AI Industry Chain Atlas¶
Extracted 269 structured supply chain relationships (cross-validated by ≥2 articles) from 1643 SemiAnalysis + Stratechery + tier-1 Substack articles, sorted by citation frequency.
Top 15 Key Edges (mermaid Network Diagram)¶
graph LR
NVDA["NVDA"] -->|GPU foundry capacity| TSM["TSM"]
OPENAI["OPENAI"] -->|GPU compute for training | NVDA["NVDA"]
INTC["INTC"] -->|Market share in X86 serve| AMD["AMD"]
TSM["TSM"] -->|advanced chip manufacturi| AAPL["AAPL"]
AMD["AMD"] -->|market share in AI GPUs| NVDA["NVDA"]
NVDA["NVDA"] -->|AI GPU market share| AMD["AMD"]
INTC["INTC"] -->|foundry services for lead| TSM["TSM"]
META["META"] -->|user attention and ad rev| TIKTOK["TIKTOK"]
GOOGL["GOOGL"] -->|AI model capabilities and| OPENAI["OPENAI"]
AMD["AMD"] -->|market share in server CP| INTC["INTC"]
NVDA["NVDA"] -->|GPU chip fabrication (CoW| TSM["TSM"]
INTC["INTC"] -->|foundry market share| TSM["TSM"]
AMD["AMD"] -->|CPU and GPU manufacturing| TSM["TSM"]
NVDA["NVDA"] -->|HBM memory| SKH["SKH"]
AMD["AMD"] -->|AI GPU market share for t| NVDA["NVDA"]Complete Top 60 Edges Table¶
| from | → | to | what flows | Citation Count | Example Context |
|---|---|---|---|---|---|
NVDA |
→ | TSM |
GPU foundry capacity | 90 | Nvidia secures capacity at TSMC for its GPU production, which is critical for me |
OPENAI |
→ | NVDA |
GPU compute for training models | 9 | OpenAI uses Nvidia GPUs but developed Triton to bypass CUDA, weakening Nvidia's |
INTC |
→ | AMD |
Market share in X86 server CPUs | 8 | Intel fell behind AMD in server CPU designs due to foundry process issues. |
TSM |
→ | AAPL |
advanced chip manufacturing capacity | 8 | TSMC's massive capital expenditure paved the way for Apple as a key customer. |
AMD |
→ | NVDA |
market share in AI GPUs | 7 | AMD competes with NVDA in the AI GPU market, but NVDA has a higher valuation mul |
NVDA |
→ | AMD |
AI GPU market share | 7 | AMD's MI300X competes with Nvidia's Blackwell in performance/TCO for AI workload |
INTC |
→ | TSM |
foundry services for leading-edge logic | 7 | Intel competes with TSMC in the foundry market, aiming to regain process leaders |
META |
→ | TIKTOK |
user attention and ad revenue | 7 | Meta competes with TikTok for short-form video engagement, adapting its algorith |
GOOGL |
→ | OPENAI |
AI model capabilities and market leadership | 7 | Google and OpenAI are direct competitors in generative AI, with Google's Gemini |
AMD |
→ | INTC |
market share in server CPUs and client CPUs | 6 | AMD is gaining CPU market share from Intel, especially in enterprise and data ce |
NVDA |
→ | TSM |
GPU chip fabrication (CoWoS packaging) | 6 | Nvidia relies on TSMC for advanced packaging like CoWoS-S, which is a bottleneck |
INTC |
→ | TSM |
foundry market share | 6 | Intel CEO Pat Gelsinger predicts Moore's Law will accelerate, positioning Intel |
AMD |
→ | TSM |
CPU and GPU manufacturing | 5 | AMD relies on TSMC for advanced chip fabrication, including EPYC CPUs and Instin |
NVDA |
→ | SKH |
HBM memory | 5 | NVIDIA uses HBM memory from SK Hynix in its GPUs for high-bandwidth memory in AI |
AMD |
→ | NVDA |
AI GPU market share for training and inference | 5 | AMD's Instinct MI350 and MI450 compete with NVIDIA's GPUs for AI workloads. |
TSM |
→ | NVDA |
advanced chip manufacturing (e.g., for CuLitho) | 5 | TSMC CEO CC Wei gave a quote supporting Nvidia's CuLitho technology. |
NVDA |
→ | AMD |
market share in datacenter GPUs | 5 | Nvidia and AMD compete in the datacenter GPU market for AI training and inferenc |
OPENAI |
→ | MSFT |
AI model access, cloud compute | 5 | Microsoft provides Azure cloud infrastructure for OpenAI's training and inferenc |
NFLX |
→ | ADVERTISERS |
ad inventory | 5 | Netflix's advertising business has turned the corner, offering ad slots. |
INTC |
→ | TSM |
chip manufacturing services for fastest chips | 5 | Intel gives up margin to TSMC to make its fastest chips while ramping its own fa |
AAPL |
→ | APP DEVELOPERS |
App Store access, iOS user base, IAP system | 5 | Apple provides distribution and payment infrastructure to developers in exchange |
META |
→ | NVDA |
GPUs for AI training and inference | 5 | Meta is a major buyer of Nvidia GPUs to power its AI infrastructure. |
META |
→ | AAPL |
advertising revenue and user data access | 5 | Apple's ATT policy restricts Facebook's ability to track conversions, impairing |
AAPL |
→ | APP DEVELOPERS |
App Store distribution, payment processing, user trust | 5 | Apple provides the App Store platform and payment system, taking 30% of transact |
NVDA |
→ | TSM |
GPU manufacturing capacity | 5 | Nvidia relies on TSMC for advanced chip fabrication, as highlighted in the conte |
INTC |
→ | TSM |
foundry services | 4 | Intel is transitioning to smaller nodes to compete with TSMC, causing internal C |
NVDA |
→ | TSM |
chip manufacturing services | 4 | NVIDIA relies on TSMC for manufacturing its graphics chips, as highlighted by a |
GOOGL |
→ | NVDA |
AI accelerator market share | 4 | Google's TPU 8 competes with Nvidia's Blackwell GPUs for GenAI training and infe |
ASML |
→ | TSM |
EUV and DUV lithography machines | 4 | TSMC is a key customer of ASML's lithography machines, which are essential for a |
OPENAI |
→ | MSFT |
AI model access and compute resources | 4 | OpenAI partners with Microsoft for Azure compute and distribution of ChatGPT. |
NVDA |
→ | TSM |
GPU manufacturing services | 4 | Nvidia relies on TSMC for advanced chip fabrication, as indicated by Trump's sto |
AMD |
→ | TSM |
chip manufacturing services | 4 | AMD also uses TSMC for production, as Trump bought AMD stock before favorable po |
NVDA |
→ | GOOGL |
AI accelerator market share | 4 | Google's TPU rack-scale systems compete with Nvidia's Oberon architecture. |
NVDA |
→ | AMD |
GPU market share for AI and HPC | 4 | AMD competes with NVIDIA in the GPU market, though NVIDIA dominates AI accelerat |
TSM |
→ | INTC |
foundry market share | 4 | TSMC surpassed Intel in revenue in 2022, competing in leading-edge manufacturing |
TSM |
→ | ASML |
EUV lithography equipment | 4 | TSMC maintains EUV orders despite delays in other equipment. |
META |
→ | ESLOY |
smart glasses design and distribution | 4 | Meta partners with EssilorLuxottica for smart glasses, utilizing their optical m |
TIKTOK |
→ | META |
user attention and creator talent | 4 | TikTok is a major competitor to Instagram, especially in breaking new talent and |
AAPL |
→ | OPENAI |
ChatGPT integration for image generation and analysis | 4 | Apple relies on OpenAI's ChatGPT for AI features like Image Playground and Visua |
AAPL |
→ | EPIC |
legal dispute over App Store policies | 4 | Epic Games sued Apple over anti-steering provisions and mandatory IAP, leading t |
OPENAI |
→ | NVDA |
GPU compute capacity | 4 | OpenAI relies on Nvidia GPUs for training and inference, and GPU shortages const |
AAPL |
→ | META |
advertising revenue and user data access | 4 | Apple's ATT restricts third-party tracking, benefiting its own Search Ads at the |
META |
→ | ADVERTISERS |
ad impressions and ad revenue | 4 | Meta's advertising business generates revenue by showing ads to users, with AI i |
META |
→ | ADVERTISERS |
ad inventory and targeting | 4 | Meta provides ad space and targeting to advertisers, generating $27 billion in q |
AAPL |
→ | DEVELOPERS |
App Store access, limited APIs | 4 | Apple's iOS is a thick platform but restricts developer freedom via App Store co |
MSFT |
→ | OPENAI |
AI model integration (ChatGPT) into Bing | 4 | Microsoft partners with OpenAI to incorporate generative AI into Bing search. |
MSFT |
→ | OPENAI |
AI models and technology | 4 | Microsoft partners with OpenAI to integrate AI into its products and Azure cloud |
AMD |
→ | INTC |
x86 server CPU market share | 3 | AMD competes with Intel in the x86 server CPU market, and AMD's growth comes par |
AMAT |
→ | ASML |
market share in lithography equipment | 3 | ASML competes with Applied Materials in advanced lithography systems. |
INTC |
→ | AMD |
CPU market share | 3 | Intel competes with AMD in the x86 CPU market. |
AMZN |
→ | NVDA |
GPU chips | 3 | Amazon uses Nvidia GPUs for AWS, but is reducing dependence via custom chips. |
AMD |
→ | NVDA |
AI accelerator market share | 3 | AMD competes with Nvidia in the data center GPU market, with AMD's Instinct GPUs |
AAPL |
→ | MEMORY SUPPLIERS |
memory chips | 3 | Apple faces rising memory costs and supply constraints that impact its gross mar |
CEREBRAS |
→ | TSM |
wafer-scale chip manufacturing (N3 process for WSE-4) | 3 | Cerebras relies on TSMC for advanced node fabrication of its wafer-scale engines |
AMD |
→ | TSM |
chip design for fabrication | 3 | AMD uses TSMC for manufacturing, competing with Intel in the CPU market. |
ASML |
→ | TSM |
EUV lithography machines | 3 | ASML supplies EUV machines to TSMC for advanced chip manufacturing. |
MSFT |
→ | NVDA |
GPU hardware for AI training and inference | 3 | Microsoft uses Nvidia GPUs in its Azure cloud for AI workloads, and Nvidia also |
NVDA |
→ | TSM |
GPU designs for fabrication | 3 | Nvidia relies on TSMC to manufacture its GPU chips. |
AMD |
→ | TSM |
GPU foundry capacity for MI350P chips | 3 | AMD relies on TSMC to manufacture its Instinct MI350P GPUs. |
OPENAI |
→ | AAPL |
ChatGPT integration into iOS (Siri, Visual Intelligence) | 3 | OpenAI integrated ChatGPT into Apple's operating systems expecting subscriber gr |
How to Read This Chart¶
- Edge frequency = number of independent articles mentioning this relationship. Higher is more robust.
- What flows = specific flow content extracted by LLM (e.g., 'GPU foundry capacity', 'HBM memory').
- Use the
corpus-search '<from_ticker> <to_ticker>'command to pull original article text.