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.
| Bloom Energy | NetApp | Marvell Technology | |
|---|---|---|---|
| Moat rating | narrow FY2025 10-K (Item 1, filed 2026-02-09): a proprietary solid-oxide platform protected by 380 active U.S. utility patents plus 183 pending, with speed-to-power the company says can deliver onsite power 'within approximately 90 days' — real but contested advantages: the same 10-K's Item 1A calls distributed generation and hydrogen 'still emerging markets' whose acceptance is uncertain, and revenue concentration was 43%/13%/12% across three customers in FY2025 (sec.gov/Archives/edgar/data/1664703/000162828026006516/be-20251231.htm). | narrow The FY2026 10-K shows a real, durable lock but not an unassailable one. On the durable side: "Our cloud storage services are based on the same ONTAP data management software that underpins our on-premises ONTAP storage infrastructure offerings", and the same filing's income statement shows the company holding a gross margin near 71% across all three reported years - $4,433M on $6,268M in FY2024, $4,613M on $6,572M in FY2025 and $4,899M on $6,925M in FY2026 - while revenue grew from $6,268M to $6,925M and income from operations widened from 19% to 24% of net revenues. Holding that margin through the memory-cost shock the same filing discloses is the commercial evidence the lock is worth something. On the limiting side, the filing says competition "is intense", that in public cloud "customers may choose native cloud services that are consumed as operating expenses", and that "New competitors or alliances among existing competitors could emerge and quickly gain significant market share" - and IDC's 1Q26 external-storage tracker (Blocks & Files, 2026-06-16, cited on the AFF/ASA product row below) ranks NetApp second behind Dell, not first. | 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 The FY2025 10-K states 'Intellectual property is an essential differentiator for our business' and details 380 active U.S. utility patents, 252 active international patents, trade secrets around cell printing, and a purpose-built copy-exact Newark manufacturing facility — the moat rests on proprietary solid-oxide technology and manufacturing know-how, not network effects or locked-in customers (Item 1, Intellectual Property / Manufacturing Facilities). | switching costs The FY2026 10-K makes the source of the advantage explicit and it is the cost of leaving the data-management layer, not a network or a patent estate. The same ONTAP software runs the on-premises arrays and the cloud services ("Our cloud storage services are based on the same ONTAP data management software that underpins our on-premises ONTAP storage infrastructure offerings"), and the AFF family "allows customers to connect to clouds for more data services, data tiering, caching, and disaster recovery". A customer's volume layout, snapshot and replication workflow and operating tools therefore carry from the array into Azure, AWS and Google rather than being abandoned at the cloud boundary — the filing describes NetApp as "the only provider of enterprise-grade storage services natively embedded in the world's largest public cloud providers", so the usual moment of escape is instead the moment the relationship renews. | 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 | clear leader Scoped to fuel-cell-based onsite power for datacenters: Energy Changemakers (2026-07-04) positions Bloom as the market leader in fuel-cell deployment for AI datacenters — Brookfield framework expanded $5B→$25B, Oracle up to 2.8 GW, Equinix 100+ MW across 19 US datacenters, AEP gigawatt-scale — with FuelCell Energy the named challenger at smaller aggregate scale (~450/380/360 MW deals). In onsite power broadly (vs. turbines and engines) Bloom is one option among several (energychangemakers.com/fuel-cells-gain-ground-in-data-centers/). | co leader IDC's 1Q26 external enterprise storage systems tracker, as reported by Blocks & Files on 2026-06-16 (cited in full on the AFF/ASA product row below), ranks NetApp second worldwide behind Dell and ahead of Everpure, Huawei and HPE, attributing the placing to "its growing all-flash business and cloud-integrated data management". Second of five ranked vendors, in a market whose leader is someone else, is a shared front rank rather than an owned one - and the distinct claim NetApp makes in the FY2026 10-K is positional rather than volumetric: being "the only provider of enterprise-grade storage services natively embedded in the world's largest public cloud providers". | 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 Q2 2026 release: GAAP gross margin 33.4%, up 668 bps YoY, with 34.3% non-GAAP and raised FY26 guidance of ~34% non-GAAP gross margin — margin expansion during a supply-constrained turbine market suggests real but cyclical pricing leverage; the FY2025 10-K frames the core competition as grid electricity and combustion OEMs on cost, which caps pricing through the cycle (sec.gov/Archives/edgar/data/1664703/000162828026050150/ex991_q226financialresults.htm). | moderate It holds price rather than raising it. On the figures filed with the FY2026 10-K, gross margin was 70.7% of revenue in FY2024 ($4,433M on $6,268M), 70.2% in FY2025 ($4,613M on $6,572M) and 70.7% in FY2026 ($4,899M on $6,925M) - flat across three years in which revenue grew from $6,268M to $6,925M - and it held that level while absorbing a component-cost shock. It is no stronger than that because the filing's own risk factor lists "competitive pricing, customer price sensitivity" and "pricing and discounting pressures" among the drivers of gross margin, and discloses that the company "experienced inflationary pressure and supply chain constraints beginning in the second half of fiscal 2026, resulting in increased costs for memory and other components, which have affected our gross margins" - a cost shock it is absorbing rather than fully passing on. | 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 | Bloom's edge is a single solid-oxide platform — 'cell printing, stack assemblies, and column configurations' — that serves both the Energy Server and the Bloom Electrolyzer, protected by 380 active U.S. patents and decades of materials know-how (FY2025 10-K, Item 1). Against its practical competition — grid power, gas reciprocating engines, small gas turbines, and other fuel-cell chemistries (PEM/MCFC/PAFC) — the 10-K claims higher efficiency, no combustion, 99.9% fleet availability on post-2020 non-redundant installs, and deployment in weeks-to-months while turbine OEMs are supply-constrained. That time-to-power window is currently being monetized in AI datacenters: the Q2 2026 release (2026-07-28) reported record revenue of $1,065.4M (+165.5% YoY) and management said all major U.S. hyperscalers have 'validated and approved' Bloom, calling it 'a standard for AI onsite power.' The moat is rated narrow, not wide: the 10-K's own risk factors stress that distributed generation is an emerging market with uncertain acceptance, three customers were 43%/13%/12% of FY2025 revenue, and industry coverage (Energy Changemakers, 2026-07-04) shows FuelCell Energy signing deals of comparable scale — the advantage is a technology-and-execution lead that rivals and turbine capacity additions can compress, not a structural lock. | NetApp sells storage hardware but the asset is ONTAP, the data-management software that has run its arrays for over three decades and now also runs inside the three largest public clouds as a first-party service. The FY2026 10-K organises the company into two segments, Hybrid Cloud (AFF and ASA all-flash arrays, AFX for AI workloads, FAS hybrid-flash, E/EF-Series, StorageGRID object storage) and Public Cloud (Azure NetApp Files, Amazon FSx for NetApp ONTAP, Google Cloud NetApp Volumes, Cloud Volumes ONTAP), and states that both rest on the same ONTAP software. That is the whole argument: an enterprise that has standardised its snapshots, replication and multiprotocol access on ONTAP carries those habits with it when it moves workloads to a hyperscaler, and NetApp is paid on both sides of the move. The evidence that the lock has commercial value is the margin's steadiness: across the three years the FY2026 10-K reports, gross margin sat at 70.7%, 70.2% and 70.7% of revenue ($4,433M on $6,268M, $4,613M on $6,572M, $4,899M on $6,925M) while revenue grew, and the filing's own percentage-of-revenue table shows no mix shift doing that work - product and services held near 46% and 54% of revenue throughout. The limits are equally in the filing. NetApp is second, not first: IDC's 1Q26 tracker puts it behind Dell in external enterprise storage, and the 10-K's competition section concedes that cloud providers are simultaneously partners and rivals, that consumption models "may reduce overall demand for our traditional on-premises offerings sold through a capital expenditure (capex) model", and that alternative architectures "may reduce or eliminate demand for some of our offerings". Component exposure is real too: the filing discloses "inflationary pressure and supply chain constraints beginning in the second half of fiscal 2026, resulting in increased costs for memory and other components, which have affected our gross margins", and names NAND among the components whose supply can tighten. This is a durable second place built on software stickiness, not a structural monopoly. | 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 | Onsite prime-power supplier to AI infrastructure: per the FY2025 10-K, AEP is procuring up to 1 GW of Bloom fuel cells for AI datacenter loads, Brookfield established a financing framework (up to $5.0B over five years at signing, per the 10-K) housed in an AI Infrastructure Fund, and SK ecoplant (a ~2.5% Class A holder) distributes and co-assembles in Korea; the Q2 2026 release adds that all major U.S. hyperscalers plus over a dozen neoclouds, AI labs, and colocation operators have approved Bloom's systems. | NetApp sits between the memory supply and the enterprise data centre. Upstream, the FY2026 10-K says "Third-party component costs make up a significant portion of our product costs" and singles out NAND as hard to manage "if supplies of certain components, including NAND, become limited relative to demand". Downstream, the hyperscalers are channel, partner and rival at once: the filing states "We both partner with and compete against cloud service providers through our cloud-based software and services offerings", while Azure NetApp Files, Amazon FSx for NetApp ONTAP and Google Cloud NetApp Volumes are delivered as those clouds' own natively embedded services. Distribution is a mix of direct sales and "an ecosystem of partners, including the leading cloud providers". | 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 | +0.20 at weight 0.20 · swarm bullish Editorial prior, not backtested. | +0.13 at weight 0.20 · swarm neutral Editorial prior, not backtested. | +0.13 at weight 0.20 · swarm bullish Editorial prior, not backtested. |