Google’s Acquisition Of On2 Not A Big Deal, Here’s Why

This morning, Google announced they would acquire On2 Technologies for $106.5M. While the deal is not expected to close until sometime in the fourth quarter, many writing about this are already declaring that Google will now disrupt the online video industry by speculating that Google will open source On2's VP8 technology. While Google could go the open source route, there are a lot of reasons why that would not disrupt the industry and I think way too many people now want to say Google is going to win the online video war, when all the did was acquire a codec. (also see part two of my post entitled "Debunking Some Myths Of The Google/On2 Deal, Questioning VP8's Quality")

While many want to speculate that Google bought On2 so that they could use On2's VP8 codec instead of H.264 for YouTube and not have to pay licensing fees to MPEG LA, the cost to license H.264 is $10,000 per website. So even if Google had 100 websites, it would only cost them $1M a year. That's not the reason to go out and spend 100x that to buy On2. So H.264 licensing is not the issue. Also, some folks have written that by buying On2, Google now doesn't have to pay Adobe a Flash license fee for YouTube streaming. As anyone in the industry knows, Google doesn't pay Adobe any fee for Flash as YouTube doesn't deliver any of their videos via the Flash streaming protocol, (RTMP) from a Flash server. All of YouTube's videos are progressive downloads which come off of Google servers, so no license fee is paid by Google to Adobe.

While Google could very well switch YouTube over to VP8 and then force Microsoft and Adobe to include support for it, that would go directly against YouTube's own support of H.264 for their HD videos and more importantly, would kill any four screen strategy for the company. Like it or not, many industries have already adopted the H.264 standard for everything from set-top-boxes to video conferencing applications. I don't see Google trying to build up VP8 to try and disrupt H.264 when it's a standard that has been well adopted. Look at all of the devices like the iPod and others using H.264. If Google truly wants to have a four screen strategy for delivering YouTube outside of the PC, VP8 is not the way to do it.

In addition, VP8 is a codec, not a platform. Today, content owners are working to solve the problems of the entire video ecosystem and need a platform and third party solutions that tie into that platform. The codec is just one small piece of the entire system that's needed. Many hardware devices today have built in support for H.264, like encoding boxes, to help content creators solve problems with encoding. None of those boxes have support for VP8 today and very few even have support for VP6.

The other big thing to keep in mind that is that today, no one has seen VP8 in action. Everyone is assuming the quality is better than H.264, but is it? On2 has yet to prove that in the market and even if it is a better technology, you wont overcome the market standards already in place for devices. Also, while YouTube has a lot of power in the market, keep in mind that consumers don't know and don't care what codecs are. While one could assume that Google will do something with VP8 and Chrome tying into the new HTML 5 standard, is that the reason to spend $100M to acquire a codec? I don't think so.

While many want to automatically assume Google will always be successful in whatever they do, simply because they are Google, they have never done anything well outside of search and advertising, as far as generating revenue goes. If they want to challenge H.264, they'll lose. If they simply acquired On2 for their own use in YouTube and Chrome, ok, could work. But it won't have any major impact on the industry.

  • Kyle Stonebrook

    Dan, since you think every reason for Google to buy On2 is NOT the reason or logical, I would like one question answered…
    Why then? If it is not worth it, why are they doing it? Why buy On2 if the VP codecs are not an issue? Why not buy some other firm developing H.264 for much less?
    I would appreciate you letting us all know since you seem to have a handle on this.

  • Dan, if they open-sourced VP8, wouldn’t that kill one of the major objections to HTML5 video (that Theora is not good enough?) in other words, the objective might not be ‘make a platform’ but ‘accelerate standards that weaken the Flash and Silverlight platforms.’ Google is very clearly in favor of HTML 5
    Otherwise, I can’t think of any rationale for the purchase. Just like they don’t need to buy ON2 to save on licensing fees, they don’t need to buy it to get back-end improvements for Youtube.

  • Codec Guy

    This is a bizarre move by Google that will probably cost them money. The risk with deploying codecs is an IP suit. This is a very, very high risk. Part of the “benefit” of deploying 264 or VC1 is that their patents are part of a patent pool and the risk is mitigated because infingement costs are a known.
    On2 doesn’t have a patent pool. The technology almost for sure infringes (codec patents are really messy and almost all codecs infrings on certain aspects of other codecs). This will limit Google’s deployments to internal uses i.e. they can’t license or distribute in a way that would require exposing the code.
    It is hard to see them buying the codec for internal purposes. Maybe this is just a big company that doesn’t know what to do with its money syndrome?

  • John Fitzpatrick

    I am not convinced by the argument above.
    First, the licensing argument: licensing rules are always subject to change. To wit, licensing a codec ‘per website’ is an outmoded metric. At some point, Google could have found itself in an unfair licensing situation. At bare minimum, owning VP8 will prevent that for years to come.
    Killing a ‘4 screen strategy’: I am confident Google has the weight to push their tech onto all four screens. And its my understanding that VP8 actually performs better than H.264 at low processing rates, making it a natural for portable devices.
    “VP8 is a codec, not a platform”: right. but Chrome is a platform. Android is a platform. Wave. etc. Google does platforms.
    “No one has seen VP8 in action”: are you presuming Google didn’t kick the tires here for a bit? I think they probably did.
    “can’t overcome standards” “Consumers don’t know/care”: Precisely why I think Google might win here. Google has privity with the consumer, more so than any competitor. Adobe? MSFT? Not the same level of engagement with the consumer.

  • Antonis, you raise a few interesting points that I’ve been mulling since I wrote the news article for early this morning. I didn’t do much analysis, due to the short timeframe to get the news up, but have spent the last few hours thinking through a few of the whys and wherefores, and Dan’s post brings a few of those to the forefront.
    I think Google’s got a nice three-base hit going on here, with the outside chance of a home run. As you mention, open sourcing VP8 could deal with the objections that have been raised more loudly since the Open Video Conference in June and the subsequent impact on the HTML 5 video tag discussions.
    As mentioned in the articles I did around both of those topics, Google’s made it plain, at least with Chrome, that it’s wiling to play both in the open-source and standards-based video worlds.
    The original open-sourcing of the VP codec that Theora is based on, VP3.2, has raised valid objections about the open-source community’s use of a “tired” code base to attempt an assault on H.264.
    The thing is, Google doesn’t have to open-source VP8 to be a part of the open-source movement toward a “free” codec; they can open-source VP7 instead, which provides a chance for the open-source community to mount a decent effort to catch up with H.264 (and for Wikimedia and Mozilla and Chrome to get what they wanted during the HTML5 debate)
    Based on the comparisons I did a few years ago (in Dec 2007, available at between VP6, VP7 and H.264, if they released VP7 as open-source it would bring open-source and standards-based codecs close to parity.
    The second-base play is that Google gets H.264 as part of its On2 purchase. That’s what the Hantro part of the On2 purchase gives Google a double benefit.
    The third-base play is that Google gets to make the open-source community happy, play in the standards-based world with H.264 and keep VP8 all to itself as a proprietary codec for live, low-latency, bi-directional streams (beyond Googletalk, think Skype, AIM Triton, the Chinese telecom that On2 has as a client for video IM, etc.)
    Will it be a home run? I think the economies of scale Google gets from buying out of an unknown licensing fee might look like a way to get that final base, but the real final sprint from third to home plate will be all about Google’s ability to execute on a vision like the one I’m conjecturing. Regardless, the gestalt of these moves means Google gets a chance to keep something for itself, while building a barrier to entry on the innovation side.

  • Roadman

    Allthough there is wide adoption of the H.264 standard, soon TV’s will come with (Chrome) Operating systems like the android Gphone Os.
    Set-top boxes cover ground in the video market flux but long term IPTV seems to fragment distribution.
    I here what you say about the ipod, but this device is loosing ground too, whats the point in having an ipod when my GPhone does the lot at the same price with more features?
    Google cant help but opensource the codec, without doing so would go against the whole brand recognition, they may also provide encoding services in the cloud.
    “While one could assume that Google will do something with VP8 and Chrome tying into the new HTML 5 standard, is that the reason to spend $100M to acquire a codec? I don’t think so.”
    Neither do I,
    They could have improved the Dirac codec which is already opensource or improved the Theora codec, very strange…

  • Codec Guy

    Open sourcing the codec would do nothing to remove the infringement risk. Further, when it comes to niche, very complex technologies such as codecs the benefits of open sourcing are diminished significantly. After all how many people around the world are there that have the math AND the video experience to really make improvements to a codec? That aren’t already employed by companies with a vested interest in compression? Not many.
    Don’t forget that while H264 isn’t open source, it is an open spec that anyone can contribute to. When developers contribute they have the opportunitiy to share in the profits from the licenses so there’s a good incentive. This drives a lot of optimization hence why mpeg2 got so much better over its lifetime.
    Now if Google got H264 as part of the deal then that makes sense. If they can distribute Android without paying further royalties then it was a bargin. That’s probably the most logical reason for them to buy On2.

  • Codec Guy, I should clarify what I said about H.264 / Hantro.
    One of the reasons, as least in my mind, that On2 bought Hantro was to hedge its bets and further embed itself into the embedded device space. What On2 got from Hantro, then, was a set of expertise that it could couple with a group of the original RPI researchers that form the core of On2’s R&D team.
    The latter group has been continually innovating for well over a decade, and while it’s not a surefire method for success, since so much money has to be rolled back into R&D, it certainly has allowed a small company like On2 to keep pace with the industry as a whole.
    Will this keep the IP lawsuits at bay? I suspect not, but since Google specifically calls out the R&D and innovation as a factor in the purchase, one wonders – if Google can keep them happy – what greenfield innovations the small team will be able to perform with somewhat limitless resources and a mobile / embedded platform that’s a potential alternative to RIM and Apple’s offerings.

  • Tom Streeter

    Hey, I have no idea what Google’s up to either, but I’m not sure anyone is quite being disruptive enough here.
    I don’t think the question necessarily *has* to be between open and closed-source. What if VP8 were released as a royalty-free open-standard? Don’t open source any On2 (now Google) optimized code, but allow the world to come in and take a shot. For free. Heck, publish a really unoptimized “demo implementation” to use as a jumping off point.
    Chrome (in all its manifestations) can be the showpieces for the Google version of the standard, but the W3C can — if it wants — put it into the HTML5 standard without blushing. Mozilla could implement it with an equal lack of embarrassment. Apple and Microsoft? It’s not 1997 anymore. If two mainstream browsers embrace standards and two don’t (in different ways, undoubtedly), who’s playing offense and who’s playing defense?
    Oh, and IP issues? Against the Googlebot Lawyer Squad? Really? Especially in an area as murky as codecs?
    Again, I don’t know if this is Google’s thinking or whether that rope above me is just getting increasingly wet. But it seems to me that there are scenarios where this is a big deal.

  • Tim, I agree with your comments. Well thought out. I will there was a way I could link to them directly.

  • If there is one thing in video you can count on, is that Rayburn’s first impression on anything is wrong. Netflix, the Flash video platform and YouTube going forward are a few easy examples. (There are dozens of others)
    Dan’s involuntary ignorance of On2’s products and business model are excusable only because even On2 the company doesn’t know where they are….leaderless for over a year and no marketing strategy has certainly left the company in question!
    ….but, your licensing reference is way off. Siglin’s ‘Labyrinth’ post misses the spec in many ways, the $5M annual cap isn’t even mentioned and it doesn’t even address the second term. Online video is about to see country if not time-zone taxation on top of a 10% increase across the board. Don’t even hint that licensing doesn’t play a role here! Look for example at a single company like DIVX, they paid $2.3M in tech licensing last Q and $2.2M the Q before that. Tell me again that licensing isn’t an issue for average companies looking to go 4 screens!! MPEGLA can’t be too excited about this development…lol
    On2 has a hundred customers paying recurring revenues, relationships with TI (DaVinci, OMAP3 & 4), Infineon, MediaTek, Freescale, Marvell and an engineering team that has readied VPx for the optimized embedded space. So, you think this was for a codec? This deal positions GOOG to kick some ass in the video space. Why ‘open source’ when they can bundle the ‘best video tech’ into their emerging Chrome and Android platforms, the ‘optimized embedded solutions’ can be purchased for a small fee for soon to be 1B devices a year. A browser and device OS for free that is bundled with the best video tech on earth and oh yea, I’ll give you the SDK’s and encoding tools for free to developers, ‘build my platforms’……, who wins in this search and ad market across 4 screens??
    Dan, who can really compete with that….just a simple question….can Flash and Silverlight counter??
    I enjoyed talking to you, Take care!!

  • “$5M annual cap isn’t even mentioned” …exactly what i was thinking, the idea that youtube will never have to pay more than $10k a year for h.264 licensing could be wishful thinking a few years down the road, and that’s just the tip of the iceberg.
    the purchase of vp8 and on2 represents leverage against onerous licensing costs, and it’s more than just the wired ‘net; if necessary, google can afford to buy it’s own wireless spectrum.

  • The bottom line is that even IF Google was spending $5M a year on licensing H.264, and MPEG LA raised their rates by double next year, since Google spent over $100M for On2 it would take Google TEN years to save money on licensing costs. Lets say MPEG LA raised their rates 5x next year, it would still take Google FIVE years to recoup their money. I know Google looks to the future, but they are not spending $100M today, so save money five or ten years from now. The simple math proves this point.
    Maybe reducing their licensing costs is one of the added benefits of acquiring On2, but it’s not the reason for the acquisition.

  • as mark cuban and techcrunch have pointed out, though, it was a stock-for-stock transaction, no cash involved, so what was the actual out-of-pocket cost for google?

  • mseenu

    With the browser emerging as the unifying platform across a fractured OS space (Windows, Android, Moblin, …) and video emerging as the killer app in the browser, Google would like to control video in the browser. With the video-in-the-browser wars heating up, having a codec that’s better than Theora (and H.264) and less expensive than H.264 helps Google create a dominant position for Android, Chrome (the browser and the OS) and YouTube. I’d be surprised if they open-sourced the codecs …

  • Anon

    This certainly seems a strange move. I wonder how many patents On2 holds – perhaps this is Google buying up patents in case of a video codec patent attack.
    I can’t see that Google gains much from opensourcing the latest On2 stuff (they didn’t open source Google Earth for example) nor do I see them rolling the new tech into something like GTalk… I think it is far more likely to be useful in some sort of set top box / small device venture.
    One thought that occurs to me is perhaps they want to use On2 engineers to improve ffmpeg?

  • Xiou

    This acquisition is all about China! On2 has strategic deals with Netease and other key Chinese players(, and its CODECS are the De-Facto standard for Chinese video delivery across ALL platforms. Besides China On2’s codecs/encoding/transcoding solutions are widely used throughout Asia and Google will now have much greater influence across the entire region. Asia has BILLIONS of potential users and On2 is THE WEB VIDEO KINGMAKER FOR ASIA. It would be foolish if the other major US players let Google steal this company, their cutting-edge tech and future explosive royalty-based revenue for a pittance. Barring other offers form the US, expect a savvy company in India or Asia to act!

  • Codec Guy

    Xiou – China has a long, long history of not paying for codecs and in fact the Chinese govt has stated publicly for years that they are developing their own codec. It is highly unlikely that they are going to now pay Google.

  • Moissinac

    Google seems to give a good support to SVG. A problem with the open approach of SVG is the lack of commonly used open video codec. Have a video codec could be a weapon to fight the hegemony of Flash and the emergence of Silverlight in multimedia on the web by pushing SVG. It’s not a way to get a lot of money; it’s a way to weaken Adobe and Microsoft, two major concurrents of Google. Perhaps…

  • Comment by NilmaBostonRio:
    About On2 VP8 encoder Merger Update and Google, I know now through this article.
    But I already knew the VP6 format that even you can try a free demo of Flix Pro encoder has On2 VP6 – Two pass encoding On2 VP6-S for High Definition output – the company’s website:
    But I still prefer the H.264 QuickTime Pro to upload some of my web videos on YouTube in HD High Definition with compact file sizes.

  • jj

    They bought this so they could control, enable and have rule and word on the licensing for firmware encode/decoding of video in micropowered SOC deployments. Alos possibly displace MPEGLA in 2014 when the pool rules on it change. IMHO very very forward thinking, the benefit will reveal itself years from now. It will be why you will see conversation — your phone, google tablet mobile whatever iPatioBrick become a video phone.

  • mac

    …Open letter to Google: free VP8, and use it on YouTube….
    great post thanks!

  • “While one could assume that Google will do something with VP8 and Chrome tying into the new HTML 5 standard, is that the reason to spend $100M to acquire a codec? I don’t think so.”
    Lol, funny to read this again now after Google released VP8 as WebM under the BSD license (the do-whatever-the-heck-you-want license). By the way, it was announced during Google IO, and the only thing they talked about was HTML5. A day later, in fact, they changed the Google logo to an HTML5 PacMan game. I think they’ve expressed a pretty big support for HTML5, and I highly doubt that their purchase and open-sourcing of VP8 wasn’t almost entirely due to HTML5, and having an open standard codec.

  • The HTML5 endorsed by Google is simply the next generation and I am sure Google will put their energy and focus behind it, driving this market segment. And the video industry? Who know what the next step is Google will take…