What happens if you are sued in Tokyo after Pantech v Google?

Takanori Abe of Abe & Partners walks through Japan’s SEP litigation and SEP Judicial Mediation guidelines in chronological order—from the petition to the judgment, and beyond.

It begins without your consent

What arrives from the Tokyo District Court may be a complaint. It may equally be a petition for SEP Judicial Mediation (SEPJM).

Until recently, the latter carried a condition. Judicial IP mediation at the Tokyo District Court required the other side’s agreement to jurisdiction; absent that agreement, referral could be refused. The amendment to the Civil Mediation Act in May 2026 removed the condition.

So long as Japanese courts have jurisdiction, an SEP holder may petition for SEPJM—one type of judicial IP mediation—at the Tokyo District Court without obtaining your consent.

For both procedures, the Intellectual Property Divisions of the Tokyo District Court published guidelines in January 2026: one set for SEP litigation, one for SEPJM.

Both are published as practice common to all four divisions that make up the Intellectual Property Divisions. Whichever division the case lands in, it proceeds the same way.

So let us take those two documents, and what actually happened in Pantech v Google, and follow them together in chronological order.

What must be decided before the Answer is submitted

At the complaint stage, the plaintiff proposes a global FRAND royalty, setting out the basis of calculation in specific terms. Because that is what the guidelines require.

What is required of the defendant in its Answer is not merely denials and defences. It must submit a counterproposal disclosing its own basis of calculation, supported by evidence of the sales volumes and sales amounts of its products.

The guidelines state this expressly: if the defendant does not voluntarily produce the evidence necessary to calculate the global FRAND royalty, the court may find that it has no intention to take a licence on FRAND terms.

Worldwide sales volumes and sales amounts—at this stage, to the court of this jurisdiction?

Produce them, and your hand is open in the first brief. Every negotiation that follows proceeds on those figures. Withhold them, and, exactly as the guidelines state, you carry the risk of being found unwilling. You face this choice before a single question of infringement or invalidity has been decided.

In SEPJM, what the Answer must contain is no different from litigation. What differs is what comes after. SEPJM is designed as a rule to be completed within three sessions. The same decision whether to produce or withhold must be taken, but with fewer opportunities than litigation allows.

Settlement is recommended at the first hearing

The guidelines provide that, in order to promote agreement on the global FRAND royalty as swiftly as possible, the court will as a rule recommend settlement at the first hearing.

In Pantech v Google, it went like this. On July 23, 2024, at the third preparatory hearing, the Tokyo District Court disclosed its view that infringement was established, recommended settlement, and asked both parties whether they intended to conduct settlement discussions before a Japanese court on the premise of the global SEP portfolio.

Pantech and Google both answered that they did. Google’s own settlement proposal covered Pantech’s worldwide patents as the licensed subject matter, and proposed a lump sum based on a per-smartphone royalty multiplied by the total projected units of PixelPhone (all models included) sold by Google, past and future.

Pantech v Google predates publication of the guidelines, and settlement was recommended at the third hearing. The guidelines have brought that forward to the first.

In the settlement room of a single Japanese court, the terms of a worldwide portfolio become the subject of discussion. That is not an exceptional course of events; it is the ordinary course the guidelines contemplate.

Produce no evidence, and you are found to have no intention to take a licence

September 30, 2024, the fourth preparatory hearing. The Tokyo District Court asked Google that, “if it has the intention to take a licence”, it submit a fresh settlement proposal, taking into account the calculation method set out in the IP High Court’s Grand Panel decision in Apple v Samsung—which takes the sales revenue of the end product as the starting point of calculation—and having regard also to the comparable approach.

Google, after internal deliberation, replied that submitting a further settlement proposal would be difficult. Two reasons. The products at issue included far more diverse models than in the Grand Panel case, so that applying the Grand Panel method as such would make the calculation excessively complex.

And its own proposal was equivalent to, or in fact more favourable to Pantech than, what the Grand Panel method would yield. It disclosed no sales amounts or sales volumes broken down by type of infringing product.

What Google used for that per-unit amount was a fixed amount common to all models, derived from the average price of smartphones generally rather than of its own products; what the court asked for was a method starting from the actual sales price and sales volume of each product. The two place the base differently. The former can be calculated without producing one’s own product figures; the latter cannot be calculated without them.

The judgment rejected Google’s answer. Google had not clarified the differences in model type or price, and had offered no concrete explanation of the degree to which the calculation would become complex.

On Pantech’s case, the smartphones at issue amounted to no more than some 28 products including derivative models; while it could not be denied that the calculation would take effort, it could not be said that the calculation itself was difficult, provided the sales price and sales volume of each product were confirmed.

The judgment then states: Google, despite having agreed to the Tokyo District Court’s settlement recommendation, refused to disclose the sales amounts and unit sales of the infringing products and, without presenting a settlement proposal under the Grand Panel method, itself eliminated the room for licence negotiations; therefore, on the literal terms of the court’s request, it cannot be found that Google has the intention to take a licence.

Google argued that its own proposal was itself reasonable. The judgment states that even if the proposal had a degree of reasonableness, Google can only be said to have itself eliminated the room for licence negotiations.

On June 23, 2025, for the first time in Japan, an injunction based on an SEP was granted.

What happened in that case was subsequently written into the guidelines: the passage providing that if the defendant does not voluntarily produce the evidence necessary to calculate the global FRAND royalty, it may be found to have no intention to take a licence. A determination in an individual case has been published as practice common to all four divisions.

Negotiating in good faith out of court is not enough

For completeness, the judgment also addressed how matters stood if only the out-of-court negotiating history between the parties were considered.

What is found there favours Google. That it proposed negotiating an NDA at the same time as expressing its intention to take a licence. That it continued studying claim charts while NDA negotiations were suspended. That it identified, with supporting reasons, why Pantech’s proposal could not be called reasonable, and presented a counterproposal. That those points were not themselves found unreasonable. That it continued negotiating after suit was filed and presented licence terms on several occasions.

The judgment concludes: Google did all it could in response to licence negotiations with Pantech. At least on the out-of-court negotiating history between the parties alone, it cannot be found that there are special circumstances showing that Google lacks the intention to take a licence on FRAND terms.

In the out-of-court negotiations, no fault was found in Google’s conduct. The injunction was granted nonetheless. The dividing point was not in the out-of-court negotiations.

Pantech v ASUS, handed down by the same presiding judge two months earlier, dismissed the injunction claim. There, the court found that the implementer executed an NDA immediately, asserted non-infringement and invalidity about two months after claim charts were provided, and presented a counterproposal about three months after the basis of calculation was provided—each finding accompanied by its period.

The number of days taken to respond is recorded in a published judgment. Whenever the parties are looked up in this jurisdiction thereafter, that record will not be missed.

In SEPJM, your response goes into the mediation record

If SEPJM is brought against you, both what is at issue and how the record is kept are different. The SEPJM guidelines are no different from litigation in requiring a counterproposal and evidence of sales volumes in the Answer.

But they contain no passage tying a failure to produce evidence to the assessment of the intention to take a licence. What goes on the record in SEPJM is whether a party has the intention to engage in the mediation.

Where the mediation fails because one party has no intention to engage in it, that fact is entered in the mediation record. Where the mediation panel has presented a mediation proposal, that proposal and the panel’s written opinion are attached to the mediation record.

And either party may submit that mediation record in subsequent litigation or preliminary injunction proceedings, on the issue of the abuse-of-rights defence—that is, on the intention to take a licence.

What carries over into the subsequent proceeding is not only the fact of failure. Who declined, and what terms the court’s side considered appropriate, carry over as an official record.

It does not end with the judgment

The injunction ordered in Pantech v Google was a domestic one based on a Japanese patent, and covered the Pixel 7, a discontinued model. On the judgment alone, the damage was limited. Google appealed.

While that appeal was pending before the IP High Court, in the month after the judgment Pantech petitioned the Tokyo District Court for a preliminary injunction against the Pixel 9, a current model. A proceeding separate from the appeal began in parallel, directed at a product still on sale.

The Tokyo District Court recommended an amount it found reasonable as a patent royalty on FRAND terms. On December 15, 2025, a settlement was concluded on the basis of that recommendation.

Google pays royalties on the 4G and 5G SEPs held by Pantech for the smartphones it sells worldwide. Fourteen related disputes in Japan, including the appeal pending before the IP High Court, and one in Germany, were all withdrawn.

A proceeding that began over a single Japanese patent determined worldwide licence terms. The German dispute ended there as well. It was the other side that chose Tokyo, not you.

Where was the dividing point?

The dividing point does not lie at the moment judgment is given. It lies at the Answer, and at the moment the court asks for something specific to be produced. It is not whether you accept or refuse the terms the court indicates, but whether you respond by not producing what was asked for, without concretely explaining why you cannot. That remains in the reasoning of the judgment in litigation, and in the mediation record in SEPJM.

The same structure, seen from the side that starts the proceeding, reads in reverse. A procedure that can be commenced without the other side’s agreement to jurisdiction, that calls for specific proposals and supporting materials from the first hearing, that records the response, and that in SEPJM allows that record to be carried into subsequent proceedings—such a procedure exists in Tokyo.

Six months have passed since the guidelines were published, two months since the amendment to the Civil Mediation Act. There is no appellate ruling. But what you must decide by the first deadline, once proceedings begin in Tokyo, is already written in the published guidelines.

The point at which to begin designing your defence in Tokyo is not when the complaint or the SEPJM petition arrives. It is before that.

Takanori Abe is a partner at Abe & Partners. He can be contacted at:  abe@abe-law.com