Zhipu AI (Z.ai) — AI supply-chain exposure
The model reads Zhipu AI (Z.ai) primarily as a producer in Models. Its most binding exposure is Inference serving (system bottleneck #4), which it consumes rather than makes — a price-taking dependency. Its largest modeled sensitivity is a shock at GPU (constraint β 25).
The structural read · model-generated
The model reads Zhipu AI (Z.ai) primarily as a producer in Models. Its most binding exposure is Inference serving (system bottleneck #4), which it consumes rather than makes — a price-taking dependency. Its largest modeled sensitivity is a shock at GPU (constraint β 25).
In depth · editorial + model · written 2026-07-13
Zhipu AI, spun out of Tsinghua University, is one of China's frontier model labs and the maker of the GLM family. It sits squarely on the model layer of the chain: it pretrains its own foundation models, builds reasoning and multimodal variants on top, and serves them for inference. Unlike closed Western labs, it leans on open-weights releases, which turn its models into a reference stack that domestic developers and downstream apps can build against directly.
Its real structural hook is where it is forced to run. Placed on the US Entity List, Zhipu is cut off from the newest Western accelerators, which pushes it toward domestic silicon and positions it inside China's sovereign-compute effort. That constraint is also its centrality: it is a producer whose open models and home-chip adaptation make it a load-bearing node for the non-US model ecosystem, which is why the model weights it heavily despite the export walls around it.
Where it has leverage
Chain footprint by layer
How it participates
Geographic concentration
Frequently asked
What is Zhipu AI (Z.ai)'s role in the AI supply chain?
The model reads Zhipu AI (Z.ai) primarily as a producer in Models. Its most binding exposure is Inference serving (system bottleneck #4), which it consumes rather than makes — a price-taking dependency. Its largest modeled sensitivity is a shock at GPU (constraint β 25).
Which parts of the AI value chain is Zhipu AI (Z.ai) exposed to?
Zhipu AI (Z.ai) is mapped to 5 parts of the AI value chain, most strongly Pretraining, Reasoning models, Inference serving. It sits primarily in the Models layer as a producer.
Does Zhipu AI (Z.ai) own an AI bottleneck?
Not in the current model — Zhipu AI (Z.ai) is exposed to constrained parts but sits downstream of them rather than producing them.
What is Zhipu AI (Z.ai)'s biggest AI supply-chain risk?
Its largest modeled sensitivity is a shock at GPU (constraint β 25). 4 nodes depend on it; pressure 78/100
Who are Zhipu AI (Z.ai)'s closest peers by AI-chain position?
By shared chain dependencies: MiniMax, StepFun, Moonshot AI, xAI.
Go live on Zhipu AI (Z.ai)
- The interactive dependency graph and full company Nexus
- The analyst bull / bear thesis and valuation lens
- Live signals, today’s movers and the read-through
- Track it in your Portfolio Cockpit — positions, P&L, valuation, thesis
model v0.7.0 · research, not advice
Chain analytics are illustrative, order-of-magnitude estimates from our model of the AI value chain — not investment advice.