Back in November, Public Knowledge and Open Markets Institute argued to the International Trade Commission that it would violate the public interest to grant Qualcomm’s request to ban iPhones that used Intel baseband technology from the U.S. market. We wrote then,
Qualcomm targeted iPhones that use Intel baseband technology with their lawsuit, conspicuously leaving out iPhones that use Qualcomm baseband technology, despite the fact that those phones also use technology that Qualcomm has claimed is infringing. By exercising its patent rights in a way tailored to harm a new competitor, Qualcomm seems more concerned with maintaining a monopoly position in the premium baseband market than in obtaining patent relief. This behavior appears targeted at pushing Qualcomm’s only competitor, Intel, to exit the market.
We therefore asked the ITC to uphold an administrative law judge’s opinion – which found, among other things, that if it granted Qualcomm’s request it was “nearly certain” that Intel would “exit the premium base band market” – that the public interest in competition meant that Apple should be permitted to continue selling iPhones with Intel, rather than Qualcomm chips.
Unfortunately that reality has come to pass, though through different means. As for the matter we filed in, the ITC, in reviewing the administrative law judge’s opinion, found that the one remaining patent in the case was, in fact, invalid. Thus it did not have to either agree or disagree with the judge’s finding (which was not at all popular among some segments of the patent bar) that yes there was infringement, but no there should still not be an import ban. This was a good outcome, though not one with the same legal effect as one might have wanted.
But, since then, Apple and Qualcomm have settled all of their litigation, and cut a deal where Apple will resume buying chips from Qualcomm and paying licensing fees. Like with a lot of settlements, it looks like the companies met in the middle, finding a pragmatic solution that satisfies their business and product roadmap needs.
I’m happy for them. But now that Apple is switching its business back to Qualcomm, Intel has exited the premium baseband market. Losing Apple’s business was indeed fatal to Intel’s baseband ambitions. I’m less happy about that.
To be clear this appears to have a lot to do with Intel’s inability to deliver the parts that Apple needs. The problem is less that one company is leaving the market, than that lots of companies have left the baseband market in the past decade: Freescale Semiconductor, EoNex Technologies, Texas Instruments, Renesas, Broadcom, ST-Ericsson, NVIDIA, and Marvell have all bowed out, ceding the market to Qualcomm. People see and interact with monopolies and companies with dangerous levels of market power every day, but monopolies are a problem behind the scenes as well, in the components of the products you buy and in the machines that make them, driving up cost and reducing choice in ways that are harder to see but are just as real.
This would be a problem even if Qualcomm achieved its market power fairly, through out-innovating and out-executing its rivals. But as competition authorities around the world have found, this is not the case. Qualcomm has abused standards-setting and patent licensing to obtain and lock in its dominance. Many patents are valuable simply because they are included in a standard, and this is why most companies are happy to agree to charge FRAND (fair, reasonable and non-discriminatory) rates in exchange for having their technologies included in a standard. This means that they charge a reasonable rate to anyone who is willing to pay – which can mean, of course, a lot more revenue than the zero they would receive if a rival’s patents were included in a standard instead. But Qualcomm has engaged in a number of tactics that have attracted legal censure around the world, leveraging patents that have become valuable due to the standard-setting process for its broader business ends. For example, it would refuse to license its FRAND-obligated patents to competitors – only to end users of the competitors’ parts. Consumer devices would therefore ultimately be licensed, but by demanding payment from downstream users of components that it has more leverage over (for instance, because it sells and licenses them other components and patents they need), it can demand higher “reasonable” fees than it otherwise would. (Fortunately, a federal judge has recently ruled that Qualcomm is required to license its FRAND patents to anyone who asks, including competitors.) Qualcomm would also demand a license that was a percentage of the final sale price of a device, meaning that it would get higher rates due to improvements to devices that its patents have nothing to do with. Also, Qualcomm’s broader strategy of bundling licensing in with the sale of chips, and charging for the same license over and over to multiple parties in the manufacturing chain, not only allow Qualcomm to extract higher payments, they muddy the waters for regulators and provide opportunities for complicated and anticompetitive “rebate” schemes designed to further shut out competitors. (It’s important to note here that under the doctrine of patent exhaustion, if a manufacturer sells a fully-licensed item, downstream users of that item don’t have to take out their own license, and if you buy a patented item from a seller that seller can’t turn around and sue you for patent infringement. A lot of Qualcomm’s actions can be seen as ways to contractually route around the patent exhaustion doctrine.)
That’s why, despite the Apple/Qualcomm settlement, we encourage the Federal Trade Commission to stay the course in its ongoing lawsuit against Qualcomm. Companies like Apple have to look out for their bottom line, and drawn-out litigation with a key business partner doesn’t usually make sense. Only public authorities like the FTC can stand up to companies like Qualcomm, and end their anticompetitive practices, hopefully lowering prices for consumers and making new market entry possible.