Kioxia vs. Viasat Patent Litigation—A Patent Attorney Explains the 37 Billion Yen Jury Verdict | The Subject Patent (US8,615,700) and Practical IPR Strategies

Lead: $229 Million Jury Verdict Against Major NAND Flash Manufacturer—But Not a “Final Judgment”

On July 16, 2026 (U.S. time), a jury in the U.S. District Court for the Western District of Texas found that Kioxia Corporation (Kioxia America) had infringed U.S. Patent No. 8,615,700(hereinafter the "'700 patent") and awarded damages in the amount of $229,025,021 (approximately 37 billion yen).

In a statement issued on July 17, 2026, Kioxia Holdings stated that “Viasat’s claims and the jury’s decision are absolutely unacceptable” and indicated its intention to pursue “all available legal remedies,” including a motion for judgment as a matter of law (JMOL) and an appeal.The company also explained that there will be no impact on the provision of its products and services (these are all claims made by the parties involved). On the Tokyo Stock Exchange that same day, Kioxia Holdings shares closed at the daily limit-down price of 52,110 yen(down 10,000 yen, or 16.1%, from the previous day). However, this decline coincided with a broad sell-off in semiconductor stocks (the Nikkei Average fell 2,700 yen) triggered by a drop in U.S. AI stocks, so the decline cannot be attributed solely to the verdict.

This case is a cross-industry litigation involving a “satellite communication-derived error correction (ECC) technology patent that extended to NAND flash products,” and it offers valuable insights for the manufacturing and semiconductor industries when considering FTO (Freedom to Operate) investigations for patents in other sectors and litigation strategies following an IPR.Please note that, similar to our analysis of the Palworld patent litigation, this article explains a pending lawsuit based on publicly available information (court records, statements by the parties, and media reports) and does not prejudge the outcome of the infringement claim, the validity of the patent, or the final resolution of the case.

1. Case Overview and Final Timeline (Lawsuit Filed in 2021 → Verdict in 2026)

Plaintiff Viasat is a major U.S. satellite communications company that acquired Efficient Channel Coding, Inc.—founded in 1996—in 2005 for approximately $16.5 million plus earn-outs, thereby obtaining error correction technology for satellite applications.Defendant Kioxia is a leading NAND flash manufacturer whose predecessor was Toshiba Memory.The complaint states that the products in question are SSDs for enterprise, data center, and client use (including the FL6, CM6, CM5, PM6, CD, XD, XG, and BG series, among others) that incorporate Kioxia’s proprietary ECC technology, “QSBC” (based on the complaint).

The verdict amount is described as “running royalties covering only past infringement through March 30, 2026” and does not include a punitive surcharge. The treatment of infringement after March 31, 2026, as well as interest and future royalties, remains undetermined.

2. Details of the Subject Patent US8,615,700—The “Cross-Field” Scenario Where a Satellite Communications Company’s Patent Was Asserted Against NAND

First, a brief note on the technical background. ECC (Error Correction Code) is a technology that adds redundant information to data for verification purposes, enabling the original data to be restored even if bits are corrupted during recording or transmission.Whether in noisy satellite communications or in NAND flash—where errors increase due to miniaturization—the challenge of “detecting and correcting errors” is common, and this core technology circulates across industry boundaries.

'700 Patent (US 8,615,700 B2 “Forward error correction with parallel error detection for flash memories,” assignee Viasat, Inc.,filed August 18, 2010; registered December 24, 2013; expiration (after PTA adjustment) January 20, 2032; 25 claims in total) relates, as its title suggests, to an FEC (forward error correction) decoder for flash memory.I will quote the original text of Claim 1 as a representative example.

To explain this in Japanese: (1) A decoding module that receives and decodes encoded data from flash memory to generate multiple “partially decoded data streams”;(2) an error detection module equipped with multiple error detection submodules operating in parallel, which forwards only the portions containing errors to the correction side; and (3) an error correction module located “physically separate” from the error detection module.In practice, it is important to note that benefits such as low power consumption and improved reliability are derived from the specification rather than the claim language.

Note that Claim 1 has been invalidated and revoked in a separate IPR proceeding. The single claim presented to the jury was the one that survived the invalidity proceedings; according to expert commentary, this is Claim 16, which describes a configuration that adaptively changes the coding rate in response to the degradation state of the memory (direct verification in the original verdict document has not yet been completed).The above citation of Claim 1 is included as a representative example illustrating the technical framework of this patent.

Viasat’s complaint paints a narrative that “FEC technology developed for satellite communications was applied to flash memory,” but this is the plaintiff’s assertion and not a fact found by the court.On the other hand, the patent itself was written from the outset as a decoder for flash memory, so one cannot simply say, “Since it is a satellite company’s patent, it has no relevance to NAND”—this is a pitfall in FTO practice.

3. The Core of the Proceedings—The Failure to Invalidate the Patent in IPR Changed the Landscape of the Jury Trial

The procedural crux of this case lies in the fact that the claims were upheld in the IPR (Inter Partes Review) filed by Kioxia, and that decision was upheld by the CAFC (surviving claims 2, 10–14, 16, 23–25;The claims that had been invalidated were deemed moot, and certain arguments were deemed forfeited), the issues in the jury trial appear to have been effectively narrowed to “whether infringement occurred and the amount of damages” (the actual composition of the issues in the trial has not been confirmed in the public record).

The underlying principle is IPR estoppel (315(e)(2) of the U.S. Patent Act).A petitioner who has reached a final written decision in an IPR cannot raise in district court the grounds for invalidity that were “raised or could reasonably have been raised” in the IPR (there are no reports explicitly stating the application of this provision to this case; this is an explanation for legal clarification purposes).Once an IPR fails to invalidate the patent and a decision upholding the patent becomes final, the path to asserting invalidity in district court is virtually closed.

A specific example of a prior art reference that was not cited is the Lee patent (US 7,865,809, Super Talent), which was cited as prior art in the IPR.The PTAB found that “buffering alone does not constitute decoding,” and the CAFC upheld this ruling, stating that “buffering alone does not satisfy the ‘decoder’ limitation of Claim 15.” This is a real-world example where the interpretation of the claim terms “decode” and “decoder” determined the success or failure of the invalidity argument.Furthermore, the plaintiff, Viasat, presented only a single claim from among the upheld claims to the jury (as confirmed in Kioxia HD’s statement: “the single patent claim that Viasat presented to the jury”).While narrowing the issues to reduce the jury’s cognitive load is a common trial tactic, I will refrain from speculating on the specific objective in this case.

From the plaintiff’s perspective, the '700 patent family includes the '347 patent (US 8,966,347), which originated from a continuation application. A strategy of developing multiple sets of claims from the original application through continuation and divisional applications is standard practice for patent holders; for details, please refer to our article explaining divisional application strategies.

4. Comparison with Western Digital and SanDisk, Which Were Sued Over the Same Patent

On the same day (November 29, 2021), Viasat also filed a lawsuit against Western Digital. However, the patents at issue are not the same.

On July 28, 2022, WD/SanDisk filed a counterclaim against Viasat in the N.D. Cal. (4:22-cv-04376,US 9,424,400, 10,447,667, and 8,504,834). However, Claim 14 of the '834 patent was found to be ineligible under Section 101 (November 15, 2023), and regarding the '400 and'667 patent resulted in a summary judgment of non-infringement in Viasat’s favor in October 2025; the record shows that none of the three patents asserted in the counterclaim were upheld as infringement claims due to determinations of Section 101 ineligibility or summary judgments of non-infringement.The fact that, of the two companies facing the same '700 patent, only Kioxia proceeded to a jury verdict may be due to a combination of factors such as counterclaims, IPRs, and negotiations; however, this article does not assess the causal relationship or the merits of the decisions made by either company.Note that the case number of the “separate IPR” that invalidated Claim 1 and other claims of the ’700 patent could not be identified from publicly available information.

5. Practical Implications for Japanese Companies—FTO for Patents in Different Industries, Settlement Decisions, and Running Royalties

Implication 1: Designing FTO Searches That Include Patents from Different Industries

Generic element technologies—such as ECC, signal processing, and power management—may be held by applicants in “industries other than one’s own,” such as satellites, telecommunications, and broadcasting.Since FTO analyses focused solely on monitoring competitors are likely to miss these, it is necessary to design clearance searches based on functions and technologies (across CPC classifications). The release of proprietary technologies, such as QSBC, presents an ideal opportunity to conduct such searches.

Insight 2: The Decision-Making Process for Settlement Following an IPR Defeat

The point at which claims are upheld by the PTAB (November 2023 in this case) and the point at which the upholding is finalized by the CAFC (December 2025) are pivotal moments when the value of settlement options fluctuates significantly.Since negotiating leverage decreases once the possibility of invalidation is eliminated, companies should have an internal process to reassess ongoing costs and trial risks at each milestone (I will not comment on the merits of Kioxia’s decision in this case).

Implication 3: Structure of Running Royalty Damages

The $229 million figure covers only “past sales through March 30, 2026”; sales thereafter, interest, and future royalties may accrue separately.When a major product line with high sales volume is involved, even a small royalty rate can result in a massive absolute amount (please note that much of the detail regarding the royalty rate and calculation basis in this case is confidential).

Implications ④–⑥: Venue, Proactive IP Strategy, and Investor Relations Response

The Western District of Texas (Waco) is known as a jurisdiction where lawsuits filed by patent holders have been concentrated; as long as a Japanese company maintains a sales network through its U.S. subsidiary, it cannot avoid the risk of being sued.On the other hand, from Viasat’s perspective, the foundation of the verdict lies in its ability to patent technologies acquired through the acquisition in adjacent fields and maintain the patent family through continuous filings; its approach to claim drafting and patent family formation—which anticipates potential applications of its own technology—serves as a valuable reference for the IP departments of Japanese companies.Furthermore, as evidenced by the stock hitting its daily price limit down the day after the verdict (July 17, Japan time), patent verdicts are also capital market events; therefore, a perspective that incorporates timely disclosure and the preparation of anticipated Q&A into intellectual property risk management is required.

6. Summary and Future Watch Points

(1) An ECC patent for flash memory held by a satellite communications company was asserted against NAND flash SSDs, and the jury returned a verdict of $229 million. (2) After failing to invalidate the patent through IPR and CAFC proceedings, the issues in the jury trial appear to have been narrowed to focus primarily on infringement and damages.③ The verdict is not a final judgment; both the conclusion and the amount may change following post-verdict motions or an appeal—these three points form the backbone of this case.

This case is pending, and this article does not prejudge the outcome regarding infringement, patent validity, or the final resolution. For related articles, please also see our piece on the Palworld patent litigation, which explains how to analyze pending litigation, and our article on divisional application strategies, which covers the practical aspects of forming patent families.

[Disclaimer] This article is a general explanation based on publicly available information (court records, statements by the parties, and media reports) as of July 17, 2026, and does not constitute legal advice. This case is pending, and the content described herein—including any findings of infringement and the amount of damages—is subject to change depending on future proceedings.We plan to add updates as new information becomes available. (Supervised by a patent attorney)