Lenovo’s major UK win over InterDigital: chart and further analysis

It’s not unheard of in complex business (and particularly patent) litigation that both sides will claim victory after a ruling. Spin doctors will always find something good among the bad in a multifaceted decision. But ip fray seeks to serve the licensing and litigation professions and can’t afford to fall for, much less to engage in, any kind of spin.

It’s not that everything in the decision is bad for InterDigital, but the bottom line is. The derivation that led to that bottom line has different aspects, and there’s also a secondary theater of war, which is the question of interest on past royalties. But the per-unit royalty was always going to be the grand prize, and Lenovo won that one.

The per-unit rate is what really matters

In the chart below, the three most important values are highlighted (reddish color):

Here’s an explanation of the numbers:

• The ECWA decision notes that InterDigital’S economic expert took the position that the appropriate per-unit royalty rate was 49.8 U.S. cents.
• That number is not inconsistent with the overall average of InterDigital’s license deals. InterDigital’s 2023 revenues were approximately $550 million. The global smartphone market has a unit volume of about 1.2 billion devices (May 30, 2024 IDC press release). If we assume that 85% of InterDigital’s license fee income comes from smartphone makers’ licenses to cellular SEPs, that means an overall average of 45.8 U.S. cents. That average is likely higher. For instance, OPPO has not been licensed yet as discussed further below, and OPPO’s volume is substantial.
• InterDigital made its official rates (the upper end of what they’d ever seek) public (InterDigital rate disclosure webpage). For 5G, that’s 0.6% of a device price that is always assumed (regardless of the actual price) to be at least $60, and capped at $200. That is a range from $0.36 (published minimum) to $1.20 (published maximum; not in the chart).
• The appeals court arrived at 22.5 U.S. cents, only 62.5% of InterDigital’s published minimum rate and 45.2% of InterDigital’s position in this litigation.
• That is only 29% more than the outcome in the lower court, and 41% more than Lenovo’s offer based on which this litigation started (it’s possible that Lenovo would have paid something more than that final official offer to obviate the need for litigation, as companies often offer one set of terms for the purposes of potential or actual FRAND litigation and another on a basis that can’t be used in litigation).

The lower court declared Lenovo the “commercial winner” because the outcome was so much closer to its proposed royalty rate than to InterDigital’s position. In principle, that is still the case. The appeals court raised the rate by 29% over the High Court’s decision, but it would have had to raise it by 185% to arrive at InterDigital’s proposed rate.

While not in the chart, it’s also worth noting that Apple agreed to pay InterDigital $134 million per year. Apple’s iPhone sales in 2023 were 231 million units. If the assumption is that 95% of what Apple pays InterDigital for relates to the use of cellular SEPs by its smartphones (as opposed to other types of devices and other technologies), that is an average per-unit royalty of 55 U.S. cents.

Ericsson comparison: $2.50 or $1-4 vs. $0.225

Prior to Friday’s decision, the last FRAND determination of this kind to have been resulted from two rounds of litigation was Ericsson v. HTC. The United States District Court for the Eastern District of Texas held a jury trial, and the outcome was that Ericsson had been consistent with its FRAND licensing obligation when seeking either a simple $2.50 per-unit royalty for its 4G patents or a corridor from $1 to $4 if HTC paid 1% of the device price, for which a minimum of $100 and a cap of $400 were defined.

If we take the $2.50 figure because it’s simple and probably a reasonable average for HTC’s 4G devices at the time, that is more than 11 times the per-unit royalty that the EWCA arrived at in the decision published on Friday.

Implications for ongoing and future negotiations

InterDigital also has litigation pending against OPPO. Having obtained a redacted InterDigital v. OPPO judgment from the Munich I Regional Court, ip fray knows that OPPO (like Lenovo) has never been licensed to InterDigital’s patents. The Munich ruling also notes that (in the late 2010s as ip fray learned from a different source) there was a period where OPPO did not respond for about a year and a half to InterDigital’s offer. Therefore, a UK ruling consistent with the outcome in the Lenovo case would result in substantial back-royalties (OPPO would not be able to argue that any of that is already time-barred) and interest on top. But the UK won’t decide that dispute. A Chinese court will. It doesn’t appear likely now that the Chinese court will arrive at a FRAND determination above the UK level. It will probably just do it in a more differentiated way, with a much lower rate for certain geographies.

In Germany, OPPO was enjoined despite pointing to the EWHC outcome. Now that there is even an appellate ruling that doesn’t arrive at a rate anywhere near what InterDigital demands, there is a possibility of the Munich Higher Regional Court ordering a stay of the enforcement of InterDigital’s German injunction against Lenovo and also of OPPO’s appeal (OPPO did not seek a stay as it had left the German market already due to some Nokia injunctions) succeeding.

As the UK ruling notes and was publicly known before, InterDigital entered into a license agreement with Xiaomi in 2021. It would be completely out of character for Xiaomi to have entered into a deal with a longer term than five years. Assuming that it was a five-year deal, it will expired in 2026 and renewal negotiations will start next year. It will be hard for InterDigital to get a lot more from Xiaomi, on a per-unit basis, than from Lenovo. And with Xiaomi already having taken a license, there isn’t even a threat of major back-royalties and interest on them.

Then there’s Samsung. Last year, InterDigital announced an arbitration agreement with Samsung (January 3, 2023 InterDigital press release). No further announcement concerning cellular SEPs (only one about other types of technology) was made. The EWCA decision may have come too late for Samsung to use in the arbitration, and this could also depend on the framework for the arbitration on which the parties agreed in 2023.

Will existing licensees seek adjustments?

Some of InterDigital’s licensees probably feel that they elected (or were coerced) to overpay. There is no precedent yet where an implementer of SEPs later achieved a downward adjustments, but Apple tried so against Qualcomm (the 2019 settlement was due to Apple not being able to rely on Intel for its 5G chips) and Deutsche Telekom is still making a long-shot attempt in Germany against IPCom (despite even having accepted a contract clause that precludes the “argument” Deutsche Telekom raises, and which failed in the Mannheim Regional Court as well as the Karlsruhe Higher Regional Court).

This does not mean to say that someone will try the same against InterDigital, but the risk is greater than zero, also in light of a more recent trend in the UK to accept free-standing FRAND claims.

Outcome by ground of appeal

What InterDigital focuses on when commenting on Friday’s outcome are two things: the much larger payment it will receive now than what Lenovo initially offered (we’ll get to that in a moment) and the claim that its appeal succeeded on all counts and Lenovo’s appeal failed all the way. Let’s fact-check that part:

Lenovo’s cross-appeal got nowhere. They wanted to shorten the period for back-royalties and to get rid of interest, or at least get a much lower interest rate set. But that was just because of a lot of money at stake. They at least had to try before forking over tens of millions of dollars without a fight. And by virtue of cross-appealing, they got more time in court.

InterDigital’s appeal was always going to be the key one:

InterDigital did not even attempt to persuade the appeals court that some other license than the LG license on which Mr Justice James Mellor based his decision should be used as the best comparable. In other words, they did not argue to the appeals court (as it would have gotten them nowhere) that a royalty rate they questionably extracted from a player like Fairphone (a small and medium-sized enterprise even by the narrow EU definition at the time, thus unable to defend itself in court). Fairphone was so small then and still is so small that InterDigital obviously had no intent other than shaking down a defenseless company (which understands the issues well, but litigation would have cost them a lot more than acceding to InterDigital’s demand) in hopes of creating a “comparable” for licensing negotiations and litigation with the likes of Lenovo. That strategy was a failure. The lower court in London rejected it, and InterDigital didn’t even dare to try that on appeal. They should do the only honorable thing now and repay Fairphone whatever it paid in excess of the UK FRAND determination.

They nevertheless tried to get (more than) 49 cents per device on appeal. The following bullet points quote the appeals court’s summary of each ground of appeal:

• “Ground A is that the judge derived the wrong dollar per unit rate from the LG 2017 licence because he failed to correct for non-FRAND effects which he found had affected past sales.” While the Court of Appeal agreed with InterDigital on that ground of appeal in principle, the impact is limited: “I would allow InterDigital’s appeal on ground A to the extent of substituting the figure of $0.30 for the figure of $0.24 per unit derived by the judge from LG 2017.” (Those per-unit numbers are then subject to an adjustment.)
• “Ground B, which is dependent on Ground A, is that, for the same reason, the judge did not correctly adjust the dollar per unit rate he derived from LG 2017 when determining the FRAND rate for Lenovo.” Here, too, InterDigital won, but the adjustment was very limited: the adjustment ratio was raised from 0.728 to 0.75. And 0.75 times 30 cents is 22.5 cents (the per-unit royalty at which the appeals court arrived). “Multiplying that figure by 792,571,429 units (the final figure used by the judge) gives a total of $178.3 million.”
• “Ground C is that the judge was wrong to reject the simpler of the two top-down crosschecks relied upon by InterDigital as being of value when determining the FRAND rate for Lenovo.” Here, the appeals court didn’t even agree with InterDigital that its top-down cross-check was good. There is no explicit holding that Ground C succeeded or failed. In a way, the appeals court treats it as a moot issue: “The judge’s principal reason for rejecting InterDigital’s top-down cross-check was that it was inconsistent with the result of the comparables analysis: […] I agree with the judge that the comparables analysis is a much more reliable basis for estimating FRAND than InterDigital’s top-down cross-check.”
• “Ground D is that the judge should have found, and declared, that InterDigital was a willing licensor.” On this one, the appeals court didn’t want to rule because it would serve no useful purpose.

Arguably, InterDigital’s Grounds C and D failed. At best, they were declared moot. Grounds A and B succeeded in a legalistic sense, but failed in economic terms. And they didn’t even dare to defend their original theory on appeal. All in all, that’s a disappointing outcome no matter what they say. It could have gone worse, and yes, as ip fray predicted, the appeals court wasn’t going to affirm the decision below and made an adjustment in InterDigital’s favor. It’s just that the adjustment doesn’t solve any problem for InterDigital.

Absolute amounts

InterDigital emphasizes that the $240 million Lenovo owes under the Friday decision is a lot more than Lenovo offered. For instance, one statement by InterDigital suggests it’s about three times what Lenovo said at the outset of that litigation.

If InterDigital had convinced the UK judiciary to award three times the per-unit rate that Lenovo described as FRAND, that would have been a major success. Lenovo said 16 cents was the right number, and 48 cents would have been almost the same as InterDigital’s demand.

But the difference in absolute numbers is mostly attributable to Lenovo’s position on back-royalties (they wanted to exclude many years from the royalty-bearing period) and interest, which Lenovo rejected.

There is no question that in purely mathematical terms, InterDigital will get some additional money that more than offsets the costs of the appeal. At the same time, Lenovo saved a lot more money by not acceding to InterDigital’s royalty demands (unless the 49 cents were just an official demand and they indicated to Lenovo that they’d settle for half as much, which is unlikely).

Both parties’ law firms did great work: Gowling WLG for InterDigital and Kirkland & Ellis for Lenovo.

But what Mr Justice Mellor said last year applies again: Lenovo is the commercial winner. InterDigital now has a problem on its hand with this UK decision, and by declaring victory it has made it even easier for third parties to argue that this royalty level is FRAND (whether those third parties negotiate new agreements or, as mentioned above as a hypothetical, try to get existing deals adjusted).

ip fray is 100% convinced that if InterDigital had known what the outcome after two rounds of litigation would be in the UK, they would not have bet on the UK but would have tried to get leverage elsewhere, such as in Germany.