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.
| Nebius Group | Applied Digital | CXMT | |
|---|---|---|---|
| Moat rating | narrow The FY2025 20-F positions Nebius as 'one of the few global, at scale, multi-tenant clouds purpose built for AI,' but the same filing names its central dependency plainly: 'We currently rely on Nvidia for the GPU chips we use' — a purpose-built neocloud renting a supply the hyperscalers it competes with also control, which is a real but narrow position. | none The FY2026 10-K (filed 2026-07-29) shows contracted revenue, not a demonstrated competitive edge. About 1,410 MW is leased under 15-year take-or-pay, non-cancellable base terms worth about $36.2 billion, but only about 100 MW of the roughly 1.5 GW that is contracted and either operating or under construction was operating and earning revenue at May 31, 2026, and Item 1A says "lessees may have the right to terminate applicable leases if there are significant delays in construction." Item 1A also concedes "We do not have the resources to compete with larger providers of similar products or services at this time," and the Competition section names 13 power-advantaged developers the company competes with. Signed leases give revenue visibility, but the filing does not show a durable advantage. | none CXMT's 2026 interim report (published 2026-08-29) names scale and IDM technology as core competencies, but also concedes gaps. It says that compared with the leading international manufacturers the Company 'still has a certain gap in overall scale, technology accumulation and customer resources' (与国际头部厂商相比，公司在整体规模、技术积累、客户资源等方面仍然存在一定差距). It describes its scale effect as 'gradually emerging' (逐步显现), and its gross-margin risk covers the case where the scale effect cannot materialise over the long term (公司规模效应长期无法显现). Its excess returns span a single shortage. The retained-earnings note shows an accumulated deficit of ¥36.65bn entering 2026, after ¥1.87bn of 2025 net profit attributable to the parent, and the first-half 2025 net loss attributable to shareholders was ¥2.33bn. Then first-half 2026 revenue rose 873.64%, which the report attributes to the global DRAM supply shortage, higher prices and sharply higher volumes, and the main-business gross margin reached 84.84%. |
| Moat type | cost scale The moat the filing describes is engineering: designs that 'optimize power and cooling efficiency, lower latency' plus 'a consistent track record of being one of the first-to-deploy the latest generation of NVIDIA GPU chips' — density and time-to-deploy at scale, not a switching lock. | none The 10-K claims three advantages: power-advantaged sites (it believes securing power and interconnection ahead of demand is 'the principal constraint on new HPC capacity and a core differentiator for us from many of our competitors'), a standardized 'franchise-style' design, and hyperscaler master service and master telecom service agreements 'that are difficult to obtain.' The filing does not show any of them to be durable. Its Competition section says competition 'centers on securing and developing sites with access to large-scale, reliable, and cost-competitive power and interconnection' and names 13 power-advantaged developers going after the same leases, and Item 1A concedes it lacks the resources to compete with larger providers. Signed leases are take-or-pay and non-cancellable, so a tenant leaving for convenience owes 'the full remaining contractual value,' but that is contractual lock-in on each lease rather than a moat source, so no moat type is assigned. | none On the report's own evidence, no candidate moat source qualifies. Cost scale: the report calls DRAM a highly standardised product (高度标准化的产品) in which the cost advantage of scale is a core competency. But CXMT ranks fourth globally by capacity, and the report, calculating on sales, puts Samsung, SK hynix and Micron at 33.96%, 34.48% and 23.41% of the 2025 global DRAM market, so the scale advantage lies with the leaders. IP: it reports 4,484 domestic patents (3,744 of them invention patents) and 3,400 overseas patents as of 2026-06-30. Yet it describes its core technology as reaching 'international advanced level' (国际先进水平) and flags possible IP disputes with competitors. Switching costs: ¥133.48bn of ¥150.31bn first-half 2026 revenue went through distributors, who under the report's revenue note decide their own resale prices. |
| Leadership | fast follower The 20-F's own claim is 'one of the FEW global, at scale' AI clouds — a differentiated challenger to the hyperscalers it names as the competitive field, not a claimed leader of it; the first-to-deploy record is a follower's speed advantage, not category leadership. | behind The 10-K makes no leadership claim and gives no ranking or share figure. Item 1A concedes "We do not have the resources to compete with larger providers of similar products or services at this time" and that some rivals have "substantially greater liquidity and financial resources than we do." Its Competition section places APLD against established operators (Digital Realty, Equinix), hyperscalers that build their own capacity, independent developers and 13 named power-advantaged developers (IREN, Cipher Digital, TeraWulf, Hut 8, Riot, CleanSpark, HIVE, Core Scientific, Bitdeer, Galaxy Digital, Fermi, Keel Infrastructure, MARA). | fast follower TrendForce (2026-09-24) reports that CXMT's global DRAM revenue share rose to 9.5% in 2Q26 from 7.6% in 1Q26, 'placing it fourth behind Samsung at 39.4%, SK hynix at 24.9%, and Micron at 23.3%'. It is catching up through generation-skipping R&D. The interim report had its fifth process platform in customer certification, and a voluntary disclosure on 2026-09-21 announced that platform's mass production, adding that its products are not yet in scale sales and yields need time to ramp. TrendForce's listing analysis (2026-07-28) still says CXMT 'still trails Samsung, SK hynix, and Micron in advanced processes, product performance, yields, and certification by high-end customers'. |
| Pricing power | weak The filing frames a 'highly competitive' market with 'frequent introduction of new or improved solutions' and a GPU cost base set by a single supplier (NVIDIA) — a renter of compute competing on efficiency has little list-price control. | weak Item 1A says "Due to the limited number of hyperscalers, we expect that a limited number of customers will continue to account for a high percentage of our revenue for the foreseeable future," and that if customers' equipment usage declines or they discontinue use of its facilities, APLD "may be compelled to lower our lease prices in some instances or risk losing a significant customer." One customer was 59% of FY2026 revenue from continuing operations. Take-or-pay, non-cancellable terms protect contracted revenue over the base term, and Note 19 reports a $39.1M HPC Hosting segment profit on $385.3M of segment revenue in FY2026, but those terms are agreed with a small group of concentrated buyers. | weak CXMT is a price taker riding the cycle. The interim report attributes first-half revenue growth to the global DRAM supply shortage, rising prices and sharply higher volumes. It records industry prices between 2015 and 2025 as high as $7.89/GB and as low as $1.78/GB in first-half 2023, and it states that the continued steep price rise is not sustainable. Main-business revenue was ¥15.22bn against cost of ¥13.29bn in first-half 2025, and ¥150.04bn against ¥22.75bn in first-half 2026. Most sales go through distributors, who set their own resale prices. |
| Summary | Nebius sells a 'unified full-stack AI cloud that spans the complete AI journey – from compute capacity to software and services,' with hardware and software built in-house — the neocloud pitch of hyperscaler reliability at purpose-built efficiency, and a first-to-deploy record on new NVIDIA silicon. Two things bound the moat, both from the filing: it 'currently rel[ies] on Nvidia for the GPU chips,' the same constraint every neocloud shares, and it is a Nasdaq 'Controlled Company' whose founding shareholder holds concentrated voting power. The three non-core segments (Toloka, Avride, TripleTen) are separate businesses, not the cloud moat. | Applied Digital designs, builds and operates purpose-built, liquid-cooled HPC data centers, which it calls 'AI factories', and leases the capacity to CoreWeave and investment-grade hyperscalers. At May 31, 2026 its 10-K lists five campuses (Polaris Forge 1-3 and Delta Forge 1-2) with about 1,410 MW contracted under roughly 15-year take-or-pay, non-cancellable leases worth about $36.2 billion over the base terms. The filing claims three sources of advantage: it controls power-advantaged sites, it uses a standardized 'franchise-style' design built to deliver about 150 MW in about 14 to 18 months, and it holds hyperscaler master agreements that are 'difficult to obtain.' The same document shows how early the company is. About 100 MW was operating and earning revenue. One customer was 59% of FY2026 revenue from continuing operations. It competes with Digital Realty, Equinix, hyperscalers that build their own capacity and 13 named power-advantaged developers, and it concedes that it lacks the resources to compete with larger providers. Signed leases give long-dated revenue visibility, but the filing does not show a durable competitive advantage. | CXMT (ChangXin) is a Hefei-based IDM that designs and fabricates DDR5 and LPDDR4X/LPDDR5/5X/LPDDR6 memory and sells DRAM wafers, chips and modules. Its 2026 interim report says it reached its fourth process platform through a 'generation-skipping' (跳代研发) R&D strategy, and that it ranks first in China and fourth globally by shipments and sales. First-half 2026 revenue was ¥150.31bn (DDR series ¥69.47bn, LPDDR series ¥78.19bn), with a main-business gross margin of 84.84%. TrendForce puts its 2Q26 DRAM revenue share at 9.5%, up from 7.6% in 1Q26. The position is real but recent. The report concedes gaps to the three leaders in scale, technology and customers. The customers it names are Alibaba Cloud, ByteDance, Tencent, Lenovo, Xiaomi, Transsion, Honor, OPPO and vivo. On 2026-06-08 the US Defense Department added its subsidiary ChangXin Memory to the Section 1260H list. TrendForce notes that export controls leave CXMT relying on DUV multiple patterning rather than EUV, and that its prospectus 'has not disclosed a clear mass-production timetable' for HBM. The report itself warns that the continued steep rise in DRAM prices is not sustainable (价格的持续大幅上涨不具备可持续性) as international makers add capacity. |
| Chain position | Layer-8 neocloud — a purpose-built AI compute provider reselling NVIDIA silicon at scale. | Developer and landlord of power-advantaged, liquid-cooled AI data-center capacity, leased long-term to CoreWeave and investment-grade hyperscalers. | Upstream DRAM IDM selling wafers, chips and modules to server, smartphone, PC and automotive makers, mostly through distributors (¥133.48bn of ¥150.31bn first-half 2026 revenue). |
| Products (share / barrier) |
|
|
|
| Long-horizon vote | -0.01 at weight 0.20 · swarm neutral Editorial prior, not backtested. | -0.20 at weight 0.20 · swarm neutral Editorial prior, not backtested. | -0.06 at weight 0.20 · swarm bullish Editorial prior, not backtested. |