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.
| Samsung Electronics | Eli Lilly and Company | Cisco Systems | |
|---|---|---|---|
| Moat rating | narrow Samsung leads three of its four principal markets — Q1 2026 DRAM revenue share 38.5% ($37.32B, TrendForce, https://www.trendforce.com/presscenter/news/20260601-13070.html), Q1 2026 NAND share 31.6% ($13.51B, TrendForce, https://www.trendforce.com/presscenter/news/20260525-13058.html), and #1 in Q2 2026 smartphones at 24% (Counterpoint via Android Authority, https://www.androidauthority.com/counterpoint-research-q2-2026-smartphone-shipment-report-3686931/) — but the moat is narrow, not wide: memory economics are violently cyclical (industry DRAM revenue swung +81% QoQ on contract-price spikes per the same TrendForce release), SK hynix holds the highest HBM bit-shipment mix among the top three (same release), and foundry trails TSMC 6.5% vs ~72% share (TrendForce data via TelecomLead, https://telecomlead.com/semiconductor/global-foundry-market-hits-record-47-95-bn-in-q1-2026-as-ai-chip-demand-drives-growth-tsmc-expands-share-to-72-126247). | wide Eli Lilly's FY2025 10-K states "Intellectual property protection is critical" and that compound patents provide roughly 20-year market exclusivity ("patents in each relevant country last for a period of 20 years"), warning that loss of patent protection causes a "severe and rapid decline in revenues." Backed by an 83.0% gross margin (up from 81.3% in 2024) and a fast-growing, patent-protected incretin franchise (Mounjaro +99%, Zepbound +175% in 2025), the economic moat is wide and durable. Source: LLY FY2025 10-K. | narrow The FY2025 10-K shows a moat that still pays but no longer compounds: total gross margin of 64.9% and remaining performance obligations of $43,533 million (up 6%) against a Networking product category that fell from $34,570 million in fiscal 2023 to $29,229 million in fiscal 2024 to $28,304 million in fiscal 2025, and a product gross-margin bridge in which 'Product pricing' subtracted 1.6 percentage points. All of the reported category growth came from Splunk-carrying lines - Security +59% and Observability +26%. Real and durable, but bounded: narrow, not wide. |
| Moat type | cost scale The durable advantage is manufacturing scale and capital intensity: the top three DRAM suppliers take roughly $8.97 of every $10 of industry revenue (Samsung 38.5% + SK hynix 28.8% + Micron 22.4% in Q1 2026 per TrendForce, https://www.trendforce.com/presscenter/news/20260601-13070.html), a structure sustained by multi-billion-dollar fab economics that has admitted no new large entrant in decades; Samsung is the largest-revenue producer in both DRAM and NAND (TrendForce, https://www.trendforce.com/presscenter/news/20260525-13058.html). | intangibles ip Per the LLY FY2025 10-K, the pharma protection "anticipated to be of most relevance" is "patents claiming the active ingredient (the compound patent)," supplemented by regulatory data protection, FDA approvals, and brands — i.e., intangible assets. The 10-K notes that without such protection, products face intense generic and biosimilar price competition that erodes revenue quickly. | switching costs The durable piece is the contracted annuity attached to installed equipment: total remaining performance obligations of $43,533 million, of which $21,961 million is services and $21,810 million is long-term (recognised beyond twelve months), carried at a services gross margin of 68.5% that has risen for three consecutive years from 66.4% and 68.1%. Customers keep paying for support on gear they already run. |
| Leadership | co leader Revenue leader in DRAM (38.5%) and NAND (31.6%) in Q1 2026 (TrendForce) and #1 in Q2 2026 smartphone shipments (Counterpoint), but not the leader where AI value concentrates: SK hynix has the highest HBM bit-shipment mix among the top three DRAM suppliers (TrendForce, https://www.trendforce.com/presscenter/news/20260601-13070.html) and TSMC dominates foundry with ~72% share to Samsung's 6.5% (TrendForce via TelecomLead). | co leader Duopoly with Novo Nordisk: Novo's own FY2025 20-F names "Novo Nordisk and Eli Lilly [as] the most significant companies measured by market share" in the global diabetes and GLP-1 markets. Lilly is ascendant — the 20-F concedes Wegovy's years-long US anti-obesity prescription lead "was overtaken by a competing GLP-1 based product in the beginning of 2025" (Lilly's tirzepatide) — but Novo still cites Ozempic as "the leading product by sales," so co-leader is the honest band. Sources: Novo Nordisk FY2025 20-F; LLY FY2025 10-K. | co leader The 10-K names no competitor and discloses no market share anywhere in Management's Discussion and Analysis. The positional facts it does give are scale and direction: Networking product revenue of $28,304 million in fiscal 2025 - by far the largest of its four categories - but down $6.3 billion from fiscal 2023, with channel partner financing volume falling from $32.1 billion to $27.1 billion to $24.9 billion across fiscal 2023-2025. |
| Pricing power | moderate Current pricing is extraordinary but cyclical, not structural: conventional DRAM contract prices rose roughly 93-98% QoQ in Q1 2026 (TrendForce, https://www.trendforce.com/presscenter/news/20260601-13070.html) — shortage-driven commodity pricing that reverses in downcycles. The set-side shows the limit: Samsung's own MX (smartphone) division posted a KRW 0.7 trillion operating loss in Q2 2026 on elevated component costs (Samsung Newsroom, https://news.samsung.com/global/samsung-electronics-announces-second-quarter-2026-results). | moderate Mixed. The LLY 10-K shows an 83.0% gross margin (up 1.7pp), evidence of strong list-price power on exclusive branded drugs; but it repeatedly notes 2025 revenue growth was "partially offset by lower realized prices" for both Mounjaro and Zepbound, alongside heavy PBM/formulary rebate pressure and default anti-obesity coverage restrictions (e.g., US self-insured employers "must generally opt in" for obesity-drug coverage). Net pricing power is therefore moderate, not strong. | moderate The product gross-margin bridge attributes fiscal 2025's 0.2-point improvement to productivity (+2.0 points) and mix (+1.1 points) while 'Product pricing' subtracted 1.6 points; a legal dispute with a supplier cost a further 0.8 points. Total gross margin of 64.9% (product 63.7%, services 68.5%) is held by cost and mix, not by price. |
| Summary | Samsung Electronics is the world's largest memory maker and a vertically integrated device company. In Q1 2026 it was #1 in DRAM (38.5% revenue share) and #1 in NAND (31.6%), and in Q2 2026 it retook #1 in global smartphones at 24% share. The AI-datacenter memory shortage produced record results: Q2 2026 company revenue of KRW 171.5 trillion and an all-time-high operating profit of KRW 89.5 trillion, with server memory reaching a record share of the sales mix and HBM4E samples shipped to major customers (Samsung Newsroom). The caveats that keep the moat narrow: the profit pool is a price cycle, not a structural annuity; SK hynix leads the highest-value HBM segment by bit-shipment mix; and Samsung Foundry, at 6.5% share versus TSMC's ~72%, remains a distant second in leading-edge logic despite expanding 2nm design wins. | Lilly's moat rests on patent-protected, FDA-approved branded medicines led by its tirzepatide incretin franchise: per the FY2025 10-K, Mounjaro ($22.97B) and Zepbound ($13.54B) together generated roughly 56% of 2025's $65.2B revenue (up 99% and 175% respectively), while compound patents confer ~20-year market exclusivity and support an 83.0% gross margin. Durable competition only emerges after patents and regulatory data protection lapse, at which point the 10-K warns generics and biosimilars drive a "severe and rapid decline in revenues." | The installed base still switches slowly - but the filing does not claim it is winning the AI datacenter on the strength of that base. The fiscal 2025 Networking decline is attributed to 'product shipments returning to normalized levels during the first half of fiscal 2025 from the elevated levels ... in the first half of fiscal 2024', to servers, and to 'a decline in campus switching' - normalisation, not disclosed share loss, so the whitebox/Arista thesis is not something this 10-K either confirms or denies. What the filing does disclose is the shape of the AI dollar Cisco is actually winning: Americas service provider and cloud growth 'driven by AI infrastructure revenue from webscale customers', funded by additional purchase commitments 'related to manufacturing Cisco Silicon One and other products to meet demand from webscale and other customers', which 'significantly increased our supply chain exposure, which has resulted in negative impacts to our product gross margin in recent periods and may result in further negative impacts in future periods'. Enterprise switching earns 63.7% product gross margin; webscale silicon costs margin to serve. The moat is intact where the annuity is and thin where the growth is. |
| Chain position | Upstream supplier of DRAM/HBM and NAND to the AI datacenter buildout (server memory at a record share of its Q2 2026 sales mix; HBM4E samples shipped to major customers) and a contract foundry, while simultaneously a downstream device maker (Galaxy) that consumes its own components (Samsung Newsroom, https://news.samsung.com/global/samsung-electronics-announces-second-quarter-2026-results). | Upstream branded-drug originator. Per the FY2025 10-K, US distribution flows through three wholesalers (McKesson, Cencora, Cardinal Health), each accounting for over 10% of consolidated revenue, into consolidated PBMs/payers whose formularies and rebate demands constrain net price; Lilly is also expanding a direct-to-patient channel, LillyDirect. | Supplies networking into AI datacenters: the 10-K ties Americas service provider and cloud growth to 'AI infrastructure revenue from webscale customers' and to Cisco Silicon One manufacturing commitments. |
| Products (share / barrier) |
|
|
|
| Long-horizon vote | +0.13 at weight 0.20 · swarm bullish Editorial prior, not backtested. | +0.30 at weight 0.20 · swarm neutral Editorial prior, not backtested. | +0.13 at weight 0.20 · swarm neutral Editorial prior, not backtested. |