Access Official Technical PDF

Share Summary:LinkedInEmail

Patent damages apportionment can be difficult to apply when an infringement claim concerns structures inside a semiconductor device. A 2019 petition by MLC Intellectual Property, LLC, in its dispute with Micron Technology, brought two related issues into focus: whether a damages expert could use negotiations behind comparable lump-sum licenses, and whether a royalty base tied to a single-component semiconductor die needed to be apportioned further. The petition asked the Federal Circuit for permission to review district court orders that excluded MLC’s damages evidence; it presented MLC’s arguments, not a final ruling on the merits.

Why damages evidence mattered in the Micron dispute

MLC alleged that Micron Flash Memory products infringed structural claims in a patent relating to multi-level memory devices. The patent had expired, and MLC sought a reasonable royalty rather than an injunction. In the district court’s view, the damages orders had left MLC without expert evidence to present on damages. The court certified the orders for interlocutory appeal, reasoning that early review could help avoid an inefficient sequence of separate proceedings.

The dispute matters to technical and commercial teams because a patent damages model connects the claimed technology to the products accused of infringement, the evidence used to value that technology, and the revenue selected as a royalty base. In this case, the proposed model relied on historical licenses and on the bare die as the smallest saleable patent practicing unit (SSPPU). Both choices were challenged.

Comparable licenses and the hypothetical negotiation

A reasonable royalty is commonly assessed through a hypothetical negotiation: an analysis of what a willing licensor and licensee might have agreed to before infringement began. The Georgia-Pacific framework provides factors for that analysis, including evidence of existing licenses and the parties’ bargaining positions.

MLC’s expert used licenses negotiated by BTG International with semiconductor manufacturers Hynix and Toshiba. According to the petition, both parties’ damages experts considered those licenses the most comparable to the hypothetical negotiation with Micron. The agreements were lump-sum licenses covering a portfolio that included the patent in suit, rather than documents stating a simple per-unit royalty rate for that patent alone.

To interpret how those lump sums might inform the hypothetical negotiation, MLC’s expert also considered evidence of the related negotiations. The petition describes, among other points, a most-favored-customer provision in the Hynix license and contemporaneous discussions associated with a 0.25% rate. MLC said this evidence helped explain how the earlier agreements were derived and how differences between those agreements and the hypothetical license should be considered.

The district court applied the parol evidence rule to exclude negotiation evidence outside the license documents. That rule generally concerns using outside evidence to modify or interpret an integrated contract. MLC argued that its expert was not using the evidence to change the agreements; rather, the expert was using it to estimate a hypothetical royalty. The petition therefore asked whether contract interpretation rules should bar relevant negotiation evidence in a Georgia-Pacific damages analysis.

For industrial companies, the underlying distinction is practical: a license document may record a final lump sum without showing the assumptions, bargaining positions or product context that produced it. Whether and how that background can be used in a damages model can affect the expert’s ability to compare agreements. The petition did not establish a general rule; it raised the question for appellate review.

Royalty-base apportionment and the semiconductor die

The second issue concerned the royalty base. Apportionment is intended to ensure that damages reflect the value associated with patented technology rather than unrelated features of a larger product. The SSPPU concept can help identify a product unit that practices the patent and is sold separately, but disputes can remain about whether that unit includes other components or functions.

MLC’s expert identified the bare die as the SSPPU, and Micron agreed that the bare die was the relevant unit, according to the petition. The expert treated the die as a single component containing the allegedly infringing structures. For other accused products that included additional items such as external leads and plastic packaging, the expert used the die’s average selling price as a reference point to reduce the revenue attributed to those products.

The district court nevertheless required further apportionment within the die to account for other functions, including error correction. MLC argued that once a single-component SSPPU had been identified, the royalty base did not need another reduction for functions that could not be separated conceptually or practically from that component. It contrasted this situation with a multi-component product, where apportionment may be needed to isolate a smaller component.

This distinction is relevant to semiconductor engineers and damages teams because a die can integrate multiple functions into one physical product. Determining whether the asserted technology can be separated from other functions may involve technical evidence as well as economic analysis. The petition argued that disputes over such assumptions could be tested through competing expert testimony and cross-examination, rather than resolved by excluding the damages opinion outright.

Practical considerations for technical decision-makers

The case illustrates several points for organizations managing patent risk or supporting litigation:

  • Preserve licensing context. Executed agreements may not explain how a lump-sum payment was negotiated. Relevant correspondence and negotiation records can provide context, subject to discovery and evidentiary rules.
  • Document comparability. Experts need to explain how differences in patent coverage, products, rights and commercial circumstances affect the use of a prior license.
  • Define the product unit precisely. Technical teams can help clarify what the accused product includes, how components relate to the claimed structure, and whether the proposed SSPPU is a single component or a larger assembly.
  • Separate technical and economic questions. Product architecture informs the analysis, but the royalty rate and base also depend on licensing evidence and the applicable legal framework.

Conclusion

MLC’s 2019 petition placed two recurring challenges in patent damages analysis before the Federal Circuit: the use of negotiation evidence to interpret comparable licenses for a hypothetical royalty, and the level of apportionment required when the SSPPU is asserted to be a single semiconductor component. The petition sought review; it should not be read as a final appellate determination of those questions. For manufacturers, integrators and technical leaders, the practical lesson is to make the product architecture, licensing record and assumptions behind any damages model clear and traceable. Those details can shape how a reasonable-royalty analysis is built and contested.

Access Official Technical PDF

About The Author