IN THE CIRCUIT COURT OF THE STATE OF OREGON FOR THE COUNTY OF MULTNOMAH
STATE OF OREGON, by and through Attorney General DIANE CALLOWAY,
Plaintiff,
v.
HALCYON SEMICONDUCTOR AMERICAS INC., a Delaware corporation,
Defendant.
Case No. 23CV-18274
Hon. Robert Castañeda
DECLARATION OF DR. FRANKLIN MORROW, PH.D., IN SUPPORT OF DEFENDANT'S OPPOSITION TO PLAINTIFF'S MOTION FOR PRELIMINARY INJUNCTION
DECLARATION OF DR. FRANKLIN MORROW, PH.D.
I, Dr. Franklin Morrow, Ph.D., hereby declare as follows:
I. INTRODUCTION AND QUALIFICATIONS
A. Identity and Engagement
1. My name is Dr. Franklin Morrow. I am the founder and Principal of Stonebridge Analytics LLC, an economic consulting firm specializing in antitrust economics, competition policy, and the economics of technology markets. Stonebridge Analytics is located at 1700 K Street NW, Suite 850, Washington, DC 20006. I submit this declaration based on my personal knowledge and professional expertise in the field of antitrust economics.
2. I have been retained by Calderwood & Harkness LLP on behalf of Defendant Halcyon Semiconductor Americas Inc. ("Halcyon Americas"), a wholly owned subsidiary of Halcyon Semiconductor AG (FSE: HLC), for the purpose of providing expert economic analysis regarding the competitive effects of Halcyon's acquisition of Meridian Photonics Inc. ("Meridian") and to respond to the economic theories of harm advanced by the State of Oregon (the "AG" or "Plaintiff") in support of its Motion for Preliminary Injunction filed on August 15, 2023. I have been asked to assess whether the AG's economic arguments and the supporting analysis of its expert, Dr. Evelyn Ng of Ridgeline Economic Consulting LLC, are well-founded as a matter of antitrust economics, and whether the competitive conditions in the relevant markets support the relief sought by the AG.
3. My compensation for this engagement is billed at standard hourly rates through Stonebridge Analytics LLC. My compensation is not contingent on the outcome of this proceeding, nor is it contingent on the substance of my opinions or the content of my testimony. I have no financial interest in Halcyon Semiconductor AG, Halcyon Semiconductor Americas Inc., or any other party to this litigation.
B. Professional Background and Qualifications
4. I hold a Ph.D. in Economics from the University of Chicago, where my doctoral research focused on industrial organization, with a particular emphasis on oligopoly theory, innovation competition, and the economics of merger analysis. I also hold a Master of Arts in Economics from the University of Chicago and a Bachelor of Arts in Economics, summa cum laude, from Williams College. My doctoral dissertation, completed under the supervision of Professor Harold R. Dietrich, examined the dynamics of innovation competition in technology markets with network effects, a subject of direct relevance to the matters at issue in this proceeding.
5. I have over twenty years of professional experience in antitrust economics, with a specialization in technology markets, the semiconductor industry, and merger analysis. Prior to founding Stonebridge Analytics LLC in 2011, I served for six years as an economist in the Bureau of Economics at the Federal Trade Commission ("FTC"), where I provided economic analysis and advice in connection with merger investigations, monopolization matters, and consent decree negotiations. During my tenure at the FTC, I was involved in the economic review of numerous transactions in the semiconductor, telecommunications, and information technology sectors.
6. I have been qualified as an expert in antitrust economics and have testified in federal and state courts, as well as before federal administrative agencies, on more than thirty occasions. My testimony has addressed issues of market definition, competitive effects analysis, merger simulation, unilateral and coordinated effects, vertical merger theory, innovation competition, and the quantification of merger efficiencies. Courts in multiple jurisdictions have accepted my testimony and relied on my analysis in reaching their conclusions.
7. I have published extensively in peer-reviewed academic journals and professional publications on subjects directly relevant to the issues in this case, including market definition methodology, merger simulation modeling, innovation competition in technology industries, and the economics of vertical integration in semiconductor markets. My publications include articles in the Journal of Industrial Economics, the Antitrust Law Journal, the Review of Industrial Organization, and the International Journal of Industrial Organization, among others. I have also served as a peer reviewer for several of these journals.
8. My curriculum vitae, which sets forth my educational background, professional experience, publications, prior testimony, and other relevant qualifications in full detail, is attached hereto as Exhibit A and incorporated herein by reference.
C. Materials Reviewed
9. In forming the opinions set forth in this declaration, I have reviewed and considered the following materials, among others:
(a) The Complaint filed by the State of Oregon on April 17, 2023;
(b) Defendant's Answer, filed May 5, 2023, and Defendant's Motion to Dismiss brief, filed May 19, 2023;
(c) The Court's Order dated July 7, 2023 denying Defendant's Motion to Dismiss;
(d) Plaintiff's Motion for Preliminary Injunction and all supporting materials filed on August 15, 2023;
(e) The Declaration of Dr. Evelyn Ng of Ridgeline Economic Consulting LLC, submitted in support of Plaintiff's Motion for Preliminary Injunction, together with all exhibits thereto;
(f) Publicly available financial data and filings for Halcyon Semiconductor AG (consolidated FY 2022 revenue of approximately €9.3 billion), Meridian Photonics Inc. (FY 2022 revenue of approximately $385 million), Luminos Systems Corp., Crestwave Optical Technologies Ltd., and other industry participants;
(g) Market research reports on co-packaged optics ("CPO") and silicon photonics transceiver markets published by recognized industry research firms, including reports from Granite Harbor Management Consulting, Meridian Enterprises, and Dunmore Consulting Group;
(h) Halcyon and Meridian business documents produced in discovery in this proceeding, including internal strategic plans, competitive assessments, customer contracts, pricing data, product roadmaps, and internal communications;
(i) Data on customer purchasing patterns, including purchases by Cascadia Cloud Services Inc. (approximately $62 million in Meridian CPO purchases in FY 2022) and Northstar Digital Infrastructure LLC (approximately $28 million in FY 2022);
(j) Meridian's patent portfolio, comprising 47 U.S. patents related to CPO technology, together with accompanying prosecution histories and patent claim charts;
(k) Optical Internetworking Forum ("OIF") 400ZR and 800ZR standards documentation, including implementation agreements and interoperability specifications;
(l) Hart-Scott-Rodino Act ("HSR") filing materials (HSR File No. 20221003-1847) and the early termination grant dated November 14, 2022;
(m) Transaction documents, including the Agreement and Plan of Merger executed September 8, 2022, the senior secured term loan facility agreement, and related closing documents; and
(n) Additional publicly available industry data, academic literature, and economic analyses relevant to the issues addressed herein.
10. I may supplement this declaration in the future as additional materials become available or as the need arises in connection with further proceedings in this matter.
D. Summary of Opinions
11. Based on my analysis and the materials I have reviewed, I have reached the following principal opinions, each of which is explained in detail in the sections that follow:
(1) The AG's proposed relevant product market of "co-packaged optics transceiver modules for data-center applications" is overly narrow and fails to account for the significant competitive constraints imposed by pluggable transceivers, on-board optics, near-package optics, and other optical interconnect solutions that are functionally substitutable for CPO modules.
(2) Even if one were to accept the AG's narrow market definition for purposes of analysis, the market shares and HHI concentration metrics that the AG relies upon are misleading indicators of competitive effects in a rapidly growing, dynamic, and nascent technology market with a projected compound annual growth rate ("CAGR") of approximately 28% through 2027.
(3) Significant competitive constraints exist from Luminos Systems Corp. (approximately 22% share), Crestwave Optical Technologies Ltd. (approximately 11% share), fringe competitors, and potential entrants, and barriers to entry are low, particularly given the open nature of the OIF 400ZR and 800ZR standards.
(4) The merger generates substantial, merger-specific procompetitive efficiencies, including an estimated $90 million in annual cost savings from vertical integration of Halcyon's analog and mixed-signal IC design capabilities with Meridian's silicon photonics platform.
(5) The AG's two primary theories of competitive harm — the "elimination of an independent innovator" theory and the "vertical foreclosure through degraded interoperability" theory — are economically unsupported and inconsistent with market evidence, industry structure, and the commercial incentives of the merged entity.
(6) The AG has not demonstrated that the acquisition will substantially lessen competition as required under ORS 646.730, and a preliminary injunction imposing behavioral restrictions on Halcyon would harm innovation, increase costs for data-center operators — including Oregon-based customers such as Cascadia Cloud Services Inc. and Northstar Digital Infrastructure LLC — and is not warranted by the economic evidence.
II. BACKGROUND ON THE TRANSACTION AND INDUSTRY
A. Overview of the Acquisition
12. On September 8, 2022, Halcyon Semiconductor AG, acting through its wholly owned U.S. subsidiary, Halcyon Semiconductor Americas Inc. (4500 Great America Parkway, Suite 300, Santa Clara, CA 95054), entered into an Agreement and Plan of Merger (the "Merger Agreement") to acquire one hundred percent (100%) of the outstanding shares of Meridian Photonics Inc. The acquisition price was $2.15 billion in cash.
13. The acquisition was funded through a $2.2 billion senior secured term loan facility led by Vereinigte Kreditbank AG (Frankfurt, Germany) as administrative agent, with Nordmark Capital Group (Glenrock Assurance, Switzerland) and Pacific Rim Commercial Bank (Singapore) serving as syndicate members. Of the $2.2 billion available under the facility, $2.15 billion was drawn at the closing of the transaction, with the remaining portion of the total purchase price funded by approximately $50 million of Halcyon's cash on hand. Callister Blackwood & Co., a New York-based investment bank, served as Halcyon's M&A advisor in connection with the transaction. The engagement was led by managing director Jonathan Reiss.
14. On October 3, 2022, the parties submitted an HSR premerger notification filing to the Federal Trade Commission and the Antitrust Division of the U.S. Department of Justice (HSR File No. 20221003-1847). On November 14, 2022, early termination of the HSR statutory waiting period was granted. Notably, neither the FTC nor the DOJ issued a second request for additional information or documentary material in connection with this transaction. The significance of this fact is discussed further in Section VIII below.
15. The acquisition closed on January 31, 2023. Following the closing, Meridian now operates as a wholly owned subsidiary of Halcyon Semiconductor Americas Inc. Meridian's principal headquarters remain at 1820 NW Everett Street, Suite 400, Portland, Oregon 97209. At the time of the acquisition, Meridian employed approximately 1,200 persons. Meridian was founded by Dr. Samuel Ashford, who served as its Chief Executive Officer through the closing. Linnea Johansson served as Meridian's Chief Financial Officer.
B. Overview of Co-Packaged Optics (CPO) Technology
16. Co-packaged optics, commonly referred to by the industry abbreviation "CPO," is an emerging technology for high-speed optical interconnects in data-center environments. CPO transceiver modules are silicon photonics-based devices that integrate optical transceivers directly with network switch application-specific integrated circuits ("ASICs") on the same substrate or within the same package. This co-packaging approach stands in contrast to traditional pluggable transceiver modules (such as those in the QSFP-DD or OSFP form factors), which are physically separate components that plug into front-panel ports on a network switch.
17. The primary application of CPO technology is in high-speed data-center optical interconnects. As data centers have grown in scale and complexity — driven by the exponential growth in cloud computing, artificial intelligence workloads, and high-performance computing — the demand for higher-bandwidth, lower-power, and lower-latency optical interconnect solutions has intensified. CPO technology promises to address these needs by reducing the electrical path length between the switch ASIC and the optical components, thereby lowering power consumption and improving signal integrity at the highest data rates.
18. Key industry standards governing CPO and related technologies include the OIF 400ZR coherent optical interface standard and the emerging OIF 800ZR standard. These standards, developed through the Optical Internetworking Forum, an industry-wide standards body with broad participation, define the optical, electrical, and protocol specifications that allow interoperability among components from different suppliers. The open nature of these standards is a critical fact for the competitive analysis in this case, as I discuss below.
19. It is essential to recognize that CPO is an emerging technology that remains in the early stages of commercialization. While industry projections suggest significant growth in CPO adoption over the next several years, traditional pluggable transceivers currently account for the vast majority of optical transceiver deployments in data centers worldwide. Additionally, other related technologies — including on-board optics ("OBO") and near-package optics ("NPO") — represent intermediate architectural approaches that directly compete with CPO for the same customer deployments. These alternatives provide varying tradeoffs between integration density, thermal management, serviceability, and cost. From the perspective of a data-center customer evaluating interconnect solutions, CPO, pluggable transceivers, on-board optics, and near-package optics are all potential solutions to the same fundamental problem: moving data optically between switches and across data-center facilities. This functional substitutability is central to the proper definition of the relevant product market.
C. Halcyon's Business Pre-Acquisition
20. Halcyon Semiconductor AG is a publicly traded semiconductor company headquartered in Munich, Germany, listed on the Frankfurt Stock Exchange under the ticker symbol FSE: HLC. The company is led by Chief Executive Officer Dr. Tobias Wendt. Katrin Schreiber serves as General Counsel, and Priya Nagarajan serves as U.S. Deputy General Counsel.
21. In FY 2022, Halcyon reported consolidated revenue of approximately €9.3 billion and employed approximately 28,500 persons worldwide. Halcyon's core business focuses on the design and manufacture of analog and mixed-signal semiconductors, power management integrated circuits ("ICs"), and related semiconductor components. These products serve a broad range of end markets, including industrial automation, automotive, telecommunications infrastructure, and data-center equipment.
22. Halcyon maintained a legacy CPO product line prior to the acquisition, but its CPO products were focused on control ICs and driver electronics — the semiconductor components that support and enable the operation of optical transceivers — rather than on the full photonics integration that defines a CPO transceiver module. Halcyon's legacy CPO line represented approximately 14% of the U.S. CPO transceiver market by revenue, accounting for approximately $122.5 million of the approximately $875 million total U.S. CPO market in FY 2022. The nature of Halcyon's pre-acquisition CPO business is important because it was in many respects complementary to, rather than directly competitive with, Meridian's core silicon photonics transceiver business. This complementarity is the source of many of the procompetitive efficiencies I discuss in Section VI.
D. Meridian's Business Pre-Acquisition
23. Meridian Photonics Inc. was founded by Dr. Samuel Ashford and is headquartered in Portland, Oregon. Meridian's core business is the design, development, and manufacture of silicon photonics transceivers and co-packaged optics modules for data-center and telecommunications applications.
24. In FY 2022, Meridian reported revenue of approximately $385 million. Based on this revenue, Meridian's estimated share of the U.S. CPO transceiver market was approximately 44% (i.e., approximately $385 million of the approximately $875 million total U.S. CPO market). Meridian holds 47 U.S. patents related to CPO technology, covering aspects of silicon photonics design, optical modulation, wavelength-division multiplexing for co-packaged applications, and related innovations. While these patents represent a significant portfolio, they constitute only a subset of the broader intellectual property landscape in CPO technology, as I discuss further below.
25. Meridian's customer base includes major data-center operators with significant Oregon presences. Key Oregon-based customers include Cascadia Cloud Services Inc., which operates large-scale data-center facilities in Prineville and The Dalles, Oregon, and which spent approximately $62 million on Meridian CPO modules in FY 2022, and Northstar Digital Infrastructure LLC, which operates a data-center campus in Hillsboro, Oregon, and which spent approximately $28 million on Meridian CPO modules in FY 2022. While these are substantial customers, they represent a fraction of Meridian's total sales and an even smaller fraction of total U.S. data-center optical interconnect spending.
III. THE AG'S RELEVANT MARKET DEFINITION IS OVERLY NARROW
A. The AG's Proposed Market Definition
26. In its Complaint and in its Motion for Preliminary Injunction, the AG alleges a relevant product market consisting of "co-packaged optics transceiver modules for data-center applications." The AG further alleges that the relevant geographic market is the United States, with particular emphasis on the impact of the transaction on Oregon-based data-center customers. Dr. Evelyn Ng of Ridgeline Economic Consulting LLC has submitted a declaration in support of the AG's motion opining in favor of this narrow market definition. I have carefully reviewed Dr. Ng's analysis and find it to be fundamentally flawed for the reasons set forth below.
27. The AG's proposed market definition is critical to its competitive harm theory. If the market is defined narrowly to include only CPO transceiver modules, the post-merger combined entity appears to hold a significant share of the market and concentration metrics appear elevated. However, if the market is properly defined to include the broader set of optical interconnect solutions that compete with CPO modules for the same customers and applications — as antitrust economics requires — the competitive picture changes dramatically. Proper market definition is therefore the threshold issue in evaluating the AG's motion.
B. The Product Market Should Include Broader Optical Interconnect Solutions
28. The AG's proposed product market ignores significant competitive constraints from adjacent product categories that directly discipline the pricing, innovation, and commercial conduct of CPO transceiver suppliers. Under well-established principles of antitrust market definition, the relevant product market must include all products that are reasonably interchangeable by consumers for the same purposes. The AG's narrow definition fails this test.
29. Traditional pluggable optical transceivers remain the dominant technology. Traditional pluggable optical transceivers in the QSFP-DD and OSFP form factors remain the dominant technology for data-center interconnects today and are the workhorses of data-center networking infrastructure. These pluggable modules are functionally substitutable for CPO modules in the vast majority of current data-center deployments. They perform the same fundamental function — converting electrical signals to optical signals and vice versa to enable high-speed data transmission over fiber-optic links within and between data-center facilities. While CPO technology offers potential advantages in power consumption and latency at the highest data rates, pluggable transceivers offer advantages in serviceability, modularity, and mature supply chains. For the vast majority of data-center applications today, either technology can fulfill the customer's interconnect requirements.
30. On-board optics and near-package optics are intermediate alternatives. On-board optics ("OBO") and near-package optics ("NPO") represent intermediate technological approaches that sit on a spectrum between traditional pluggable transceivers and full co-packaged optics. OBO integrates optical engines onto the switch motherboard near the ASIC but not within the same package, while NPO positions optical components adjacent to the ASIC package. These approaches directly compete with CPO for the same customer budgets and deployment architectures, offering different tradeoffs between integration density, thermal management, serviceability, and cost. Both OBO and NPO are being actively developed and deployed by major switch and transceiver vendors. To exclude them from the relevant market while including only CPO is to draw an arbitrary technological boundary that does not correspond to the actual competitive dynamics observed in the marketplace.
31. Customer evidence confirms cross-technology substitutability. The customer evidence available in this case confirms that data-center operators, including Cascadia Cloud Services Inc. and Northstar Digital Infrastructure LLC, do not evaluate CPO modules in isolation from other optical interconnect options. Rather, these sophisticated purchasers evaluate CPO modules alongside pluggable transceivers, on-board optics, and near-package optics as part of their procurement processes. Their purchasing decisions are driven by total cost of ownership, power consumption per port, latency, thermal design constraints, ease of field service, and multi-vendor qualification — not by whether a module uses co-packaging architecture. The evidence demonstrates that these customers actively weigh the relative merits of CPO against pluggable transceivers and other solutions for each deployment decision, which is precisely the kind of demand-side substitutability that determines the relevant product market.
32. A hypothetical monopolist test confirms the market is too narrow. The standard analytical tool for defining relevant product markets in antitrust analysis is the hypothetical monopolist test, also known as the SSNIP (small but significant and non-transitory increase in price) test. Under this test, the relevant market is defined as the narrowest set of products such that a hypothetical monopolist of those products could profitably impose a small but significant (typically 5-10%) non-transitory increase in price. Applied to the AG's proposed market of CPO transceiver modules alone, the hypothetical monopolist test would fail. A 5-10% increase in CPO module prices by a hypothetical CPO monopolist would cause substantial substitution to pluggable transceivers, which are currently cheaper on a per-port basis and which can fulfill the same interconnect function for most current deployments. This price-induced substitution would render the hypothetical price increase unprofitable, demonstrating that CPO modules alone do not constitute a relevant antitrust market. Dr. Ng's analysis does not adequately address this substitution dynamic, and her attempt to distinguish CPO modules on the basis of their "unique" integration characteristics conflates product differentiation with market definition.
33. Including broader optical transceiver solutions dramatically alters market shares. If the relevant product market is properly defined to include pluggable transceivers, on-board optics, and near-package optics in addition to CPO modules — reflecting the true scope of competitive alternatives available to data-center customers — the combined entity's market share would be dramatically lower than the figures cited by the AG. The total U.S. optical transceiver market (including all form factors and architectures) is estimated to exceed $7 billion annually. Against this larger market, the combined Halcyon/Meridian revenues of approximately $507.5 million would represent a share of approximately 7%, far below any level that could reasonably be considered anticompetitive. Even if one were to define a somewhat narrower market focused on "high-speed optical interconnects for hyperscale data centers" (excluding enterprise and telecom applications), the combined entity's share would remain well below the levels at which competitive concerns would ordinarily arise.
C. The Geographic Market Is Properly National or Global
34. The AG alleges a U.S. geographic market with particular emphasis on Oregon-based customers. While I do not dispute that the United States is an appropriate geographic frame of reference as an outer bound, the AG's emphasis on Oregon-based customers as a particular locus of competitive harm is misplaced and analytically unsound.
35. CPO transceiver modules are lightweight, high-value products that are shipped worldwide, with transportation costs that are negligible relative to the value of the product. There are no meaningful tariffs, regulatory barriers, or logistical constraints that would prevent Oregon-based data-center operators from sourcing CPO modules from any supplier globally. Indeed, the evidence shows that Oregon data-center operators currently source CPO modules from suppliers located across the United States and internationally, including Crestwave Optical Technologies Ltd., which is headquartered in Shenzhen, China, and ships product directly to U.S. customers.
36. There are no meaningful regional price differences or supply constraints within the United States that would justify a geographic market narrower than the nation as a whole. CPO module pricing is substantially uniform across U.S. customers, regardless of their location. Oregon-based customers such as Cascadia Cloud Services and Northstar Digital Infrastructure pay the same prices and receive the same products as data-center operators located in Virginia, Texas, Iowa, or any other state with significant data-center concentrations.
37. Even within the AG's proposed national geographic market, the emphasis on Oregon-based customers is misplaced. Cascadia Cloud Services and Northstar Digital Infrastructure, while important Meridian customers, represent a small fraction of total U.S. demand for CPO modules and an even smaller fraction of the total U.S. optical transceiver market. There is no basis in antitrust economics for singling out customers in a particular state as a relevant geographic submarket when those customers face no unique supply constraints, pay no premium prices, and have full access to the same set of global suppliers as customers elsewhere in the United States.
IV. MARKET SHARES AND CONCENTRATION METRICS IN CONTEXT
A. Summary of Market Share Data
38. For purposes of completeness, and without necessarily accepting the AG's narrow product market definition as correct, I present below the estimated market share data for the U.S. CPO transceiver module market in FY 2022 as relied upon by the AG. I note that my analysis of these figures does not constitute an endorsement of the AG's market definition; rather, I address these figures to demonstrate that even under the AG's own framework, the competitive harm theory is not well-supported.
39. Under these figures, the post-merger combined Halcyon/Meridian share would be approximately 58% (44% + 14%), representing combined revenues of approximately $507.5 million in the U.S. CPO transceiver market.
B. HHI Analysis
40. The Herfindahl-Hirschman Index ("HHI") is a commonly used measure of market concentration in antitrust analysis. The HHI is calculated by summing the squares of the market shares of all firms in the market. Using the AG's market share figures, the HHI calculations are as follows:
Pre-Merger HHI: 44² + 22² + 14² + 11² + 9² = 1,936 + 484 + 196 + 121 + 81 = 2,818
Post-Merger HHI: 58² + 22² + 11² + 9² = 3,364 + 484 + 121 + 81 = 4,050
Change in HHI (Delta): 4,050 − 2,818 = 1,232
41. I acknowledge that these figures exceed the thresholds set forth in the U.S. Department of Justice and Federal Trade Commission Horizontal Merger Guidelines (2010), which provide that mergers resulting in a post-merger HHI above 2,500 with a delta exceeding 200 are presumed to be likely to enhance market power. These thresholds are the basis for the AG's and Dr. Ng's reliance on concentration metrics as evidence of competitive harm.
42. However, it is well-established in antitrust economics — and recognized by the Merger Guidelines themselves — that HHI thresholds are screening tools that create a rebuttable presumption, not dispositive proof of anticompetitive effects. The Guidelines expressly recognize that market-specific factors may override the structural presumption created by concentration metrics. This is particularly true in nascent, rapidly growing technology markets where market shares are volatile, entry is likely, and current revenue-based shares are poor predictors of future competitive dynamics. As I explain in the following subsection, each of these conditions obtains in the CPO transceiver market.
C. Why Static Market Shares Are Misleading in This Market
43. The AG's reliance on static, backward-looking market share figures as the centerpiece of its competitive harm theory is misguided in the context of the CPO transceiver market. There are several reasons why current market shares are particularly poor indicators of future competitive significance in this market.
44. Rapid market growth. The U.S. CPO transceiver market grew by approximately 45% between FY 2021 and FY 2022. Industry forecasts from multiple independent research firms project a compound annual growth rate ("CAGR") of approximately 28% through 2027, implying that the total U.S. CPO market could exceed $3 billion by 2027. In a market growing at this rate, the total addressable market will more than triple over the next four years. In such an environment, no single firm can maintain a dominant share simply by serving its existing customer base; the overwhelming majority of future demand will come from new deployments and new customer relationships, which will be contested by all competitors. Revenue-based market shares calculated from a single year's data in a market of this character are inherently unreliable as predictors of future market structure.
45. Technology markets frequently exhibit volatile shares. The history of technology markets is replete with examples in which high initial market shares dissipated rapidly as competitors entered, scaled, and innovated. The transition from one generation of technology to the next frequently reshuffles competitive positions. For example, in the pluggable transceiver market itself, the leading firms of the early 2000s are not the leading firms today; successive waves of innovation in form factors (SFP, SFP+, QSFP, QSFP-DD, OSFP) created opportunities for new entrants and allowed smaller competitors to gain substantial share at the expense of incumbents. There is no reason to expect the CPO market to be different, particularly given the open nature of the relevant standards and the ongoing investment by multiple well-capitalized competitors.
46. Competitors are gaining share and expanding capacity. Within the AG's own narrow market, competitive dynamics are already shifting. Luminos Systems Corp. has been gaining market share and has publicly announced plans for expanded manufacturing capacity at its Austin, Texas facility. Crestwave Optical Technologies Ltd. is investing heavily in next-generation CPO platforms, including 800ZR and beyond, and is actively expanding its presence in the U.S. and European markets. These facts, which are documented in publicly available announcements and industry reports, demonstrate that the competitive landscape is dynamic and that Halcyon/Meridian faces real and intensifying competition. The current market share snapshot captures a single moment in a rapidly evolving market, and reliance on that snapshot as evidence of durable market power is economically unsound.
V. COMPETITIVE CONDITIONS: ENTRY, EXPANSION, AND RIVAL VIABILITY
A. Luminos Systems Corp.
47. Luminos Systems Corp. is a Delaware corporation headquartered in Austin, Texas. Luminos is the second-largest CPO transceiver supplier in the United States, with an estimated market share of approximately 22%, corresponding to approximately $192.5 million in FY 2022 U.S. CPO revenue. Luminos is a well-established company with a full-spectrum silicon photonics technology platform, substantial engineering talent, and a growing customer base among hyperscale and enterprise data-center operators.
48. Luminos has publicly announced plans for significant expansion of its manufacturing capacity, including a new fabrication line at its Austin campus that is expected to approximately double its production output for CPO transceiver modules by mid-2024. This expansion is financed through a combination of internally generated cash flow and a recently completed equity offering. Critically, Luminos has an independent silicon photonics technology platform that is not dependent on Halcyon or Meridian intellectual property. Luminos holds its own portfolio of patents and proprietary technology, and it has developed its CPO modules in-house using its own design methodology and fabrication processes.
49. Luminos is well-positioned to compete aggressively for data-center CPO contracts, including contracts with Oregon-based customers. There is no evidence that the acquisition has impaired or will impair Luminos's ability to compete. To the contrary, industry analysts have noted that the Halcyon/Meridian transaction may create opportunities for Luminos to gain share, as some customers may seek to diversify their supply base in response to the merger — a competitive dynamic that directly undercuts the AG's theory of harm.
B. Crestwave Optical Technologies Ltd.
50. Crestwave Optical Technologies Ltd. is a limited company organized under the laws of the People's Republic of China, headquartered in Shenzhen, China. Crestwave holds an estimated 11% share of the U.S. CPO transceiver market, corresponding to approximately $96.3 million in FY 2022 revenue. Crestwave is a substantial and growing competitive force in the CPO market, with access to significant capital resources and lower manufacturing costs than U.S.-based competitors.
51. Crestwave is actively expanding its presence in U.S. and European markets. The company has established a U.S. sales and technical support office and has been qualifying its products with major hyperscale data-center operators. Crestwave is also investing heavily in next-generation CPO platforms, including development of 800ZR and beyond coherent optical CPO modules. These next-generation products will be directly competitive with the combined Halcyon/Meridian product offering and will provide customers with additional sourcing alternatives.
52. The AG's analysis gives insufficient weight to Crestwave as a competitive constraint. While Dr. Ng acknowledges Crestwave's presence in the market, she discounts its competitive significance on the grounds that it is a foreign-domiciled firm. This reasoning is economically unsound. The relevant question is not where a competitor is incorporated, but whether it can and does supply U.S. customers at competitive prices and quality levels. The evidence demonstrates that Crestwave does exactly that. Oregon-based data-center operators have sourced and continue to source CPO modules from Crestwave, and there is no supply chain, logistical, or quality barrier that would prevent Crestwave from expanding its U.S. sales in response to a price increase by the merged entity.
C. Other Potential Entrants and Fringe Competitors
53. In addition to Luminos and Crestwave, other fringe competitors currently account for approximately 9% of the U.S. CPO transceiver market, or approximately $78.7 million in FY 2022 revenue. These fringe competitors include several smaller but technologically capable firms that are actively developing and commercializing CPO transceiver products.
54. More significantly, multiple well-capitalized semiconductor and photonics companies that are not currently active in the CPO market have publicly announced CPO development programs. These potential entrants include companies with existing capabilities in silicon photonics, coherent optical technology, and advanced semiconductor packaging — the key technological competencies required to design and manufacture CPO modules. For firms that already manufacture pluggable transceivers using silicon photonics technology, the transition to CPO is an incremental step that involves adapting existing photonics dies for co-packaging with switch ASICs. This adaptation leverages existing design expertise, fabrication infrastructure, and customer relationships, significantly reducing the time and investment required for entry.
55. The open nature of the OIF 400ZR and 800ZR standards is a critical factor facilitating entry. These standards define the optical interface, modulation format, forward error correction, and management interfaces for coherent optical transceivers. Compliance with these standards is independently verifiable through standardized testing procedures. Because the standards are open and publicly available, any firm with the requisite silicon photonics design and manufacturing capabilities can develop OIF-compliant CPO modules without the need for proprietary technology licenses from incumbents. This substantially reduces a barrier to entry that might otherwise deter new competitors.
56. Additionally, the major hyperscale data-center operators — companies like Cascadia Cloud Services, Northstar Digital Infrastructure, and their global peers — exercise significant countervailing buyer power. These sophisticated purchasers typically qualify multiple suppliers for each component category and actively manage their supply chains to avoid excessive dependence on any single vendor. This multi-sourcing strategy creates strong incentives for data-center operators to support and encourage new entrants, including by providing qualification opportunities, sharing technical requirements, and offering volume commitments to enable new suppliers to achieve scale. The countervailing buyer power of these large customers is a significant competitive constraint that limits the ability of any supplier — including the merged Halcyon/Meridian entity — to exercise market power through price increases or reduced innovation.
D. Ease of Switching
57. CPO modules conforming to OIF standards are substantially interchangeable. The purpose of the OIF 400ZR and 800ZR standards is precisely to ensure interoperability among modules from different suppliers, so that data-center operators can deploy equipment from multiple vendors without compatibility concerns. A CPO module from Luminos, Crestwave, or any other OIF-compliant supplier can be substituted for a Halcyon/Meridian module with minimal technical effort.
58. Data-center operators routinely multi-source optical components, including transceivers, and maintain qualified vendor lists with multiple approved suppliers for each component category. The switching costs associated with moving from one CPO supplier to another are low relative to the value of purchases. The primary switching cost is the time and effort required to qualify a new supplier's product, which typically takes three to six months for a new transceiver product. This qualification period, while not trivial, is a modest barrier that is readily overcome, particularly given the typical multi-year planning horizons of data-center operators and their strong incentives to maintain multiple qualified supply sources.
59. The combination of open standards, functional interchangeability, routine multi-sourcing, and low switching costs means that the merged entity faces strong competitive discipline from existing competitors and potential entrants. Any attempt by Halcyon/Meridian to raise prices, reduce quality, or slow innovation would result in rapid customer substitution to alternative suppliers, as customers can and do switch between qualified suppliers without incurring significant costs.
VI. PROCOMPETITIVE EFFICIENCIES OF THE MERGER
A. Vertical Integration Efficiencies
60. The combination of Halcyon's analog and mixed-signal IC design expertise with Meridian's silicon photonics platform creates unique vertical integration efficiencies that are substantial, cognizable, and merger-specific. These efficiencies represent a significant procompetitive benefit of the transaction that must be weighed against any alleged competitive harm.
61. Halcyon designs and manufactures the power management ICs, transimpedance amplifiers, and driver electronics that are critical companion components to CPO transceiver modules. These semiconductor devices provide the electrical power, signal conditioning, and control functions that enable the optical components within a CPO module to operate. Prior to the merger, Halcyon sold these companion ICs to various transceiver manufacturers, including Meridian, on an arm's-length basis. The integration of these two complementary capabilities within a single firm allows the co-design and co-optimization of the electrical and optical components on a single integrated platform, eliminating the interface inefficiencies and design compromises inherent in a multi-vendor approach.
62. I have estimated the annual cost savings from vertical integration at approximately $90 million, derived from three principal sources:
(a) Elimination of redundant R&D spending on parallel interface designs: approximately $35 million annually. Prior to the merger, both Halcyon and Meridian independently invested in developing interface technologies for the boundary between the electrical driver/control ICs and the optical photonics components. This required both companies to maintain duplicative engineering teams, develop parallel test and validation infrastructure, and engage in iterative design-revision cycles to ensure interoperability across the electrical-optical interface. The merger eliminates this duplication by allowing a single, unified engineering team to design the electrical and optical components as an integrated system from the outset, rather than designing them independently and then working to ensure compatibility.
(b) Manufacturing efficiencies from co-packaging Halcyon driver ICs with Meridian photonics dies: approximately $30 million annually. Co-packaging the driver electronics and optical components within the same module requires precise alignment, wire bonding or flip-chip attachment, and shared thermal management. When these components are designed by different companies, the packaging process involves additional design margins, compatibility testing, and yield losses. When the components are designed as a unified system by a single company, the packaging process can be optimized holistically, reducing material waste, improving manufacturing yields, and increasing throughput.
(c) Supply chain consolidation and procurement savings: approximately $25 million annually. The combination of Halcyon's and Meridian's supply chains allows the merged entity to consolidate purchasing of shared raw materials and components (such as silicon wafers, packaging substrates, and fiber-optic materials), negotiate volume-based discounts, and reduce logistics and inventory carrying costs. These procurement savings are a direct result of the increased scale and scope of the combined entity's operations.
63. These cost savings are merger-specific. They could not be achieved through arm's-length contracting, licensing arrangements, or strategic partnerships between Halcyon and Meridian as independent companies. The efficiencies arise from the integration of complementary engineering teams, the co-design of components on a unified platform, and the consolidation of manufacturing and supply chain operations — none of which can be fully realized without common ownership and unified management control. The AG's suggestion that these efficiencies could be achieved through alternative means such as supply agreements or joint ventures does not withstand scrutiny. Arm's-length contracts cannot replicate the information sharing, organizational alignment, and unified decision-making that are necessary to capture the full co-design and co-packaging benefits described above.
B. Innovation and Product Development Benefits
64. Beyond the quantifiable cost savings described above, the merger generates significant innovation and product development benefits. The merged entity is positioned to accelerate the development of next-generation CPO products — including 1.6 terabit-per-second (1.6T) and beyond platforms — by combining Halcyon's world-class IC design capabilities with Meridian's portfolio of 47 CPO-related U.S. patents and its deep expertise in silicon photonics design and fabrication.
65. Prior to the merger, both Halcyon and Meridian were independently developing interface technologies that needed to interoperate across the electrical-optical boundary. This required each company to devote substantial engineering resources to understanding and accommodating the design constraints of the other company's components, often with imperfect information. The merger eliminates this duplicative effort and enables a unified product roadmap in which the electrical and optical components are designed together as an integrated system, accelerating development timelines and enabling performance optimizations that would not be possible in a multi-vendor architecture.
66. The acceleration of next-generation CPO product development has direct benefits for data-center customers, who will gain earlier access to higher-performance, lower-power interconnect solutions. Given the critical role of optical interconnects in the overall performance and energy efficiency of modern data centers, faster innovation in CPO technology translates into meaningful improvements in data-center economics and environmental sustainability. These innovation benefits are a procompetitive consequence of the merger that the AG's analysis entirely ignores.
C. Pass-Through of Efficiencies to Consumers
67. The competitive pressure from Luminos Systems Corp., Crestwave Optical Technologies Ltd., and other existing and potential competitors ensures that the cost savings generated by the merger will be substantially passed through to customers in the form of lower prices and improved products. In a market with multiple well-positioned competitors, open standards facilitating entry, and sophisticated buyers with countervailing market power, the merged entity cannot simply retain efficiency gains as increased profit margins; competitive dynamics compel the firm to share those gains with customers.
68. Indeed, the post-merger pricing evidence confirms this competitive dynamic. Since the acquisition closed on January 31, 2023, Halcyon has not raised prices on Meridian CPO modules. To the contrary, per-unit pricing for Meridian CPO modules has declined approximately 8% since the closing, reflecting both the realization of early integration efficiencies and ongoing competitive pressure from rivals. This pricing trend is inconsistent with the AG's theory that the merger has enhanced Halcyon's market power or its ability to raise prices.
69. Oregon-based customers, including Cascadia Cloud Services and Northstar Digital Infrastructure, continue to have access to competitively priced CPO modules from multiple sources. There is no evidence that these customers have experienced any deterioration in pricing, quality, service, or innovation as a result of the acquisition. To the contrary, Halcyon has maintained Meridian's Portland headquarters, retained the vast majority of Meridian's engineering talent, and continued to invest in the Meridian CPO product line. These facts are inconsistent with the AG's theory that the acquisition has harmed or will harm Oregon-based customers.
VII. RESPONSE TO THE AG'S THEORY OF HARM
A. The "Elimination of Independent Innovator" Theory
70. The first principal theory of competitive harm advanced by the AG is that the acquisition eliminates Meridian as an independent, disruptive innovator in the CPO market, thereby reducing the competitive pressure that Meridian independently exerted on Halcyon and other market participants. This theory is a variant of the "loss of a maverick" or "elimination of a disruptive competitor" theory that has been advanced in certain merger challenges. As applied here, the theory overstates Meridian's pre-acquisition competitive independence and understates the competitive vitality of the remaining market participants.
71. The AG's characterization of Meridian as a uniquely independent competitive force is not supported by the economic evidence. Prior to the acquisition, Meridian was a component supplier that depended on sales to a relatively concentrated group of large data-center customers. It was not an independent competitive force that constrained Halcyon in any meaningful way, because Halcyon and Meridian operated in largely complementary (not directly competing) product spaces: Halcyon in driver ICs and control electronics, and Meridian in silicon photonics modules. To the extent there was overlap in the CPO transceiver market (where Halcyon held a 14% share and Meridian held a 44% share), the competitive constraint on each firm came not uniquely from the other, but from the full range of competitors, including Luminos, Crestwave, pluggable transceiver alternatives, and the threat of entry.
72. Innovation competition in the CPO market is robust and multi-firm. Luminos Systems Corp., Crestwave Optical Technologies Ltd., and numerous other firms are actively investing in CPO research and development, filing patents, and bringing new products to market. The elimination of Meridian as an independent firm does not eliminate the competitive incentive to innovate in the CPO market, because that incentive comes from the full spectrum of competitive pressure — from direct CPO rivals, from alternative optical interconnect technologies (pluggable, on-board, near-package), and from the demanding requirements of sophisticated customers who continuously push their suppliers to deliver better performance at lower cost.
73. Meridian's 47 U.S. patents, while important, represent only a subset of the broader intellectual property landscape in CPO technology. A search of the U.S. Patent and Trademark Office database reveals hundreds of patents and patent applications related to co-packaged optics, silicon photonics, and coherent optical transceiver technology, held by dozens of companies and research institutions. The notion that Meridian's patent portfolio is so uniquely important that its acquisition by Halcyon would foreclose innovation competition is not supported by the evidence.
B. The Vertical Foreclosure / Degraded Interoperability Theory
74. The second principal theory of competitive harm advanced by the AG is a vertical foreclosure theory. The AG alleges that Halcyon's vertically integrated position — combining Meridian's CPO transceiver modules with Halcyon's control ICs and driver electronics — would allow Halcyon to degrade the interoperability between Meridian CPO modules and third-party switch ASICs, or alternatively, to degrade the compatibility of Halcyon's companion ICs with rival CPO modules, thereby raising rivals' costs and foreclosing competition. This theory is economically implausible for several independent reasons.
75. Halcyon has strong commercial incentives to maintain interoperability. Data-center customers demand multi-vendor compatibility as a fundamental requirement of their procurement processes. Degrading interoperability — whether by making Meridian CPO modules work less well with third-party ASICs, or by making Halcyon companion ICs work less well with rival CPO modules — would directly cause Halcyon/Meridian to lose sales to Luminos, Crestwave, and other competitors whose products interoperate fully with the broader ecosystem. The cost to Halcyon of lost sales would far exceed any potential benefit from raising rivals' costs, making such a strategy commercially irrational. Dr. Ng's analysis fails to grapple with this fundamental incentive problem.
76. OIF standards are open and independently verifiable. The OIF 400ZR and 800ZR standards are open industry standards, publicly available, and compliance with them is independently verifiable through standardized conformance testing procedures. Any attempt by Halcyon to deviate from these standards — whether by introducing proprietary extensions, modifying interface specifications, or otherwise degrading interoperability — would be readily detectable by customers, competitors, and the standards community. Such deviation would be commercially self-defeating, as it would disqualify Halcyon/Meridian products from consideration by customers who require OIF-compliant components, which is to say virtually all customers in the data-center CPO market.
77. Bundling or degrading interoperability would jeopardize Halcyon's broader IC business. Halcyon's power management ICs and driver electronics are sold to a wide range of customers across multiple end markets, including many customers who also purchase optical components from non-Meridian suppliers. If Halcyon were to degrade the interoperability of its companion ICs with rival CPO modules, it would risk alienating these customers and jeopardizing its broader IC sales, which represent a far larger revenue base than the CPO transceiver business. The AG's vertical foreclosure theory fails to account for this cross-market retaliation effect, which would discipline any attempt at foreclosure.
78. Halcyon has no history of exclusionary conduct. Halcyon's legacy CPO business (14% market share pre-merger) was itself interoperable with third-party switch ASICs from multiple vendors. There is no history, pattern, or evidence of Halcyon engaging in exclusionary conduct through interoperability degradation, proprietary lock-in, or discriminatory pricing in any market in which it participates. The AG's foreclosure theory is a speculative prediction unsupported by any historical evidence of the type of conduct it predicts.
C. HHI Metrics Do Not Establish Competitive Harm
79. As I have discussed in Section IV.B above, the post-merger HHI of 4,050 and the delta of 1,232 exceed the screening thresholds set forth in the DOJ/FTC Horizontal Merger Guidelines. However, these are presumptive screening levels, not proof of anticompetitive effects. The Merger Guidelines themselves recognize that concentration metrics must be interpreted in light of the full range of competitive conditions in the relevant market, including entry, expansion, product substitution, buyer power, and efficiencies.
80. In a rapidly growing technology market with a projected 28% CAGR through 2027, static concentration metrics calculated from a single year of revenue data are particularly poor predictors of future competitive conditions. As I have explained, market shares in such markets are inherently volatile, and the HHI is a static snapshot that does not capture the dynamic competitive forces — entry, expansion, innovation, and technological substitution — that will shape the market going forward.
81. Moreover, the HHI calculation is entirely dependent on the AG's overly narrow product market definition. If the relevant market is properly defined to include broader optical transceiver solutions, the HHI figures would be substantially lower — well below the levels at which competitive concerns would ordinarily arise. The AG's concentration analysis therefore rests on a circular foundation: it yields alarming numbers only because the market has been defined too narrowly, and the narrow market definition is defended in part by reference to those same alarming numbers.
82. Dr. Evelyn Ng's analysis for the AG fails to account for the dynamic competitive effects, ease of entry, and rapid technological evolution of this nascent market. Dr. Ng's analysis relies heavily on static market share snapshots and concentration metrics while giving insufficient attention to the robust and growing competitive constraints from Luminos, Crestwave, and potential entrants. Her analysis also fails to properly weigh the procompetitive efficiencies of the merger or to address the post-merger pricing evidence, which shows an 8% decline in per-unit pricing since the acquisition closed. In my professional opinion, Dr. Ng's analysis does not provide a reliable basis for concluding that the acquisition will substantially lessen competition.
VIII. THE FEDERAL REVIEW PROCESS CORROBORATES THE ABSENCE OF COMPETITIVE HARM
83. The transaction was reviewed by the federal antitrust enforcement agencies under the Hart-Scott-Rodino Antitrust Improvements Act of 1976. The parties submitted their HSR premerger notification filing on October 3, 2022 (HSR File No. 20221003-1847). On November 14, 2022, early termination of the statutory waiting period was granted by the FTC and DOJ.
84. The grant of early termination is significant in several respects. First, it demonstrates that the federal antitrust agencies — which have deep expertise in merger review and access to the full range of investigative tools, including the ability to issue second requests for extensive additional information and documentary material — reviewed the same transaction that the Oregon AG now challenges and determined that it did not warrant further investigation. The agencies chose not to issue a second request, which is the mechanism by which the agencies would have extended the review period and obtained additional information had they identified competitive concerns warranting further scrutiny.
85. I recognize that the grant of early termination under the HSR Act is not a formal approval of the transaction, nor does it constitute a determination by the federal agencies that the transaction is procompetitive or competitively neutral. The agencies may grant early termination for range reasons, including the conclusion that the transaction does not present competitive concerns that warrant the expenditure of additional agency resources. Nevertheless, the fact that the same transaction now challenged by the Oregon AG was reviewed and cleared for early termination by the federal antitrust agencies is relevant context