Compare moats
Up to three covered companies, band by band. Every call is a curated editorial judgment, never a disclosed figure — and every band carries its cited basis.
| Lumentum Holdings | Super Micro Computer | Marvell Technology | |
|---|---|---|---|
| Moat rating | narrow FY2025 10-K (filed 2025-08-19) grounds a narrow rating: on one side ~1,020 U.S. and ~1,100 foreign patents with expirations through 2045 plus vertically integrated wafer fabs (US, UK, Japan, Thailand) behind its laser chips and photonic ICs; on the other, Item 1A states optical communications products 'have increasingly become subject to commoditization' with 'predominantly Asia-based competitors' intensifying pricing pressure and a historical 'continued decline of average selling prices' expected to continue. | narrow The FY2026 10-K describes a genuine but bounded advantage. On the asset side the filing claims a modular Server Building Block Solutions architecture, over 3,500 R&D employees, in-house design control over many sub-systems, and that Supermicro believes it is 'the only major server, storage, and accelerated compute platform vendor that designs, develops, and manufactures a significant portion of its systems in the United States.' Against that, the same filing calls its market 'highly competitive, rapidly evolving,' concedes that 'most of our competitors have longer operating histories, significantly greater resources, greater name recognition, or deeper market penetration,' reports intensified competition from 'predominantly Asia-based' entrants 'leading to pricing pressure,' and states that pricing pressure has produced 'a continued decline of average selling prices across our business and we expect that these historical trends will continue.' It also disclaims patent dependence outright. A durable-but-shallow advantage in time-to-market and integration, sitting on top of a commoditizing box business, is narrow rather than wide. | narrow FY2026 10-K (filed 2026-03-11): differentiated platform IP — over 10,000 issued patents and pending applications as of 2026-01-31, plus a proven custom ASIC platform leveraging ultra-high-speed SerDes, silicon photonics, co-packaged optics and custom HBM — but Marvell itself calls its markets 'intensely competitive' with 'pricing pressures', notes customers 'have chosen to develop certain semiconductor products internally', and discloses two >=10% customers with the ten largest at 82% of FY2026 net revenue. Real, defensible IP in a concentrated, contestable customer base = narrow, not wide. https://www.sec.gov/Archives/edgar/data/1835632/000183563226000011/mrvl-20260131.htm |
| Moat type | intangibles ip Per the FY2025 10-K, the moat rests on intangibles/IP and process know-how: patent portfolio (~1,020 U.S. / ~1,100 foreign, ~780 pending), InP laser and photonic-integrated-circuit technology (Oclaro, NeoPhotonics acquisitions), coherent DSP/ASIC lines (IPG telecom transmission acquisition), and its own wafer fabrication, assembly and test facilities — the company states it does not broadly license its IP absent adequate consideration or cross-licenses. | cost scale The filing locates the advantage in design-and-manufacturing scale, not in IP or lock-in. It says the company manufactures the majority of its systems at its San Jose headquarters with assembly, test and QC also in Taiwan, the Netherlands and Malaysia, and that this structure lets it 'reduce time to delivery, mitigate the impact of tariffs and regional costs, and reduce overall manufacturing costs.' Its own list of principal competitive factors includes 'cost-effective design and manufacturing' and 'sufficient manufacturing capacity necessary to support market demand.' Intangibles are explicitly ruled out as the source: 'neither our business as a whole nor any of our principal businesses are materially dependent on a single patent,' with reliance placed 'primarily on trade secrets, technical know-how.' Switching costs are weak by the filing's own account: it 'typically sell[s] products pursuant to purchase orders rather than long-term purchase commitments,' and customers 'have, and others may in the future, cancel or defer purchase orders on short notice without incurring a significant penalty.' | intangibles ip The moat rests on hard-to-replicate mixed-signal IP: the 10-K describes the custom ASIC platform built on ultra-high-speed SerDes, ARM compute, security, storage, silicon photonics and advanced packaging (die-to-die interconnects, chiplets, CPO, custom HBM), with multiple 5nm designs executed, 3nm in progress and a 2nm platform in development; a secondary switching-cost element comes from multi-year custom design wins co-developed to individual customer specifications. https://www.sec.gov/Archives/edgar/data/1835632/000183563226000011/mrvl-20260131.htm |
| Leadership | co leader Mixed by layer: the FY2025 10-K self-describes Lumentum as 'a leading provider of optical and photonic products... recognized as an industry leader based on revenue and market share' (company's own characterization), while LightCounting's May 2026 optical vendor landscape places Lumentum near the bottom of its transceiver-only Top 10 — noting the list 'accounts for sales of optical transceivers only, excluding all other products, which explains lower rankings of Coherent and Lumentum' and that 'Neither CIG nor Lumentum plan to stay at the bottom of the list for much longer.' Co-leader in optical components/chips; follower in merchant transceiver modules behind Innolight and Eoptolink. | at parity The filing claims no market-share leadership and gives no share figure. It states 'we believe that we compete favorably with respect to most of these factors' while immediately conceding that 'most of our competitors have longer operating histories, significantly greater resources, greater name recognition, or deeper market penetration,' naming Cisco, Dell, Hewlett-Packard Enterprise and Lenovo plus ODMs Foxconn, Quanta Computer and Wiwynn. The one leadership claim it does make is narrow and structural — that it believes it is the only major vendor in its class manufacturing a significant portion of its systems in the United States — plus a stated goal (not an achievement) to 'be the first to market with superior product designs.' Competing favorably against larger, better-penetrated rivals without asserting share leadership is parity, not leadership. | co leader Leader in its optics niche, #2 in custom silicon: ~60% of high-end PAM4 DSP share (36kr, 2026-06-27, https://eu.36kr.com/en/p/3870758441178373) but an estimated 20-25% of custom AI ASIC design services versus Broadcom's ~70% (hashrateindex, 2026-05-13, https://hashrateindex.com/blog/design-partners-ai-asic-market-part-2/) — net, a co-leader in AI data-center connectivity/custom silicon behind Broadcom overall. |
| Pricing power | moderate The May 5, 2026 Q3 FY2026 earnings release shows current pricing traction — CEO cites 'operational rigor, pricing discipline, and product mix' behind a 540 bps sequential non-GAAP gross-margin gain to 47.9% — but the FY2025 10-K states pricing pressures 'have led to a continued decline of average selling prices across our business and we expect that these historical trends will continue,' so power is cyclical/mix-driven rather than structural. | weak The filing states pricing pressure explicitly and repeatedly. 'Historically, these pricing pressures have led to a continued decline of average selling prices across our business and we expect that these historical trends will continue.' On large orders: 'Large orders are generally subject to intense competition and pricing pressure which can have an adverse impact on our margins and results of operations.' It concedes it has 'accepted customer orders with various types of component pricing protection' which 'increased our exposure to component pricing fluctuations and have adversely affected our financial results in certain quarters,' and that FY2026 industry supply constraints in memory, storage, GPUs and CPUs affected 'the pricing of these items.' Its stated remedy is aspirational rather than realized: it 'must continue to develop more advanced, differentiated products that command a premium.' The filing gives no gross-margin trend in Item 1/1A; margin appears there only as a downside — excess or obsolete inventory 'would reduce our gross margin.' | moderate The 10-K characterizes Marvell's markets as having 'pricing pressures' and intensifying competition (https://www.sec.gov/Archives/edgar/data/1835632/000183563226000011/mrvl-20260131.htm), yet the Q1 FY2027 release reports 52.1% GAAP / 58.9% non-GAAP gross margin on record revenue (https://www.sec.gov/Archives/edgar/data/1835632/000183563226000014/q127_8kx522026ex-991.htm) — differentiated-IP margins, tempered by hyperscaler buyer power. |
| Summary | Lumentum is an optical and photonics supplier whose FY2025 10-K describes two segments: Cloud & Networking ($1,410.8M of $1,645.0M FY2025 revenue) spanning datacom transceivers (Cloud Light), laser chips/EML transmitters, photonic ICs, co-packaged optics, optical circuit switches, and telecom transport (ROADMs, tunable and coherent components, pump lasers); and Industrial Tech ($234.2M) in fiber/ultrafast/diode lasers and 3D-sensing VCSELs. AI datacenter demand has inflected the business: the May 5, 2026 Q3 FY2026 release reports record revenue of $808.4M, up 90.1% YoY, with non-GAAP gross margin of 47.9% and the CEO crediting laser-chip strength, 'scale-across' components (pump lasers, narrow-linewidth laser assemblies), and pricing discipline, with co-packaged optics and optical circuit switches named as the next growth drivers. The moat is real but narrow: differentiation lives in photonic chip IP and vertical fab integration, while the transceiver-module layer commoditizes — LightCounting's May 2026 vendor landscape ranks Lumentum near the bottom of its transceiver-only Top 10 behind Innolight ($5.3B 2025 revenue) and Eoptolink, and Cignal AI (April 21, 2026) sizes 2025 datacom optical component revenue above $19B, up more than 70%, a market where scale rivals are growing as fast or faster. Customer concentration (top customers 16.0% and 15.4% of FY2025 revenue) and the 10-K's ASP-decline language cap the rating below wide. | Supermicro's edge, as its FY2026 10-K frames it, is speed and integration built on a common modular parts bin: it works with NVIDIA, Intel and AMD to 'align the design of our new products with their product release schedules,' then 'quickly assemble a broad portfolio of solutions by leveraging common building blocks across product lines.' During FY2026 it pushed that advantage up a level, growing Data Center Building Block Solutions to 'more than ten key subsystems' — coolant distribution units, heat exchangers, power shelves, switching, management software — and deploying DLC-2 liquid cooling the filing credits with up to 98% per-rack heat capture and up to 40% lower data center power draw. But the filing's competition and risk discussion is candid that this sits inside a price-taking market: ASPs decline structurally, one customer exceeded 10% of net sales in FY2026, and ODMs 'benefit from their scale and very low-cost manufacturing' while increasingly selling their own brands. | Marvell is a fabless data-infrastructure silicon supplier whose center of gravity has shifted decisively to the AI data center: the data center end market was $6,100.3M, 74% of FY2026 revenue, up from 40% two fiscal years earlier (FY2026 10-K). Its strongest position is electro-optics — in high-end PAM4 optical DSPs for 400G+ transceivers it holds roughly 60% share on Inphi-inherited SerDes/FEC IP, with Broadcom above 30%, the two together over 90% (36kr, 2026-06-27). In custom AI silicon it is the structural #2 design partner at an estimated 20-25% of the custom AI ASIC design-services market versus Broadcom's ~70%, anchored by AWS Trainium and Microsoft Maia wins (hashrateindex, 2026-05-13). The Q1 FY2027 release (2026-05-27) shows the flywheel turning — record $2.418B revenue (+28% YoY), Q2 guided to $2.7B mid-point (+35% YoY), management citing 'exceptional AI-related bookings' across 800G/1.6T optics, 51.2T switches, CPO/NPO and custom XPU — and the Celestial AI (Photonic Fabric) and XConn (PCIe/CXL switching) acquisitions closed in February 2026 extend the interconnect moat toward scale-up fabrics. The offsets that keep the moat narrow are in Marvell's own filing: intense competition (AMD, Alchip, Astera, Ayar, Broadcom, Credo, GUC, Lightmatter and others), hyperscaler in-housing risk, and heavy customer concentration. |
| Chain position | Upstream photonics supplier into AI/cloud datacenter and telecom infrastructure: sells transceivers, laser chips, CPO components and OCS to cloud operators, AI infrastructure providers and NEMs, and also supplies components to transceiver makers who compete with it ('Some of these competitors are also our customers' — FY2025 10-K). Concentrated demand: Customer A 16.0% and Customer B 15.4% of FY2025 net revenue; manufacturing concentrated in Thailand, China, UK, Slovenia, Japan and US with sole/limited-source component exposure. | Downstream system integrator: it converts third-party accelerators and CPUs (NVIDIA Blackwell/GB300 NVL72, AMD Instinct MI350, Intel Xeon 6) into validated racks, liquid cooling and full data-center building blocks sold to cloud service providers and enterprises. The AI angle is central, not incidental — the filing devotes a risk factor titled 'The AI industry has driven a significant portion of our recent success' and states 'A portion of the recent success of our server and storage solutions has been dependent on the integration of our products and services within the AI industry.' | Fabless supplier spanning 'data center core to network edge': it sits between hyperscaler AI compute (custom XPU/XPU-attach ASICs) and the optical layer (PAM4/coherent DSPs, CPO/LPO, DCI, AEC, PCIe retimers), outsourcing fabrication to independent CMOS foundries; the Feb-2026 Celestial AI and XConn acquisitions push it further into scale-up photonic fabric and PCIe/CXL/UALink switching (FY2026 10-K, https://www.sec.gov/Archives/edgar/data/1835632/000183563226000011/mrvl-20260131.htm). |
| Products (share / barrier) |
|
|
|
| Long-horizon vote | +0.13 at weight 0.20 · swarm bullish Editorial prior, not backtested. | -0.01 at weight 0.20 · swarm neutral Editorial prior, not backtested. | +0.13 at weight 0.20 · swarm bullish Editorial prior, not backtested. |