Samsung: Bigger, Ahead on Layers, and Owns Its FabsThin moat

Sandisk (SNDK) — moat facet

Samsung is two and a half times the size, ahead on layers, owns its fabs, and sits inside a conglomerate that can lose money for as long as it likes.

Samsung is two and a half times Sandisk's size in NAND, ahead on layers, owns its fabs, and does not need the business to make money.

Samsung against Sandisk32.3%NAND share Samsung12.4%NAND share Sandisk400Layers Samsung (x10)332Layers BiCS10 (x10)Layer counts at one-tenth scale to sit alongside share. Samsung owns its fabs and funds its roadmap.
Bigger, ahead on layers, vertically integrated, and inside a conglomerate that can lose money for as long as it chooses.

The share table: Samsung 32.3%, Sandisk 12.4%.1 The technology: Samsung's tenth-generation V-NAND at 400 layers on a 1 Tbit die against BiCS10's 332.2 The structure: Samsung designs, manufactures and packages in its own plants, funds its roadmap alone, and sits inside a conglomerate whose semiconductor division is one part of a much larger enterprise.

Each of those is a durable advantage and the third is the one that matters most in a downturn. A pure-play NAND producer facing a price collapse must cut utilisation, take write-downs and hope — which is exactly what Sandisk did in 2023, 2024 and 2025 while losing money in all three. A memory division inside a conglomerate can price below cash cost for as long as its parent tolerates, and its parent has historically tolerated a great deal. This is the same argument Micron's pages make about Samsung and it applies with more force to a smaller competitor with no fab.

Sandisk's counterweights are real but narrow. It has the consumer brand Samsung competes with rather than dominates, an enterprise SSD franchise built over years of qualification, and a cost position that is genuinely competitive because it pays half of a shared development and capital bill rather than all of its own.

What it does not have is the ability to outlast anybody.

Watch the layer gap at each node. At 400 against 332 it is real and not yet decisive; if it persists across two more generations, it becomes cost per bit and there is no answer to that.

Moat trajectory: Narrowing

Samsung is at 400 layers against BiCS10's 332 and holds 32.3% of NAND against Sandisk's 12.4%. Neither gap has closed, and the new Japanese capacity Sandisk is funding half of will not arrive until the back half of the decade.

The number that tests this moat
Reported
Net income, latest quarter
$6,903M in Q4 fiscal 2026, against a $23M loss a year earlier

Being behind on layers has not cost Sandisk profit in a shortage; the test is the next price decline, when the lower-cost producer keeps earning.

Source: Sandisk fiscal Q4 2026 results release (Exhibit 99.1, 5 August 2026) ↗
References
  1. ReportedThe share table: Samsung 32.3%, Sandisk 12.4%. The technology: Samsung's tenth-generation V-NAND at 400 layers on a 1 Tbit die against BiCS10's 332.
    Coverage of Kioxia and Sandisk's BiCS10 3D NAND. BiCS10 is a 332-layer technology; the interface speed of 4.8 Gbit/s is 33 percent faster than BiCS8, achieved using the Toggle DDR6.0 interface and a Separate Command Address protocol. Mass BiCS10 production is expected to start in 2027 at the Kitakami plant 2 in Iwate Prefecture, Japan, and because Sandisk and Kioxia share the output of the fab through their joint venture both are sample shipping BiCS10 chips; a QLC version of the 332-layer technology would increase chip capacity by a third. The BiCS10 technology involves stacking three 100-plus layer NAND strings together rather than building a monolithic 332-layer chip. Kioxia has an intervening BiCS9 technology which uses BiCS8 218-layer 3D NAND cells with a separate CMOS logic layer providing more performance than the BiCS8 logic circuitry. For comparison, SK hynix has 321 layers in its ninth-generation 3D NAND, also using a triple string stack design; Samsung has 400 layers with its tenth-generation V-NAND, made with separate logic and NAND cell wafers producing a 1 Tbit die in what Samsung calls a Cell-on-Periphery architecture; Micron is at the 276-layer level; and China's YMTC is expected to announce 300-layer class technology. — 2026 · publ. 2026-07-03 · source ↗
  2. ReportedThe technology: Samsung's tenth-generation V-NAND at 400 layers on a 1 Tbit die against BiCS10's 332. The structure: Samsung designs, manufactures and packages in its own plants, funds its roadmap alone, and sits inside a conglomerate whose semiconductor division is one part of a much larger enterprise.
    Coverage of Kioxia and Sandisk's BiCS10 3D NAND. BiCS10 is a 332-layer technology; the interface speed of 4.8 Gbit/s is 33 percent faster than BiCS8, achieved using the Toggle DDR6.0 interface and a Separate Command Address protocol. Mass BiCS10 production is expected to start in 2027 at the Kitakami plant 2 in Iwate Prefecture, Japan, and because Sandisk and Kioxia share the output of the fab through their joint venture both are sample shipping BiCS10 chips; a QLC version of the 332-layer technology would increase chip capacity by a third. The BiCS10 technology involves stacking three 100-plus layer NAND strings together rather than building a monolithic 332-layer chip. Kioxia has an intervening BiCS9 technology which uses BiCS8 218-layer 3D NAND cells with a separate CMOS logic layer providing more performance than the BiCS8 logic circuitry. For comparison, SK hynix has 321 layers in its ninth-generation 3D NAND, also using a triple string stack design; Samsung has 400 layers with its tenth-generation V-NAND, made with separate logic and NAND cell wafers producing a 1 Tbit die in what Samsung calls a Cell-on-Periphery architecture; Micron is at the 276-layer level; and China's YMTC is expected to announce 300-layer class technology. — 2026 · publ. 2026-07-03 · source ↗
Sources
Generated September 23, 2026