NVIDIA
Public (NVDA) · $4.9T mkt cap · Founded 1993
Supply Chain Control Room
A more visual map of how NVIDIA depends on infrastructure, partners, and operational chokepoints.
The stack below turns tier data into an exposure view: where dependencies cluster, which layer carries the most vendors, and how many live risk signals are already mapped.
Mapped tiers
6
Critical Elements
Node clusters
15
24 listed dependencies
Risk signals
18
6 formal assessments
Highest severity
critical
12 flagged nodes
Exposure Mix
Operational risk by signal density
Dominant concentration
Critical Elements carries the widest cluster with 4 mapped nodes.
Dependency Lanes
Each tier is rendered as a lane, with grouped nodes and dependency payloads surfaced directly inside the mapped layer.
Tier 01
THE SILICON KING
Products & Platforms
NVIDIA designs the chips and software that power 80%+ of AI training worldwide.
Node clusters
3
Dependencies
8
GPU Architectures
Items
3
- Blackwell (B200/B300) — shipping now; 71% of shipments
- Vera Rubin (R200) — H2 2026; 10x perf/watt
- Rubin Ultra (VR200) — 2027; 500B transistors
Software Platform
Items
3
- CUDA — 4M+ developers locked inMOAT
- NIMs (Inference Microservices) — recurring software revNEW
- DGX Cloud — AI-as-a-serviceGROWING
Revenue Machine
Items
2
- $1T+ Blackwell + Rubin orders through 2027PIPELINE
- $68.1B single quarter revenue (Q4 FY26)RECORD
Tier 02
THE ACHILLES HEEL
Manufacturing
NVIDIA designs chips but fabricates NOTHING itself. 100% outsourced.
Node clusters
2
Dependencies
3
TSMC — Sole Foundry
Items
2
- CoWoS advanced packaging is bottleneckCONSTRAINED
- No Samsung/Intel backup for leading-edgeSINGLE SOURCE
ASML Dependency
Items
1
- Indirect but absolute dependency
Tier 03
THE BOTTLENECK
Memory & Packaging
HBM4 memory and advanced packaging are the #1 supply constraint in 2026.
Node clusters
2
Dependencies
4
HBM4 Memory
Items
3
- SK Hynix 60% — primary Blackwell/Rubin supplier
- Samsung — HBM4 production ramping
- Micron 24% — third source
CoWoS Packaging (TSMC)
Items
1
- Chip-on-Wafer-on-Substrate — only TSMC does at scaleMONOPOLY
Tier 04
THE LOCK-IN LAYER
Networking & Systems
Proprietary interconnects create switching costs competitors can't match.
Node clusters
2
Dependencies
3
NVLink & InfiniBand
Items
2
- NVLink 6 coming with Rubin — proprietary GPU-to-GPUPROPRIETARY
- InfiniBand (Mellanox acquisition 2020) — cluster networkingACQUIRED
Spectrum-X Ethernet
Items
1
- Competing with Broadcom for open network market
Tier 05
SHARED CHOKEPOINT
Foundry & Materials
Same TSMC/ASML dependency as every AI company — but NVIDIA IS the demand.
Node clusters
2
Dependencies
2
TSMC
Items
1
- All leading-edge GPUs fabbed hereMONOPOLY
ASML
Items
1
- High-NA EUV — sole supplierNO ALT
Tier 06
THE INGREDIENTS
Critical Elements
The Silicon Curtain of 2026.
Node clusters
4
Dependencies
4
Gallium & Germanium
Items
1
- 25% revenue tax on chip salesGEOPOLITICAL
Neon Gas
Items
1
- ~50% of semiconductor-grade neonWAR ZONE
Copper
Items
1
- Record prices hit DC costsCOST
Silicon Wafers
Items
1
- Shin-Etsu & SUMCODUOPOLY
Risk Assessments
Formal assessments are separated from node-level flags so the page reads like a dashboard first and a report second.
Severity Stack
Immediate read
The mapped risk posture currently peaks at critical severity, with concentration strongest in Critical Elements.
TSMC Single-Foundry Dependency
100% of advanced GPUs fabbed at TSMC. Taiwan Strait blockade = NVIDIA stops shipping. No backup foundry for 3nm/2nm.
China Export Restrictions
US export controls limit China sales. Hopper series share dropping from 10% to 7%. Geopolitical wildcard.
Custom Silicon Competition
Google TPU, Amazon Trainium, Broadcom/OpenAI co-designed chip, AMD MI300X. Customers actively seeking alternatives.
Gross Margin Pressure
75% gross margins are extraordinary but face pressure from rising input costs (HBM4, CoWoS packaging) and customer pushback.
Rubin Delay Risk
HBM4 validation, CX9 networking transition, and power consumption challenges could delay Rubin shipments. TrendForce cut share from 29% to 22%.
Circular Investment Scrutiny
NVIDIA invests $30B in OpenAI → OpenAI buys $30B+ in GPUs. Regulators may view as anti-competitive tying.