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.
| Digital Realty Trust | Marvell Technology | NVIDIA | |
|---|---|---|---|
| Moat rating | narrow The FY2025 10-K's own competitive record cuts both ways. Item 1 says a "high-quality, highly interconnected global portfolio such as ours could not be easily replicated today on a cost-competitive basis," yet Item 1A concedes that competitors have "significantly greater financial, marketing and other resources and more ready access to capital" and that as rivals keep developing space, "rental rates may be reduced or we may face delays in leasing." A durable advantage that its own filing says new supply can price against is bounded, not unassailable. | 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 | wide Wide: the CUDA software platform and its developer installed base create high switching costs across the AI-training stack, defended by $76.7B cumulative R&D and reinforced by ~71% FY2026 gross margins and Data Center revenue (~90% of total) up 68% YoY — durable, hard-to-replicate advantages per the FY2026 10-K. |
| Moat type | network effects Item 1 attributes the hard-to-replicate part of the portfolio to connectivity rather than to real estate: "the network density, interconnection infrastructure and connectivity-centric customers in certain of our data centers have led to the organic formation of densely connected data communities that are difficult for competitors to replicate." That community sits on over 232,000 cross connects in over 55 metros, so each network and cloud that lands makes the same building worth more to the next tenant. | 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 | intangibles ip The durable edge rests on proprietary IP and software: the full-stack CUDA development platform running on all NVIDIA GPUs plus hundreds of proprietary domain libraries/SDKs/APIs, and $76.7B cumulative R&D yielding "inventions that are essential to modern computing" (NVIDIA invented the GPU in 1999). This is reinforced by developer-ecosystem network effects and CUDA switching costs, per the FY2026 10-K Business section. |
| Leadership | co leader Item 1 claims the title of "the largest global provider of cloud- and carrier-neutral data center, colocation and interconnection solutions," but the Competition section names Equinix and NTT as operators of properties "similar to ours in some of the same metropolitan areas," plus Global Switch and regional operators abroad — a shared top tier on the company's own telling, not a solitary one. | 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. | clear leader FY2026 Data Center revenue was $193.7B (~90% of the $215.9B total), up 68% YoY on the Blackwell ramp, and NVIDIA describes itself as "a data center scale AI infrastructure company reshaping all industries." The 10-K frames named rivals as parties who "provide or intend to provide" GPUs/accelerators — incumbent-leader positioning. |
| Pricing power | strong FY2025 renewals signed re-priced upward in every bucket — +27.0% on greater-than-1 MW space ($146 to $186 per square foot), +4.6% on 0-1 MW ($268 to $280) and +43.0% on other ($49 to $71) — and MD&A expects average aggregate rental rates on 2026 renewals to be positive against the rates currently paid for the same space "on a GAAP basis and on a cash basis." On costs, the filing says utilities expense "is our largest expense category" and that "the vast majority of the expense is passed directly through to our customers," which it credits with significantly mitigating exposure to power-cost increases rather than removing it. The cap: Item 1A warns competitor development could still force rates down. | 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. | strong FY2026 gross margin was 71.1% (75.0% in FY2025); per the 10-K MD&A the ~3.9pt decline reflects the Hopper HGX→Blackwell full-system mix shift and a one-time $4.5B H20 excess-inventory/purchase-obligation charge, not competitive price erosion. A low-70s% hardware gross margin evidences strong pricing power. |
| Summary | Digital Realty rents space, power and connectivity rather than compute: at 2025 year-end its portfolio held 310 data centers and roughly 57.6 million rentable square feet across more than 55 metros in over 30 countries, about 84.7% leased, serving more than 5,000 customers. Two different businesses sit inside that footprint. The greater-than-1 MW wholesale side is a capital-and-power race — 769 MW of projects underway with 64% pre-leased, and land that "could accommodate over 3,500 megawatts of additional data center capacity" — where the 10-K names Equinix, NTT, Global Switch and "various private operators" as rivals and warns that added supply can push rents down. The colocation and interconnection side is the defended half: over 232,000 cross connects and the "densely connected data communities" Item 1 says competitors cannot easily replicate, reinforced by contracts the filing describes as generally running 5-10+ years on large deployments and by improvements "installed at our customers' expense." FY2025 leasing supports that read — renewals signed re-priced +27.0% on greater-than-1 MW space and +4.6% on 0-1 MW — while customer concentration is the offsetting exposure, with the largest customer at roughly 11.7% of annualized recurring revenue. | 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. | NVIDIA pairs market-leading accelerated-computing hardware (the Blackwell data-center platform) with a proprietary full-stack software moat — CUDA plus hundreds of domain libraries — funded by $76.7B of cumulative R&D, and its "large and growing number of developers and installed base... strengthens our ecosystem and increases the value of our platform for our customers" (FY2026 10-K). Competition is intensifying from AMD, Intel and Huawei, and from hyperscalers (Alphabet, Amazon, Microsoft) designing internal AI silicon, but rivals must overcome NVIDIA's entrenched CUDA software ecosystem to displace it. |
| Chain position | Landlord to the AI stack — sells the space, power and interconnection that cloud, network and enterprise tenants run compute in (Oracle, IBM, Meta Platforms, AT&T, Comcast and Lumen are among the customers named in Item 1), with roughly 2.9 GW of total in-place IT capacity. | 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). | Upstream compute-platform supplier: NVIDIA sells full-stack data-center systems (GPU + Arm CPU + DPU + NVLink/InfiniBand networking + CUDA software) to cloud providers and enterprises, and relies on third-party foundry/assembly-test-packaging partners (e.g., SPIL, Amkor, Wistron, Fabrinet). Per the FY2026 10-K, several of its largest customers (hyperscalers such as Amazon, Alphabet, Microsoft) are simultaneously customers and emerging competitors developing internal accelerated-computing silicon. |
| Products (share / barrier) |
|
|
|
| Long-horizon vote | +0.17 at weight 0.20 · swarm neutral Editorial prior, not backtested. | +0.13 at weight 0.20 · swarm bullish Editorial prior, not backtested. | +0.42 at weight 0.20 · swarm bullish Editorial prior, not backtested. |