Monday, June 12, 2017

Qualcomm's NXP deal raises chip- and patent-related concerns: in-depth review by European Commission

More than five years ago, Google's acquisition of Motorola Mobility was delayed significantly by merger reviews on both sides of the Atlantic and U.S. regulatory approval was subject to certain promises related to patent enforcement. At the time, Motorola Mobility (the acquisition target) was aggressively asserting FRAND-pledged standard-essential patents against Apple and Microsoft. Against that background of blatant FRAND abuse, competition enforcers weren't prepared to grant fast-track approval.

Qualcomm's planned $45 billion acquisition of NXP Semiconductors, a leader in NFC and secure elements (SE) chips, is now also undergoing an in-depth review by the European Commission and possibly also in other jurisdictions (though the deal surprisingly got fast-track clearance in the U.S.). I'm just in the process of trying to find out more about NXP's patent dealings. But it appears that, unlike in the Google-Motorola case, it's the acquirer's conduct that adds to concerns over what might happen post-transaction. That is even more problematic since the acquirer, not the acquisition target, will make the decisions post-transaction. The press release the Commission published late on Friday contains a few keywords that sound all too familiar in the ears of anyone following the current flurry of antitrust activity relating to Qualcomm:

  • "bundling": Presumably this is about chips that might be the equivalent of having a baseband processor and an NFC/SE chipset in a single product. Bundling is a sensitive issue in the EU and the Microsoft Media Player case created some case law.

  • "tying": This is also a key issue in the U.S. FRAND abuse case. The FTC would want Qualcomm to offer its chipsets without requiring a patent license on devices that use other companies' chipsets, and as far as Qualcomm's own chipsets are concerned, patent exhaustion, which is stronger in the U.S. than ever after a recent Supreme Court decision, should take care of the licensing question.

    Intel raised the issue of a dual, mutually-reinforcing monopoly in a recent amicus brief. I had written about that kind of dynamic before. The worst-case scenario in the NXP context is now that Qualcomm might expand on its monopoly. A highly simplified way to put it would be that Qualcomm will go from a dual to triple or quadruple monopoly, forcing customers to buy chipsets of one kind if they want access to others, and/or designing its patent license terms (including "rebates") in such a way that companies will end up sourcing various types of chips from Qualcomm.

  • "increased royalties for customers": Qualcomm is synonymous with out-of-this-world royalty rates. Last month I quoted an Apple letter (which Qualcomm had attached to a court filing) according to which "Qualcomm forces the contract manufacturers and Apple to pay many times more in royalty payments than all the other cellular patent licensors combined!"

    Why has Qualcomm been able to command such royalty levels? It's not just a matter of innovation. They do a lot of R&D, without a doubt, but without the two mutually reinforcing monopolies, even Qualcomm couldn't collect many times the amount of royalties of the rest of the industry combined. If Qualcomm now gets more leverage on both the chip side and the patent side because of the NXP deal, things will get even worse.

    Qualcomm already holds more NFC patents and applications than any other company such as Sony (#2) or Samsung (#3)--almost 1,000, or roughly 5.5% of the pool. After acquiring NXP, Qualcomm's position will go up to approximately 1,350 patents and applications, putting Qualcomm far ahead of the rest (almost twice as many as Samsung, for example). While consolidation of patent ownership positions is still preferable over Nokia- and Ericsson-style privateering, it does raise issues when a company is known to overcharge.

  • "exclusion of competitors": NXP's competitors, just like Qualcomm's, are chipset makers. The only remedy that could address this concern would be that the combined company would have to license other chipset makers on FRAND terms.

While the nature of the concerns is familiar, the NXP deal involves different technologies than the other Qualcomm cases--and it affects additional industries. Mobile device makers will be affected since an increasing number of smartphones come with NFC. But NXP also appears to be a key supplier to automotive companies; otherwise the Commission's press release wouldn't "particularly" mention that industry. Even independently of its contemplated acquisition of NXP, Qualcomm is trying to position itself as a technology licensor to automotive companies such as in connection with wireless electric vehicle charging. Qualcomm's inductive charging road is impressive.

In Europe, automotive companies have a lot of political clout. Maybe some of them have, directly or indirectly (through trade associations and national governments) made the EU Commission's Directorate-General for Competition (DG COMP) aware of their industry-specific concerns.

Reuters reported on Friday that Qualcomm is confident it can address the EU's concerns. I'm sure that the EU doesn't want to block the deal if it can be avoided, but any remedies would really have to have teeth. Negotiations are likely going to continue. The next key juncture is when the Commission will have to decide whether to issue a Statement of Objections (SO), which it will likely begin drafting soon in case it needs to take that step. It's been almost eight years since the SO against Oracle's acquisition of Sun Microsystems (the only SO against a merger during that entire year). At the time, I was a consultant to a complainant. Now I'm just an app developer and blogger, and I don't know how much time I'll find to dig into the details of this process, so if you can support my efforts with information, please make use of the contact form here. I protect my sources, of course.

Share with other professionals via LinkedIn: