Wednesday, September 21, 2011

Without settlement, Google might pay more for Java license from Oracle than for Motorola Mobility

If Google ends up having to pay a per-unit royalty to Oracle for a Java license, it may within a few years pay even more -- possibly much more -- under such a settlement than the $12.5 billion price for its proposed acquisition of Motorola Mobility. In a worst-case scenario, which is however not too unrealistic, Google would have to pay billions of dollars to Oracle every year, which after several years could easily exceed the $12.5 billion offer Google made for Motorola.

Even a relatively high per-unit royalty such as the $15-$20 per-unit fee Oracle purportedly requests from Android device makers wouldn't be a mistake on Oracle's part. That amount wouldn't kill Android singlehandedly, and other patent holders will also ask for per-unit royalties, so if Oracle granted a license on any other terms, it might later find that other patent holders get a much bigger piece of the overall Android IP royalties cake. Oracle won't want to get paid a negligible amount for each copy of Android while others try to collect double-digit dollar amounts. It would be very frustrating for Oracle if in a couple of years the total amount of royalties collected per unit of an Android device exceeded $100 and Oracle, as the owner of the programming language in which almost all Android apps are written, received only a tiny piece of the action. Oracle won't let that happen. This is a once-in-a-lifetime opportunity for Oracle to secure itself a lucrative piece of the wireless platforms action.

"Little headway" in Monday's talks

After the Oracle-Google summit on Monday, led by both companies' CEOs, didn't result in a settlement, Magistrate Judge Paul S. Grewal ordered the parties to return to the San Jose Division of the United States District Court for the Northern District of California today. If there's still no settlement today, additional mediation sessions could take place between now and the end of the month.

Given how much is at stake, it's not surprising that there wasn't a first-day result. However, Bloomberg quotes an insider as saying that the parties made "little headway" in Monday's full-day talks.

Meanwhile, I see lots of off-base commentary from people who don't understand even the most basic parameters under which this settlement negotiation is taking place. It's remarkable how many people get to speak out on this high-profile case in the names of significant organizations even though they are completely clueless concerning how patent litigation and settlements work.

Contrary to what a misguided majority of commentators may say or imply, Oracle's original damages computation, which amounted in Google's representation to a maximum of $6.1 billion and in Oracle's own representation to $2.6 billion and has meanwhile had to be revised anyway, does not at all constitute a ceiling for how much Google will have to pay to get out of this. It constitutes neither a worst-case scenario for Google nor a best-case scenario for Oracle. It's detached from the real commercial issue, which is all about the terms on which Oracle would sell Google a license if the court orders an injunction of sufficient scope to be fatal to Android unless Google accedes to Oracle's demands.

Flowchart shows different decision-making steps and vastly divergent outcomes

I previously discussed possible settlement terms and I'd now like to present a couple of charts that make it very clear how this is going to play out. Let's start with this flowchart (click to enlarge):

We're now in a pre-trial situation. They are having those court-ordered talks, and those may fail simply for the lack of a firm trial date, which makes it all too tempting for Google to still hope that it can stall. Maybe the judge will set a definitive date before the end of mediation. Even if mediation fails, there's still a chance that they may work it out between mediation and trial, maybe just the night before the trial.

If they settle before a decision is handed, the parties will take into consideration the various possible outcomes and their probabilities. The basic outcomes are shown in that flowchart (the light-bluish rectangles with rounded corners). It wouldn't be a simple "weighted average" kind of calculation for the parties. A lot would depend on how the negotiations work out, and both parties -- especially Google -- would also have to consider the effect of any such settlement on other business interests they have. But at a high level it's correct to say that a pre-trial settlement would reflect how likely the various possible outcomes are to the parties.

If they don't settle, there'll have to be a trial (as early as Halloween, though not sure). In that case, the first major question for the court to decide will be whether Google infringes valid intellectual property rights held and asserted by Oracle. On Sunday I already explained that there isn't much left of the 20 affirmative defenses Google's lawyers presented last year. Still, a trial could end with a complete victory for Google. Maybe some of Oracle's intellectual property rights will be deemed invalid, and the remaining ones not to be infringed. This scenario is Oracle's worst-case and Google's best-case scenario. No money would change hands.

More realistically, there will be some infringement finding. Its scope is another question, but there will probably be something.

Possibility of injunction and workaround

In the event of an infringement finding, the court has to decide whether to order a permanent injunction. If the court believes that it's totally sufficient for Google to pay damages to Oracle, then Google can continue to make Android available anyway. That scenario would amount to a compulsory license: Oracle would have to tolerate Google's incessant infringement, but in exchange Oracle would receive whatever the court determines.

However, there are strong indications for willful infringement, which increases the likelihood of an injunction. Also, Google's non-compliant adoption of (parts of) Java results in what Oracle calls "fragmentation". The judge appears to share Oracle's concern that Android is, at best, a mixed blessing for Java, stating in a recent decision on a copyright-related Google motion that "[o]n the present record, a reasonable fact finder could disagree with Google's rosy depiction of Android's impact on the Java market". That would also be an argument in favor of an injunction.

Whether or not an injunction is ordered, Google will (have to) ask itself at this stage whether a workaround is viable. If the infringement finding is technically so narrow that Google can engineer around it, it will have the same effect for the outcome of the process as no injunction: Google would have to pay Oracle for past damages, but it would be fine going forward.

If Google can't work around the scope of the infringement finding, then we're talking about two fundamentally different scenarios depending on whether or not the court ordered an injunction. If there isn't an injunction, then Google can continue to infringe. The amount to be paid would depend on the damages ordered by the court. The court wouldn't just order a single damage amount. It would be more like a table that states an amount for each year during which the infringement continues, or the court might also order a per-unit damage amount. Anyway, the amount would depend on the court ruling, not on Oracle.

By contrast, if Oracle obtains an injunction that Google can't work around, then the two will again be negotiating a license agreement. But the situation will be largely detached from damages. Formally, Oracle would still be entitled to past damages. However, those past damages would be dwarfed by the running royalties Oracle could impose on Google in this situation. Oracle would hold an injunction in its hands with which it can shut Android down. It could make Google a take-it-or-leave-it offer, and in that scenario I can't imagine Google would rather give up Android than pay $15-$20 per unit. In that situation, there wouldn't be an argument over whether Oracle can also enforce its patent rights in the rest of the world, or over whether Google will be able to work around the injunction further down the road. Oracle would likely impose a per-unit royalty over a multi-year licensing term regardless of technical changes Google may make later.

At the current run rate, more than 200 million Android devices are activated per year. Android is experiencing phenomenal growth, and so is the overall market. In a few years Oracle could be receiving an annual payment from Google to the tune of many billions of dollars, based on a per-unit royalty that is multiplied with the applicable unit volume.

For the sake of accuracy, I'd like to point out that the viability-of-workaround question doesn't necessarily have a binary answer. There can be gray areas in which, for example, Google believes that Android would take a significant hit for a certain period of time, such as six months. In such scenarios, Google may have to decide between taking a temporary hit (in a rapidly-evolving market) and accepting Oracle's terms under what would likely be a multi-year license agreement. Also, even if the district court orders an injunction, Google might hope that the Federal Circuit could stay the injunction pending an appeal. There are some additional possibilities here, but the flowchart shows the basic permutations and outcomes.

Table of possible outcomes

For a summary and a point of reference, I've also created this table, which sums up the possible scenarios in the event that Google is found to infringe valid intellectual property rights owned by Google (click to enlarge):

In that table, the "workaround viable" column sums up all scenarios in which Oracle claims a victory so narrow in scope that Google can engineer around it. In that case, Oracle would get something for past damages (from the release of Android until the trial date) but Google would be fine going forward. Oracle's own damages estimate appears to amount to roughly a billion dollars for past infringement (the total amount also includes future infringement). I believe the court would most likely establish an amount in the hundreds of millions of dollars for past infringement. That amount could, however, be tripled in the event that Google is deemed to have infringed Oracle's rights intentionally. Post-tripling, that amount could therefore very well exceed $1 billion.

The table and the flowchart show the same basic results. The flowchart is more useful for seeing how this process would arrive at a particular outcome. The table makes it easier to compare the possible terms in a situation in which Google comes out on the losing end of a trial.

You can also download PDF versions of the flowchart and the outcome table from Scribd.

If you'd like to be updated on the smartphone patent disputes and other intellectual property matters I cover, please subscribe to my RSS feed (in the right-hand column) and/or follow me on Twitter @FOSSpatents and Google+.

Share with other professionals via LinkedIn: