company blog article obliquely referring to me as a “false soothsayer.”
As 2011 comes to a close I wanted to take an opportunity to update you on the topic we overwhelmingly receive the most questions about – direct green lasers (DGLs). Everyone wants us to tell them when they will be available! Lest you be led astray by false soothsayers, based on our periodic discussions and latest updates from three direct green laser developers we anticipate that Nichia, Osram and Soraa will release commercial versions of their lasers in 2012 and two of the three should have commercial direct green laser released by mid-2012.
I felt compelled to address their issues and since they dubbed me a “soothsayer” I decided to wear the title proudly. In Microvision’s “Soothsayer(?)” for their “Number One Question” I responded to Microvision’s 8K and I went through and explained how they were being deceptive IMO. I then decided to write a series of “Soothsayer” articles detailing the many issues with Microvision’s laser beam scanning. Lance Evans, Microvision director of business development, admitted in the April 12, 2012 MIT Technology Review that the green laser alone was costing $200, far above the price required for a consumer product.
Bernard Kress had this to say –
What role does Microsoft play in this market?
Kress: HoloLens is part of the intelligent edge which gives enterprises access to our cloud services. That is our core business. Microsoft will not be a hardware company. But in the beginning of the market nobody else makes that level of quality in hardware. We are forced to heavily invest and make it on our own. We hope that contract manufacturers and other companies will get started and will produce excellent hardware themselves. I think it will be necessary for several big players to join forces—as they did in the CD/DVD market—to bring down the hardware-costs. In AR/VR and Mixed Reality most of the revenues will be accounted for by software and content. The AR/MR market is expected to grow over $ 100 Billion after 2020. Our goal is, that Microsoft´s Mixed Reality Cloud Services can be on every single device on the planet. Market analysis expect that the main part of the AR/MR market will happen in Asia.
Makes more sense to me that MSFT would want someone else to purchase MVIS and develop another LBS hardware path which could lead to more revenue for their core Cloud Services business rather than hoard the IP to themselves . MSFT already seems to have adequate licenses, patents, and engineers to kickstart things . Preventing hardware competition goes against their strategic goals and common sense .
Thanks for the article reference. I have read many articles with and by Bernard, but I don’t remember seeing that one.
First, Bernard Kress is a treasure to the AR/VR display industry. At the same time, he is more of a researcher/theoretician AND at the same time he still works for Microsoft. It is a delicate dance that not many people could pull off as well as he does.
All the above said you have to remember that Microsoft is trying to justify the money spent on Hololens. They have spent so far several billions of dollars (likely more than Magic Leap) so far and only shipped about 50K to 75K HL1 and maybe 10K to 20K HL2 with each until probably selling at a loss. It helps to justify Hololens in terms of helping a marketing tool for cloud sales.
Maybe I am misunderstanding your point, but your conclusion that IF Microsoft’s strategy is to develop the software and cloud services while leaving someone else to buy MVIS seems flawed. UNLESS Microsoft’s contract with Microvision, has the right to sublicense, then only Microsoft and one other company had the patent rights to build a Hololens like product? If Microsoft has the right to sublicense, there would be much less incentive for someone else to buy MVIS.
Interestingly your question “Do I think that Microvision’s Laser Beam Scanning is the AR Technology of the Future? ” Leads to a more fundamental question – is there really a future for AR? Or at least a future in the way that it has been hyped. In short, in your experience, is it more realistic to expect AR devices to be primarily for military (where they were first developed) and industrial applications? Also referring to Bernard Kress – in one of his Photonics West Talks, he questioned and doubted why consumers would want AR headsets. I’ve done my own mini study of people who are not in the tech industry for a more unbiased view, and whilst many think AR is a “cool” idea, few people would actually like to wear one everyday let alone ever replacing smartphones. It seems many in the AR industry are ignoring two highly popular uses of smartphones: taking selfies and writing messages / posts/tweets. There is no convenience or utility of doing these simple actions with a pair of glasses, no matter how small and “trendy” you can try and make them (and most young people hate wearing glasses, using contact lenses if needed!). Consumers still bought mobile phones when they first came out and were ropey with poor sound quality and short battery life, because they offered convenience. Given this, and the number of AR companies that have wound down over the past 2 years, I can’t help but wonder how or why there is still optimism that there will be a consumer market for AR.
AR is a MUCH harder problem than cell phones which could be viewed as a merger between the upward migration of smarter phones and the scaling down of computers. Before the iPhone, there were already semi-smart phones that supported web browsing. The iPhone did found a way to triple the screen size over a Blackberry by using a touch screen rather than a keyboard. The iPhone also bucked the smaller and cheaper trend. We even found that most people would prefer a bigger phone to have a larger display.
AR is much more up against physical limits than iPhone/smartphone/cellphone. When you cram things as small as you do in an AR headset, you run into a lot of physics problems. The pixels on display devices are on the order of 5 to 10 wavelengths of light which means that things like diffraction cause serious image quality issues. You are not facing just 1 or 2 problems but there are dozens or unsolve problems.
Just as you wrote about, even if engineers can solve all the technical challenges, there are massive human factors and social interaction problems. As I wrote in 2012 (yes over 8 years ago) out the parody videos for Google Glass:
Some of my writings on the subject:
Hi Karl. Nothing really surprising with both Microvision and ML.
LetinAR and Kura is what’s still a mystery to us and unknown if it is mostly hype or actually possible and with good visual performance. They claim quite a lot for a maxwellian display but both seem to do something new.
Hello Karl, I am a fan from China. I have followed you for a long time. Thank you for your selfless sharing. Regarding AR’s display engine, some MicroLED technology companies have recently made new progress. Do you think MicroLED is the ultimate solution for AR’s display engine? If not, then which technology can become the ultimate winner of display engine?
I would think the majority of knowledgable people in the AR industry expect MicroLEDs to become the leading technology and I am working on a series about the promise and issues with MicroLEDs. Most recently Facebook/Oculus has placed a second “bet” on MicroLEDs with their exclusive arrangement with Plessey, having purchased InifiLED in October 2016. Microsoft, by contrast, appears to only be licensing laser beam scanning technology from Microvision for the Hololens 2. There are still many challenges with getting full color. It is also not clear that they will work well with waveguides which have driven some companies, including Microsoft’s HL2, to look at laser scanning in spite of its many problems.
LCOS seems to be the leading technology for headsets expecting to go to market in the next few years with pupil-expanding waveguides. Waveoptics and Digilens appearing to move from DLP on their older designs to LCOS on their newer designs. The reason given for the move from DLP to LCOS is a combination of resolution, size, and cost.
Micro-OLEDs are also finding interest in companies like Nreal, Tooz, ActiveLook, and Everysight. MicroLEDs have to have much more efficient optics than pupil-expanding waveguides. This tends to limit them to smaller FOV applications (Tooz, ActiveLook, and Everysight) or have bulkier optics that are less transparent such was with birdbaths (Nreal).
[…] image projector technology solves this problem. As you can see below in these zoomed in images produced by kguttag.com Microvision’s technology is heavily integrated throughout this early version of the […]
[…] laser image projector technology solves this problem. As you can see in these magnified images produced by kguttag.com, Microvision’s tech is heavily integrated throughout early versions of the Hololens, […]
[…] is just a quick new flash as a follow up on last week’s blog entry, Microvision Worth More Dead Than Alive – Stock Jumps 147% in a Day. The stock continued going up that day before plunging back down again but still at more than 2X […]
[…] to Microsoft), and not long ago, Microvision laid off 60% of it workforce. As I pointed out in Microvision Worth More Dead Than Alive – Stock Jumps 147% in a Day, Microvision has pretty much put the company up for […]
Hi Karl, since you mentioned Bosch, did you see the smart glass live demo? Seems it’s like North Focals but with full RGB, wondering how they could guarantee the optical performance.
[…] May 5th, 2020, I wrote about what seemed like Microvision’s meteoric climb to $1.10/share in Microvision Worth More Dead Than Alive – Stock Jumps 147% in a Day, and it is now up over 10X from that […]