ATTORNEY WORK PRODUCT — PRIVILEGED AND CONFIDENTIAL — PREPARED IN ANTICIPATION OF LITIGATION
CALDERWOOD & HARKNESS LLP 350 South Grand Avenue, 48th Floor Los Angeles, California 90071
INTERNAL MEMORANDUM
ATTORNEY WORK PRODUCT — PRIVILEGED AND CONFIDENTIAL This memorandum is protected by the attorney work-product doctrine and the attorney-client privilege. It was prepared in anticipation of litigation and reflects the mental impressions, conclusions, opinions, and legal theories of counsel. Do not distribute outside the litigation team without prior authorization from the Responsible Partner.
TO: Alan Ngo, Responsible Partner, Calderwood & Harkness LLP Marcus Delgado, Billing Partner, Calderwood & Harkness LLP
FROM: Wendy Schultz, Associate, Calderwood & Harkness LLP
CC: Omar Siddiqui, Associate, Calderwood & Harkness LLP
DATE: May 20, 2024
RE: Lost Profits Analysis Under Panduit — Optiwave Semiconductor Corp. v. Lumicore Technologies, Inc., Case No. 2:24-cv-00319-JRG (E.D. Tex.)
MATTER NO.: 1014-00004-OWS
TABLE OF CONTENTS
Right-click to update Table of Contents
I. EXECUTIVE SUMMARY
This memorandum analyzes Optiwave Semiconductor Corporation's ("OWS") entitlement to lost-profits damages under the four-factor framework established by the United States Court of Appeals for the Sixth Circuit in Panduit Corp. v. Stahlin Bros. Fibre Works, Inc., 575 F.2d 1152 (6th Cir. 1978), as subsequently adopted and refined by the Federal Circuit in a line of decisions spanning more than four decades. The purpose of this memorandum is to provide the analytical foundation for our retained damages expert, Dr. Lawrence Keating of Meridian Economic Consulting Group, LLC, whose expert report is due on June 15, 2025, and to frame the damages component of the complaint demand in this matter.
Conclusion. Based on our review of the record developed to date — including OWS's internal financial records, audited financial statements prepared by Gresham & Whitfield LLP, discovery produced by Lumicore Technologies, Inc. ("Lumicore"), deposition testimony, and Dr. Keating's preliminary analysis — we conclude that OWS can satisfy all four Panduit factors for a substantial share of the accused-product sales. Specifically, OWS can demonstrate: (1) strong and growing demand for products embodying the asserted patents; (2) the absence of acceptable non-infringing substitutes during the infringement period; (3) sufficient manufacturing and marketing capability to have exploited the diverted demand; and (4) quantifiable lost profits attributable to the infringement.
The analysis covers the infringement period from Q3 2022 — the quarter in which Lumicore launched its LumiLink 400G-ZR+ module, the first accused product — through the projected end of Q4 2025, which represents the anticipated trial horizon based on the current scheduling order setting a trial date of February 2, 2026, before the Honorable Judge J. Rodney Gilstrap.
Dr. Keating's preliminary analysis projects total accused-product revenue over this period at approximately $714.6 million. This figure encompasses revenue from all three accused product lines: the LumiLink 400G-ZR+, the LumiLink 800G-DR8, and the LumiEdge 1600G-LR.
Our lost-profits theory applies to the 38.2% market-share displacement fraction — that is, the share of Lumicore's accused-product sales that OWS would have captured "but for" the infringement. The remaining 61.8% of accused-product sales will be addressed under a reasonable-royalty theory, which is analyzed separately in the Comparable License Analysis memorandum prepared in July 2024 and Dr. Keating's forthcoming Georgia-Pacific analysis.
The resulting lost-profits estimate of approximately $128.9 million derives from the following computation:
$714.6M (projected accused-product revenue) × 38.2% (displacement fraction) × 47.3% (OWS gross margin) ≈ $128.9M
This memorandum is designed to guide Dr. Keating's expert report and to frame the damages component of the complaint demand of approximately $210.0 million. That aggregate demand figure reflects a blended view of lost profits ($128.9M), reasonable royalty at an 8.5% rate applied to the remaining accused-product sales ($37.5M), and a selected willful-infringement enhancement component ($43.6M). The demand figure is a rounded, conservative number — not a simple arithmetic sum of the individual theory maxima — and is intended to present a credible and well-supported damages position.
II. LEGAL FRAMEWORK
A. The *Panduit* Four-Factor Test
Under the Patent Act, 35 U.S.C. § 284, a prevailing patentee is entitled to "damages adequate to compensate for the infringement, but in no event less than a reasonable royalty for the use made of the invention by the infringer." The Federal Circuit has long recognized that lost profits, where proven, are the preferred measure of compensatory damages because they most accurately reflect the harm suffered by the patentee. See Rite-Hite Corp. v. Kelley, Inc., 56 F.3d 1538, 1545 (Fed. Cir. 1995) (en banc).
The standard framework for establishing entitlement to lost-profits damages in patent cases is the four-factor test articulated in Panduit Corp. v. Stahlin Bros. Fibre Works, Inc., 575 F.2d 1152, 1156 (6th Cir. 1978). Under Panduit, a patentee seeking lost-profits damages must establish, by a preponderance of the evidence, the following four factors:
1. Demand for the patented product;
2. Absence of acceptable non-infringing substitutes;
3. The patentee's manufacturing and marketing capability to exploit the demand; and
4. The amount of profit the patentee would have made.
The Federal Circuit has consistently applied this framework in the decades since Panduit was decided. See, e.g., Rite-Hite, 56 F.3d at 1545; Crystal Semiconductor Corp. v. TriTech Microelectronics Int'l, Inc., 246 F.3d 1336, 1353 (Fed. Cir. 2001); BIC Leisure Prods., Inc. v. Windsurfing Int'l, Inc., 1 F.3d 1214, 1218 (Fed. Cir. 1993). The burden of proof rests squarely on the patentee to establish each of the four factors. Crystal Semiconductor, 246 F.3d at 1353. Failure to establish any one factor precludes an award of lost profits, though the patentee remains entitled to at least a reasonable royalty as a damages floor. Panduit, 575 F.2d at 1156.
Importantly, the Panduit test is not the only permissible method for proving lost profits; it is one accepted analytical framework that the Federal Circuit has endorsed. See Rite-Hite, 56 F.3d at 1545 ("Panduit is not the sine qua non for proving 'but for' causation."). Nevertheless, it remains the most widely used and judicially accepted framework, and it provides a clear, structured approach that is well-suited to the facts of this case.
Where a patentee cannot demonstrate entitlement to lost profits on all infringing sales — for example, because it lacks a competitive position at certain customers or because some portion of the infringing sales would have gone to non-infringing competitors rather than to the patentee — a reasonable royalty under 35 U.S.C. § 284 serves as a damages floor for the remaining sales. See State Indus., Inc. v. Mor-Flo Indus., Inc., 883 F.2d 1573, 1577 (Fed. Cir. 1989). A patentee may thus recover a combination of lost profits on some infringing sales and a reasonable royalty on others. Id. This is precisely the approach we propose in this case: lost profits on the 38.2% displacement share, and a reasonable royalty on the remaining 61.8%. The reasonable-royalty component is analyzed in the separate Comparable License Analysis memorandum and will be presented in Dr. Keating's forthcoming Georgia-Pacific analysis.
B. Two-Supplier Market vs. Multi-Supplier Market
The structure of the relevant market has significant implications for the Panduit analysis, particularly under Factor Two (absence of acceptable non-infringing substitutes). Where the patentee and the infringer are the only two meaningful competitors in the relevant market — a so-called "two-supplier market" — a rebuttable presumption arises that the patentee would have made the infringer's sales absent the infringement. See Micro Chemical, Inc. v. Lextron, Inc., 317 F.3d 1387, 1392 (Fed. Cir. 2003) ("Where the patent owner and the infringer are the only two suppliers of a product, the patent owner is entitled to a strong presumption that it would have captured the infringer's sales.").
The silicon-photonics transceiver market for hyperscale data centers, however, does not present a pure two-supplier market. Beyond OWS and Lumicore, additional competitors participate in the market with varying degrees of market share and product overlap. As discussed in detail below under Factor Two, at least two other silicon-photonics-based transceiver suppliers hold meaningful market share in the 400G-ZR and 800G-DR8 segments. Their presence forecloses a simple "but for" presumption that OWS would have captured 100% of Lumicore's accused-product sales.
Nevertheless, the market dynamics strongly favor OWS's lost-profits claim. For the specific product categories at issue — 400G-ZR coherent transceivers, 800G-DR8 modules, and 1600G-LR modules — OWS and Lumicore compete directly and head-to-head at twelve of the top twenty hyperscale data-center operators. In these competitive situations, OWS is the principal alternative to Lumicore, and the evidence shows that customer purchasing decisions in these competitive situations were driven by the patented features that distinguish silicon-photonics-based solutions from alternative technologies. The 38.2% displacement fraction reflects a rigorous, customer-by-customer analysis of competitive dynamics that accounts for the presence of other competitors while recognizing OWS's strong competitive position at the specific accounts where overlap exists.
C. The Incremental Profits Standard
Lost-profits damages are measured by the incremental profit that the patentee would have earned "but for" the infringement. This means the relevant profit figure is the patentee's revenue minus its variable or incremental costs — not its fully burdened or absorbed cost structure. See Grain Processing Corp. v. American Maize-Products Co., 185 F.3d 1341, 1348 (Fed. Cir. 1999) ("Lost profits must be determined by reference to the patentee's incremental profit margin on the lost sales."); Minco, Inc. v. Combustion Eng'g, Inc., 95 F.3d 1109, 1119 (Fed. Cir. 1996).
The incremental-profit standard recognizes that the patentee's fixed costs (e.g., rent, executive salaries, depreciation on existing facilities) would have been incurred regardless of whether the additional sales were made. Accordingly, only the variable costs attributable to producing and selling the incremental units are subtracted from revenue to determine lost profits. See Micro Chemical, 317 F.3d at 1393 ("A patentee may not recover for overhead costs that would have been incurred even without the lost sales.").
OWS's gross margin on its competing products — the OWS-400ZR and OWS-800XR transceiver modules — is 47.3%, as documented in OWS's internal financial records produced in discovery and verified by OWS's independent auditor, Gresham & Whitfield LLP. This gross margin reflects revenue minus cost of goods sold ("COGS"), which includes wafer fabrication costs paid to foundry partners, packaging and module assembly costs paid to OSAT providers, testing costs, and direct materials. Because OWS is a fabless semiconductor company, its COGS consists predominantly of variable costs that scale with production volume, making gross margin a reasonable and conservative proxy for incremental margin. Dr. Keating may refine this figure in his expert report to reflect a true incremental margin — which could be higher than 47.3% — by excluding any fixed-cost components that may be embedded in COGS. For purposes of this memorandum, we use the conservative 47.3% figure.
III. PANDUIT FACTOR ONE: DEMAND FOR THE PATENTED PRODUCT
A. Overview of Market Demand
The first Panduit factor requires OWS to demonstrate demand for the patented product. Panduit, 575 F.2d at 1156. This factor asks whether there is a market for the patented invention and whether the patentee actually sold products embodying the patented technology. The demand inquiry also encompasses whether the patented features — as opposed to unpatented aspects of the product — drove customer purchasing decisions. See Rite-Hite, 56 F.3d at 1548-49.
OWS is the third-largest U.S. supplier of silicon-photonics transceiver modules for hyperscale data-center applications. According to its Gresham & Whitfield LLP audited financial statements for fiscal year 2023, OWS recorded total revenue of $487.3 million and net income of $61.4 million. These figures reflect a well-established market position in a rapidly growing sector of the optical communications industry.
OWS's technological position in the silicon-photonics space is further evidenced by its substantial intellectual property portfolio. As of the date of this memorandum, OWS holds 142 issued U.S. patents and 58 pending U.S. patent applications directed to various aspects of silicon-photonics design, fabrication, integration, and packaging. This portfolio reflects sustained and significant investment in research and development over more than a decade, and it underscores OWS's position as a technology leader — not merely a market participant — in the silicon-photonics transceiver space.
The broader market context strongly supports a finding of robust demand. The hyperscale data-center market is experiencing extraordinary growth in demand for high-bandwidth, low-power optical interconnects. This growth is driven by several converging forces: the rapid expansion of artificial intelligence and machine learning workloads, which require massive inter-server bandwidth; the continued expansion of cloud computing infrastructure by hyperscale operators such as major cloud service providers; and the buildout of 5G network infrastructure, which requires high-capacity backhaul and fronthaul optical links. Industry analysts project the silicon-photonics transceiver market for data-center applications to grow at a compound annual growth rate ("CAGR") of approximately 25-30% over the next several years, making it one of the fastest-growing segments of the broader semiconductor market.
B. Demand for Products Practicing the Asserted Patents
OWS's key product families — the OWS-400ZR, OWS-800XR, and OWS-1600DR — directly embody the innovations claimed in the four asserted patents in this case. The relationship between OWS's products and the asserted patents is as follows:
• OWS-400ZR module: This product practices the claims of U.S. Patent No. 10,234,847 (the "'847 Patent," directed to waveguide multiplexer architecture with low insertion loss), U.S. Patent No. 10,891,193 (the "'193 Patent," directed to coherent optical transceiver integration on silicon), and U.S. Patent No. 9,678,562 (the "'562 Patent," directed to micro-channel thermal management for silicon-photonics transceivers). The OWS-400ZR was launched prior to Lumicore's entry into the market and has been widely adopted by Tier 1 hyperscale data-center operators as a critical component of their optical interconnect infrastructure.
• OWS-800XR module: This product practices all four asserted patents — the '847, '193, '562, and U.S. Patent No. 11,402,015 (the "'015 Patent," directed to a polarization-diverse grating coupler for silicon-photonics transceivers). The OWS-800XR is a key growth product for OWS, representing the company's next-generation platform for 800-gigabit-per-second optical interconnects.
• OWS-1600DR module: Currently in ramp-up phase, this product practices the '193 and '015 Patents. The OWS-1600DR represents OWS's entry into the emerging 1.6-terabit-per-second transceiver market and is designed to meet the bandwidth demands of next-generation AI/ML training clusters.
Approximately $310 million of OWS's $487.3 million in FY 2023 revenue is attributable to products that practice one or more of the asserted patents. This figure — which accounts for approximately 63.6% of OWS's total revenue — will be confirmed and refined by Dr. Keating in his expert report based on detailed product-level financial data. The relationship between OWS's commercial products and the patented technology is also relevant to the analysis of secondary considerations of non-obviousness, which is addressed in a separate memorandum.
Critically, the patented features are not incidental design choices; they are the core performance differentiators that drive customer purchasing decisions in the hyperscale data-center market. Specifically:
• The waveguide multiplexer architecture claimed in the '847 Patent enables low insertion loss, which is critical for achieving the reach and link-budget performance required by hyperscale operators deploying 400G-ZR coherent transceivers over distances of up to 120 kilometers.
• The coherent optical transceiver integration on silicon claimed in the '193 Patent enables the monolithic integration of transmitter and receiver functions on a single silicon-photonics die, reducing cost, size, and power consumption — all critical metrics for hyperscale deployments where per-port economics are paramount.
• The micro-channel thermal management claimed in the '562 Patent addresses one of the key challenges in high-density silicon-photonics transceivers: managing the heat generated by high-speed electronic and photonic components in compact form factors. This technology enables reliable operation at the elevated temperatures and power densities encountered in hyperscale data centers.
• The polarization-diverse grating coupler claimed in the '015 Patent enables efficient coupling of light between the silicon-photonics chip and external optical fibers, regardless of the polarization state of the incoming light. This is essential for achieving the performance and reliability specifications required by hyperscale operators.
These features are not merely desirable enhancements; they are prerequisites for qualification and deployment in the demanding hyperscale data-center environment. Customer procurement specifications — copies of which were produced in discovery — confirm that the patented features correspond to critical performance parameters that must be met for a transceiver to be considered for deployment.
C. Demand for the Accused Lumicore Products
The strength of market demand for products embodying the patented technology is independently confirmed by the commercial success of Lumicore's accused product lines. In fiscal year 2023 alone, Lumicore's accused products generated aggregate revenue of $264.9 million, broken down as follows:
• LumiLink 400G-ZR+ (launched Q3 2022; accused of infringing the '847, '193, and '562 Patents): $94.8 million in FY 2023 revenue.
• LumiLink 800G-DR8 (launched Q1 2023; accused of infringing the '847, '193, '562, and '015 Patents): $128.4 million in FY 2023 revenue.
• LumiEdge 1600G-LR (launched Q4 2023; accused of infringing the '193 and '015 Patents): $41.7 million in FY 2023 revenue.
The total accused-product revenue from Q3 2022 through Q1 2024 — representing the infringement period through the filing of the complaint — is approximately $378.2 million. Projected accused-product revenue through Q4 2025, the projected trial horizon, is approximately $714.6 million, based on Dr. Keating's preliminary analysis of Lumicore's revenue trajectory and market growth assumptions. The rapid uptake and scale of Lumicore's accused products — which grew from zero to over $264 million in annual revenue in approximately 18 months — powerfully confirms the existence of robust and growing market demand for the patented technology.
It is also noteworthy that the accused products constitute the overwhelming majority of Lumicore's business. Lumicore's FY 2023 total revenue was approximately $312.6 million; accused-product revenue of $264.9 million therefore represents approximately 84.7% of Lumicore's total revenue. The centrality of the infringing products to Lumicore's business further underscores the commercial value of the patented innovations and the magnitude of the harm inflicted on OWS.
D. Conclusion on Factor One
Factor One is strongly satisfied. Demand for products embodying the patented inventions is established by both OWS's substantial revenue history and Lumicore's rapid and significant accumulation of accused-product sales. The hyperscale data-center market is experiencing sustained growth in demand for the specific type of high-bandwidth, low-power silicon-photonics transceivers that OWS's patents enable, and the patented features are critical performance differentiators that drive customer purchasing decisions.
In his expert report, Dr. Keating should detail market growth projections, quantify the causal nexus between the patented features and customer demand, and present data from customer procurement specifications and purchasing decision documents. Gregory Falk, our retained industry expert — an independent consultant who formerly served as Vice President of Engineering at Solara Optics, Inc. — can provide testimony at trial on the broader market dynamics and the importance of the patented innovations to the data-center interconnect market. Mr. Falk's testimony will be particularly valuable in connecting the technical features of the asserted patents to the commercial realities of the hyperscale procurement process.
IV. PANDUIT FACTOR TWO: ABSENCE OF ACCEPTABLE NON-INFRINGING SUBSTITUTES
A. Legal Standard
The second Panduit factor requires the patentee to demonstrate the absence of acceptable non-infringing substitutes. Panduit, 575 F.2d at 1156. The purpose of this inquiry is to establish "but for" causation: if acceptable non-infringing alternatives were available, customers who purchased the infringing product might have purchased one of those alternatives rather than the patentee's product, breaking the causal link between the infringement and the patentee's lost profits.
The question is whether, "but for" the infringement, customers who purchased the accused products would have purchased the patentee's products instead, or whether they would have switched to a non-infringing substitute. The patentee need not show that its product was the only option in the market; rather, it must show that there were no acceptable non-infringing alternatives that would have absorbed the demand that the accused product served. See Rite-Hite, 56 F.3d at 1548.
The Federal Circuit has emphasized that non-infringing substitutes must be "available" — that is, actually on the market or capable of being quickly brought to market during the relevant period — not merely theoretically possible. See Grain Processing Corp. v. American Maize-Products Co., 185 F.3d 1341, 1353-54 (Fed. Cir. 1999). A hypothetical product that could have been designed and manufactured at some future date does not constitute an "available" non-infringing substitute. Id. at 1354. The focus of the inquiry is on the infringement period (here, Q3 2022 onward), not on future design-arounds or products that could have been developed in response to the litigation.
Moreover, a purported substitute must be "acceptable" to the relevant customer base. A product that is technically non-infringing but that does not meet the performance, cost, form-factor, or reliability requirements of the customers who purchased the accused product is not an "acceptable" substitute. See Standard Havens Prods., Inc. v. Gencor Indus., Inc., 953 F.2d 1360, 1373 (Fed. Cir. 1991).
B. Analysis of Potential Non-Infringing Alternatives
Calderwood & Harkness has identified four potential alternative products or technologies that Lumicore or its customers might point to as non-infringing substitutes for the accused products. Each is analyzed below. Our conclusion is that none of the identified alternatives — individually or collectively — constitutes an acceptable non-infringing substitute sufficient to defeat OWS's lost-profits claim. However, the partial availability of other silicon-photonics competitors (Alternative 3 below) is appropriately accounted for through the 38.2% market-share displacement methodology.
1. Alternative 1: Indium Phosphide (InP)-Based Transceivers
Several suppliers in the optical communications industry offer coherent transceivers based on indium phosphide ("InP") semiconductor technology as alternatives to silicon-photonics-based modules. InP technology has a long history in optical communications and remains in use for certain applications. Some market participants might suggest that InP-based transceivers could have served as substitutes for Lumicore's silicon-photonics-based accused products.
Analysis. InP-based transceivers do not practice the asserted patents, which are specific to silicon-photonics platforms. However, this does not resolve the inquiry; the question is whether InP transceivers are functionally equivalent substitutes that would be acceptable to the relevant customer base.
The evidence overwhelmingly shows that they are not. InP-based transceivers lack the integration density, power efficiency, and cost scalability of silicon-photonics platforms. Silicon photonics leverages the existing CMOS fabrication infrastructure — the same high-volume manufacturing processes used to produce billions of silicon chips each year — to achieve per-unit costs that InP-based solutions cannot match at scale. This cost advantage is particularly significant in the hyperscale data-center market, where operators deploy hundreds of thousands of transceiver modules and per-port economics are a critical purchasing criterion.
Beyond cost, InP-based transceivers suffer from several technical disadvantages relative to silicon-photonics solutions in the hyperscale context. InP die are smaller and more fragile, limiting integration density. InP fabrication processes are less mature and lower-yielding than silicon CMOS processes. And InP-based modules typically consume more power per unit of bandwidth, a critical disadvantage in data centers where power and cooling budgets are tightly constrained.
The documentary evidence confirms these commercial realities. Customer qualification documents produced in discovery show that at least eight of the twelve overlapping hyperscale customers — i.e., customers at which both OWS and Lumicore compete — specify silicon-photonics-based transceivers in their procurement requirements. These specifications effectively exclude InP-based alternatives from consideration for the deployments at issue.
Conclusion. InP-based transceivers are not an acceptable non-infringing substitute for the accused products. They do not meet the performance, cost, and integration requirements of the hyperscale customer base.
2. Alternative 2: Externally Modulated Laser (EML)-Based Transceivers
Externally modulated laser ("EML")-based transceivers represent an older-generation technology that has been used in optical communications for decades, primarily in shorter-reach, lower-bandwidth applications such as metro and access networks.
Analysis. EML-based modules typically do not achieve the per-lane bandwidth (400 gigabits per second and above), reach, or power efficiency required for the data-center interconnect applications targeted by both OWS's and Lumicore's products. EML technology is fundamentally designed for different applications — it uses discrete optical components rather than the monolithically integrated silicon-photonics approach claimed in the asserted patents.
EML-based transceivers are not interchangeable with 400G-ZR, 800G-DR8, or 1600G-LR form factors. They do not comply with the Multi-Source Agreement ("MSA") specifications that govern the mechanical, electrical, and optical interfaces used by hyperscale data-center operators. A hyperscale operator cannot simply "swap in" an EML-based module in place of a silicon-photonics-based 400G-ZR module; the two technologies are designed for different speed classes, different reach requirements, and different deployment architectures.
Conclusion. EML-based transceivers are not an acceptable non-infringing substitute. They represent a fundamentally different technology class that does not serve the same market or meet the same customer requirements as the accused products.
3. Alternative 3: Other Silicon-Photonics-Based Transceivers (From Competitors Other Than OWS and Lumicore)
In addition to OWS and Lumicore, at least two other companies — identified generically here pending completion of Dr. Keating's detailed market analysis — offer silicon-photonics-based transceivers in the 400G and 800G product categories. These competitors hold a combined market share of approximately 22% in the 400G-ZR segment and approximately 15% in the 800G-DR8 segment.
Analysis. These competitors' products are the most plausible alternative to both OWS's and Lumicore's offerings, as they are based on silicon-photonics technology and target the same hyperscale data-center applications. They thus present the most significant challenge to OWS's Panduit Factor Two showing.
However, several important qualifications apply:
First, these competitors' products may themselves practice one or more of the asserted patents. Our technical team is currently investigating whether these products incorporate the claimed waveguide multiplexer architecture, coherent transceiver integration, micro-channel thermal management, or polarization-diverse grating coupler technologies. If so, they are not "non-infringing" alternatives, and their sales represent additional instances of infringement that may be addressed through FairPoint Licensing Group's broader licensing campaign. This investigation is ongoing and should be completed before Dr. Keating finalizes his expert report.
Second, even assuming arguendo that these competitors' products do not infringe the asserted patents, their limited production capacity and customer qualification status during the infringement period (Q3 2022 through the present) meant they could not have absorbed the full demand served by Lumicore's accused products. The evidence shows that these competitors had constrained foundry capacity during 2022-2023, were not qualified at several of the twelve overlapping hyperscale accounts, and experienced supply-chain disruptions that limited their ability to fulfill orders during the critical early phase of the infringement period.
Third, and most importantly, Dr. Keating's market-share displacement analysis expressly accounts for the presence of these competitors. The 38.2% displacement percentage reflects the portion of Lumicore's accused-product sales that would have been captured by OWS "but for" the infringement — not the entirety of Lumicore's sales. The remaining 61.8% of Lumicore's sales are allocated to these other competitors (and to customers where OWS did not have a competitive position), and those sales are addressed under the reasonable-royalty theory rather than the lost-profits theory. This approach is conservative, methodologically sound, and consistent with Federal Circuit precedent permitting a "split" damages analysis. See State Indus., 883 F.2d at 1577.
Conclusion. Other silicon-photonics competitors are partially a substitute for some customers, but their presence is fully accounted for in the 38.2% displacement figure. Their existence does not defeat OWS's lost-profits claim; rather, it reduces the share of Lumicore's sales to which lost profits apply.
4. Alternative 4: Lumicore's Own "Non-Infringing" Design-Around
Lumicore may argue at trial that it could have designed around the asserted patents and still competed in the market, and that therefore OWS would not have captured the accused-product sales even in the absence of infringement. Under this theory, Lumicore would contend that a hypothetical non-infringing version of its products constitutes an "acceptable non-infringing substitute."
Analysis. This argument fails on both the facts and the law.
On the facts, the evidence shows that Lumicore has no non-infringing design-around — and never had one. Dr. Sandra Michaud, Lumicore's Chief Technology Officer, testified at her deposition on November 19, 2024, that she could not identify any specific non-infringing design alternatives for the accused products. When asked directly whether Lumicore had any design-around options that would avoid the asserted patents while maintaining comparable performance, Dr. Michaud testified that she was "not aware of any such alternatives at this time." Raymond Bautista, Lumicore's General Counsel, admitted at his deposition on December 3, 2024, that no formal design-around analysis was commissioned before the launch of the LumiEdge 1600G-LR in Q4 2023 — the last of the three accused products to enter the market. Mr. Bautista acknowledged that the decision to launch the LumiEdge 1600G-LR was made without any analysis of whether the product might infringe OWS's patents, despite the fact that Lumicore had received two notice letters from FairPoint Licensing Group (in January 2023 and May 2023) identifying the specific patents at issue.
Lumicore's own internal estimates, produced in discovery, indicate that a redesign to avoid the asserted patents would require 18 to 24 months of development time and would result in measurable performance degradation, including higher insertion loss (due to the need to avoid the claimed waveguide multiplexer architecture), reduced thermal management efficiency (due to the need to avoid the claimed micro-channel thermal management technology), and loss of polarization diversity in the grating coupler (due to the need to avoid the claimed polarization-diverse grating coupler design). These performance trade-offs would likely render the redesigned product unacceptable to the hyperscale customer base.
On the law, the Federal Circuit has squarely held that a hypothetical future design-around is not an "available" non-infringing substitute. Grain Processing, 185 F.3d at 1353-54. The Grain Processing court emphasized that the availability inquiry focuses on what was actually available in the market during the infringement period, not on what could have been developed in response to litigation. A design-around that would take 18-24 months to develop and that did not exist during the infringement period does not qualify.
Conclusion. Lumicore's hypothetical design-around is not an available non-infringing substitute during the relevant period. Both the factual record (deposition testimony and internal documents) and controlling precedent (Grain Processing) foreclose this argument.
C. Overall Conclusion on Factor Two
Factor Two is satisfied. No single non-infringing substitute was available during the infringement period that could have absorbed the full demand served by Lumicore's accused products. InP-based and EML-based transceivers are not acceptable substitutes because they do not meet the performance and cost requirements of the hyperscale customer base. Other silicon-photonics competitors existed but could not have absorbed all of the diverted demand; their presence is already accounted for in the 38.2% displacement methodology. And Lumicore has no available non-infringing design-around.
The 38.2% market-share displacement methodology — detailed in the separate Market Share Displacement Analysis memorandum (Appendix A) — appropriately accounts for the partial availability of other silicon-photonics competitors. This approach ensures that OWS claims lost profits only on the sales it would have actually captured "but for" the infringement, while the remaining sales are addressed under the reasonable-royalty theory. Dr. Keating should address each of the four alternatives analyzed above in his expert report and explain why the 38.2% displacement figure is conservative and well-supported by the customer-overlap data and competitive dynamics evidence.
V. PANDUIT FACTOR THREE: MANUFACTURING AND MARKETING CAPABILITY
A. Legal Standard
The third Panduit factor requires the patentee to demonstrate that it had the manufacturing capacity and marketing infrastructure to have made the additional sales that were diverted to the infringer by the infringement. Panduit, 575 F.2d at 1156. This factor serves a causal function: if the patentee could not have produced or sold the additional units, then the infringement did not cause the loss, and the patentee is not entitled to lost profits on those sales. See Wechsler v. Macke Int'l Trade, Inc., 486 F.3d 1286, 1293 (Fed. Cir. 2007).
The showing required is not that the patentee had idle capacity sitting on the shelf; rather, it is that the patentee had the capability — including the financial resources, supply-chain relationships, and operational infrastructure — to ramp production to meet the incremental demand. See Gyromat Corp. v. Champion Spark Plug Co., 735 F.2d 549, 554 (Fed. Cir. 1984) (patentee need not show "excess capacity" but rather the "ability to have produced the additional volume").
B. OWS's Manufacturing Capability
OWS operates as a fabless semiconductor company, a business model common among leading semiconductor firms. Under this model, OWS designs its silicon-photonics circuits and transceiver modules in-house and contracts with third-party foundry partners for wafer fabrication and with outsourced semiconductor assembly and test ("OSAT") providers for packaging, module assembly, and testing. This fabless model provides OWS with significant flexibility to scale production volumes up or down in response to demand, because it can draw on the capacity of large foundry and OSAT partners rather than being constrained by its own fixed manufacturing assets.
OWS's foundry contracts — copies of which were produced in discovery — include capacity reservation and ramp-up provisions that are specifically designed to accommodate increased production volumes on reasonable notice. These provisions, which are standard in the semiconductor industry, guarantee OWS access to a certain level of foundry capacity per quarter, with the ability to increase orders by a specified percentage (typically 20-30%) with advance notice periods ranging from 30 to 90 days. Based on our review of these contracts, the available capacity reservations would have permitted additional production volumes corresponding to the displaced Lumicore sales — i.e., the incremental units OWS would have produced to serve the 38.2% share of Lumicore's customers that would have turned to OWS "but for" the infringement.
OWS also maintained substantial financial resources that could have funded the working-capital increases necessary to ramp production. As of the end of FY 2023, OWS had approximately $125 million in available credit under its revolving credit facility with Weston Fidelity Bank, N.A. This credit facility could have been drawn upon to fund the purchase of additional wafer starts, packaging materials, and other inputs needed to increase production volumes. The availability of this credit facility demonstrates that OWS's manufacturing capability was not constrained by financial limitations.
Dr. Nikolai Petrov, OWS's Vice President of Engineering for the Photonics Division, is prepared to testify at trial regarding OWS's ability to increase production of the OWS-400ZR and OWS-800XR modules within existing foundry capacity and contractual commitments. Dr. Petrov has direct responsibility for managing OWS's relationships with its foundry and OSAT partners and has detailed knowledge of the production capacity available to OWS during the infringement period. His testimony will establish that OWS could have ramped production to fulfill the incremental demand without any significant delays, capital expenditures, or disruptions to its existing product supply.
OWS has approximately 1,250 employees, including a dedicated operations and supply-chain management team with extensive experience in managing high-volume semiconductor production across multiple foundry and OSAT partners. This team has successfully managed significant production ramps in the past — for example, during the initial launch of the OWS-400ZR in 2021, OWS ramped production from initial qualification volumes to full-scale production within six months — and possesses the organizational capability to manage the incremental volumes at issue in this case.
C. OWS's Marketing Capability
OWS is the third-largest U.S. supplier in its market segment, with well-established relationships with the major hyperscale data-center operators that constitute the target customer base for both OWS's and Lumicore's products. OWS's marketing capability is not hypothetical or aspirational; it is demonstrated by OWS's existing market position and customer relationships.
OWS's sales and marketing team, led by Vice President of Sales Karen Whitfield, maintains active customer relationships at twelve of the top twenty hyperscale data-center operators — the same twelve customers at which OWS and Lumicore compete head-to-head. At each of these twelve accounts, OWS's products have been qualified and approved for deployment, meaning that the products have undergone the rigorous evaluation and testing process that hyperscale operators require before permitting a component to be used in their production networks. No additional qualification would have been needed for OWS to serve incremental demand at these accounts; OWS could have simply fulfilled additional purchase orders for products already approved and in use.
This is a critical point. In the hyperscale data-center market, customer qualification is a significant barrier to entry. The qualification process can take six to twelve months and involves extensive testing, performance validation, and reliability assessment. Because OWS's products were already qualified at the twelve overlapping accounts, OWS was in a position to capture additional demand immediately — without any delay for qualification, testing, or deployment planning.
OWS's investment in its sales and marketing infrastructure is substantial. In FY 2023, OWS spent approximately $42 million on sales and marketing, per its Gresham & Whitfield audited financial statements. This expenditure funded a global sales organization with dedicated account teams for each of the major hyperscale operators, a marketing team responsible for product positioning and competitive intelligence, and a field applications engineering team that provides technical support to customers during the evaluation, qualification, and deployment phases.
D. Conclusion on Factor Three
Factor Three is satisfied. OWS had both the manufacturing capacity — through its foundry and OSAT partners, supported by contractual capacity reservations and the $125 million Weston Fidelity revolving credit facility — and the marketing infrastructure to capture the additional sales diverted to Lumicore. OWS's established customer relationships, existing product qualifications, and proven supply-chain management capabilities confirm that it could have served the incremental demand without significant delays or additional investment.
Dr. Petrov and Karen Whitfield are anticipated trial witnesses who can provide firsthand factual testimony on OWS's manufacturing and marketing capabilities, respectively. David Leitner, OWS's Chief Financial Officer, can additionally testify to the financial capacity available to support a production ramp, including the terms and availability of the Weston Fidelity credit facility.
Dr. Keating should address Factor Three in his expert report with specific reference to OWS's production capacity data, foundry contract provisions, customer qualification records, and the credit facility documentation. This factor is unlikely to be seriously contested by Lumicore, but a thorough presentation will foreclose any opportunity for Lumicore to create doubt on this point.
VI. PANDUIT FACTOR FOUR: AMOUNT OF PROFIT
A. Legal Standard
The fourth and final Panduit factor requires the patentee to quantify the profit it would have made on the lost sales. Panduit, 575 F.2d at 1156. As discussed above in Section II.C, the relevant measure is incremental profit — the revenue the patentee would have earned minus the incremental or variable costs it would have incurred to produce and sell the additional units. See Grain Processing, 185 F.3d at 1348; Minco, Inc. v. Combustion Eng'g, Inc., 95 F.3d 1109, 1119 (Fed. Cir. 1996). This incremental-profit approach excludes fixed costs that would have been incurred regardless of whether the additional sales were made, and it captures the true economic harm suffered by the patentee.
The patentee bears the burden of establishing the amount of lost profits with reasonable certainty. While mathematical precision is not required, the patentee must provide a non-speculative, well-supported damages estimate grounded in reliable evidence. See Lam, Inc. v. Johns-Manville Corp., 718 F.2d 1056, 1065 (Fed. Cir. 1983).
B. OWS's Profit Margin on Competing Products
The key input for the profit calculation is OWS's profit margin on its competing products — the OWS-400ZR and OWS-800XR transceiver modules, which are the products OWS would have sold to the customers who instead purchased Lumicore's accused products.
OWS's gross margin on its competing products is 47.3%, as established by OWS's internal financial records produced in discovery and consistent with Gresham & Whitfield LLP's audited financial statements for FY 2023. The gross margin reflects revenue minus cost of goods sold, which for OWS includes the following cost components:
• Wafer fabrication costs: Payments to foundry partners for the fabrication of silicon-photonics wafers containing OWS's circuit designs.
• Packaging and assembly costs: Payments to OSAT providers for the assembly of silicon-photonics die into transceiver modules, including the attachment of optical components, fiber coupling, and module enclosure.
• Testing costs: Payments for wafer-level testing, module-level testing, and final quality assurance testing prior to shipment.
• Direct materials: Costs of optical fibers, lenses, connectors, printed circuit boards, and other materials incorporated into the finished transceiver module.
The 47.3% gross margin does not include allocated overhead costs such as research and development expenses, selling, general and administrative ("SG&A") expenses, or depreciation on corporate assets. Because these costs are largely fixed and would have been incurred regardless of whether OWS made the additional sales, their exclusion is appropriate under the incremental-profit standard. The gross margin is therefore a reasonable — and conservative — proxy for incremental margin.
We note that for purposes of this memorandum, we use gross margin as the incremental margin proxy. Dr. Keating may refine this figure in his expert report by performing a more granular analysis of OWS's cost structure to determine whether certain components of COGS are truly fixed rather than variable. For example, certain testing costs or quality assurance overhead may be partially fixed and should be excluded from the incremental cost calculation. If Dr. Keating's analysis identifies such fixed-cost components within COGS, the resulting incremental margin would be higher than 47.3%, which would increase the lost-profits estimate. By using the 47.3% gross margin figure, we present a conservative baseline.
C. Lost-Profits Calculation
The lost-profits calculation follows a four-step methodology:
Step 1 — Accused-Product Revenue Base
The starting point is the total revenue generated by Lumicore's accused products over the infringement period. Dr. Keating's preliminary analysis projects total accused-product revenue from Q3 2022 through Q4 2025 at approximately $714.6 million.
This projection encompasses the cumulative revenue from all three accused product lines:
The product-level projections will be finalized by Dr. Keating based on Lumicore's actual financial data (including quarterly revenue figures produced in discovery), market growth rate assumptions, and any adjustments for product lifecycle dynamics (e.g., the LumiLink 400G-ZR+ may experience declining growth rates as the market transitions to 800G and 1600G products, while the LumiEdge 1600G-LR is in a rapid growth phase). We anticipate that the Second Set of Requests for Production served on November 4, 2024, will yield additional granular quarterly revenue data that will enable Dr. Keating to refine these projections.
Step 2 — Market-Share Displacement
Not all of Lumicore's accused-product sales would have been captured by OWS "but for" the infringement. The market-share displacement analysis — detailed in the separate Market Share Displacement Analysis memorandum (Appendix A) — determines the percentage of Lumicore's sales that OWS would have made in a hypothetical "but for" world without infringement.
The displacement percentage is 38.2%. This figure is derived from a customer-by-customer analysis of the twelve hyperscale accounts where OWS and Lumicore compete head-to-head, taking into account:
• OWS's competitive position and qualification status at each account;
• The relative market share of OWS, Lumicore, and other competitors at each account;
• Customer-specific purchasing patterns, including the extent to which customers dual-source or multi-source transceiver modules;
• The availability and qualification status of alternative suppliers at each account; and
• The proportion of Lumicore's sales made at accounts where OWS was not an approved or competitive vendor.
Applying the 38.2% displacement fraction to the projected accused-product revenue yields the following:
$714.6M × 38.2% = approximately $273.0 million in displaced OWS revenue
This $273.0 million represents the revenue OWS would have earned on additional sales of OWS-400ZR, OWS-800XR, and OWS-1600DR modules to the customers who instead purchased Lumicore's accused products.
Step 3 — Profit Margin Application
Applying OWS's 47.3% gross margin to the displaced revenue yields the lost-profits estimate:
$273.0M × 47.3% = approximately $129.1 million
Rounding and minor adjustments in Dr. Keating's preliminary model — reflecting, among other things, product-mix effects (different products carry slightly different margins) and timing effects (margins may vary across quarters) — yield the $128.9 million estimate used in the complaint demand.
Step 4 — Cross-Check: Reasonable Royalty on Remaining Sales
The remaining accused-product revenue not addressed by the lost-profits theory is:
$714.6M − $273.0M = approximately $441.6 million
This remaining revenue is addressed under the reasonable-royalty theory, at the proposed royalty rate of 8.5% derived from the Georgia-Pacific analysis in the separate Comparable License Analysis memorandum:
$441.6M × 8.5% = approximately $37.5 million
The combined lost-profits and reasonable-royalty damages are therefore:
$128.9M + $37.5M = $166.4 million (before any willful-infringement enhancement)
This combined figure represents the total compensatory damages OWS would seek at trial, before consideration of enhanced damages for willful infringement under 35 U.S.C. § 284.
D. Comparison to Lumicore's Own Margins
Lumicore's product-level financial data — produced in discovery under the designation "Highly Confidential — Attorneys' Eyes Only" — provides an instructive comparison and may be relevant at trial in several respects.
Lumicore's gross margins on the accused products are as follows:
These margins are higher than OWS's 47.3% gross margin. Lumicore's superior margins may be relevant to several issues at trial:
(a) Rebutting a price-driven-market argument. If Lumicore argues that the silicon-photonics transceiver market is primarily price-driven (i.e., that customers purchase on price rather than technology), Lumicore's own high margins undercut that argument. A margin of 52-58% is inconsistent with a commodity market where price is the sole competitive differentiator; it is consistent with a technology-driven market where the patented features command premium pricing.
(b) Supporting willful-infringement enhancement. Lumicore earned substantial profits from the infringing products — estimated gross profits of approximately $143.8 million on FY 2023 accused-product revenue of $264.9 million. These profit levels reinforce the inference that Lumicore had a powerful economic incentive to continue infringing despite receiving actual notice of OWS's patent rights, supporting the case for enhanced damages under § 284.
It is also noteworthy that the accused products represent approximately 84.7% of Lumicore's total FY 2023 revenue ($264.9M of $312.6M). This demonstrates that the accused products are not a peripheral product line for Lumicore; they are the overwhelming core of Lumicore's business. This fact underscores both the magnitude of the harm to OWS and the strength of Lumicore's economic incentive to infringe.
E. Conclusion on Factor Four
Factor Four is satisfied. OWS's lost profits on the displaced sales are approximately $128.9 million, based on a 47.3% gross margin applied to $273.0 million in displaced revenue.
This estimate is conservative in several important respects:
(a) It uses gross margin (47.3%) rather than a potentially higher incremental margin that may result from Dr. Keating's more detailed cost analysis.
(b) It limits the displacement fraction to 38.2% of accused-product sales, reflecting a rigorous customer-by-customer analysis rather than a more aggressive assumption.
(c) It does not include price-erosion effects — i.e., the profits OWS lost on sales it actually made but at reduced prices due to Lumicore's competitive pressure. Price-erosion damages are addressed in the separate Price Erosion Analysis memorandum (Appendix B), which documents an 18.7% decline in average selling prices for 400G-ZR transceivers from Q3 2022 to Q4 2023 attributable in part to Lumicore's market entry with infringing products.
(d) It does not include lost profits on convoyed sales of complementary products (e.g., optical cables, management software, and support services), which are discussed below in Section VII.B.
VII. ADDITIONAL CONSIDERATIONS
A. Price Erosion
In addition to lost-volume profits — i.e., profits on sales that OWS lost entirely to Lumicore — OWS may also be entitled to recover profits lost due to price erosion on sales that OWS actually made. Price erosion occurs when the presence of an infringing competitor in the market forces the patentee to reduce its prices in order to retain customers and maintain market share. The result is that the patentee earns lower margins on the sales it does make, which constitutes an additional element of compensable harm. See Rite-Hite, 56 F.3d at 1548 (recognizing price erosion as a compensable component of lost-profits damages).
The separate Price Erosion Analysis memorandum, prepared in January 2025, documents that the entry of Lumicore's infringing products into the market caused average selling prices ("ASPs") for 400G-ZR transceivers to decline by 18.7% from Q3 2022 (the quarter of Lumicore's market entry) to Q4 2023. While not all of this price decline is attributable to Lumicore's entry — some portion reflects natural market maturation and volume-driven cost reductions — the analysis identifies a substantial portion that is causally linked to Lumicore's competitive pressure.
Under Rite-Hite, a patentee may recover both lost-unit profits (on diverted sales) and profits lost due to price erosion (on retained sales). These are distinct and non-overlapping damages categories. The $128.9 million lost-profits estimate presented in this memorandum focuses exclusively on lost-volume profits and does not include any price-erosion component. Dr. Keating should address price erosion in his expert report as a supplemental damages component, quantifying the additional profits OWS lost due to reduced ASPs on its own sales during the infringement period.
B. Convoyed Sales
OWS may also have lost sales of complementary products that are typically sold alongside transceiver modules. These include OWS-branded optical cables, network management software, and support services. In the hyperscale data-center market, transceiver module sales frequently drive associated purchases of these complementary products, as customers prefer to source an integrated solution from a single vendor to simplify deployment, ensure compatibility, and streamline support.
Under Rite-Hite, convoyed sales — i.e., sales of unpatented products that the patentee would have made alongside the patented product — are recoverable if there is a functional relationship between the patented product and the convoyed product and the patentee would have made those convoyed sales "but for" the infringement. Rite-Hite, 56 F.3d at 1550. The functional-relationship requirement ensures that convoyed-sales damages are limited to complementary products that are genuinely associated with the patented product, rather than unrelated products that happen to be sold by the same company.
This analysis is preliminary. Dr. Keating should evaluate convoyed-sales data provided by OWS's sales team — including historical attach rates for cables, software, and services on transceiver module orders — to determine the magnitude of any convoyed-sales damages and whether they can be established with sufficient evidentiary support.
C. Interaction with Reasonable Royalty Theory
As noted throughout this memorandum, the lost-profits theory and the reasonable-royalty theory are applied to different segments of Lumicore's accused-product sales:
• Lost profits (38.2% of accused-product revenue): Applied to the sales OWS would have made "but for" the infringement — i.e., sales to customers where OWS was an approved vendor, competed head-to-head with Lumicore, and would have captured the sale in a world without infringement.
• Reasonable royalty (61.8% of accused-product revenue): Applied to the remaining sales — i.e., sales to customers where OWS was not an approved vendor, did not have a competitive position, or would not have captured the sale for other reasons. For these sales, OWS cannot show that the infringement caused a lost sale; however, OWS is still entitled to a reasonable royalty for Lumicore's unauthorized use of the patented technology.
The two theories are complementary, not duplicative. There is no double-counting: each dollar of accused-product revenue is assigned to either the lost-profits category or the reasonable-royalty category, but not both.
The reasonable-royalty rate of 8.5% is derived from the Georgia-Pacific analysis presented in the separate Comparable License Analysis memorandum (Appendix C). The comparable licenses supporting this rate include:
• OWS-Terralink Semiconductor License Agreement (6.5% running royalty)
• OWS-NovaWave Corp. License Agreement (9.0% running royalty)
• OWS-Helix Integrated Photonics License Agreement (7.75% running royalty)
• Stratos Photonics v. ArcLight Systems settlement (8.0% effective royalty rate)