AlphaVectorBETA
WatchlistTrendingAI ChainSymbolsBlogVoicesPricing
US Markets Closed
δΈ­ζ–‡
Sign inSign up
AI industry chain

Map the AI value chain from power to applications

A practical research page for separating durable AI infrastructure demand from cyclical inventory, pricing pressure, and application monetization risk.

Framework only, not financial advice.

8
chain layers
32
sample tickers
5
signal groups

Drilldown

Where AI infrastructure demand flows

Click any source or destination to drill into the estimate. Public hyperscaler capex, cloud commitments, and private AI lab cluster estimates are modeled separately.

Scope overview

One pool, four source types

This is the top-level demand number before it is routed through operators and suppliers. The cards below show what is inside L1, and what must not be counted again.

Single-counted demand pool
L1. AI CAPEX demand pool
L1 AI CAPEX demand pool
$795-1047B
midpoint model
$921B
low-high spread
$252B

Method: use a single AI CAPEX demand pool built from company-specific 2026 AI/cloud infrastructure estimates, cloud compute commitments, and directional cluster estimates. L2 to L4 are routing and attribution layers; they do not add new spend on top of L1.

Demand mix
100%
Estimate stack
18 source rows
public guidance
6
commitments / leases
3
directional estimates
9
Do not add L2-L4 on top

L2 is the order route, L3 is allocation of the same pool, and L4 is supplier pass-through inside L3 revenue.

Accounting guardrail
How to read it

Read the chain as one money-flow system

Start with who creates AI demand, then separate who places orders, who receives the first dollar, and which upstream suppliers benefit inside that dollar.

Selected path
US hyperscaler AI programsHyperscaler cloudAI accelerators / systems
L1 demand origin
$795-1047B
L2 order path
routing layer, not extra spend
L3 first-dollar recipients
100%
L4 upstream suppliers
not counted again
AI accelerators / systems

Direct spend paid to accelerator and AI server vendors. Foundries, HBM, and advanced packaging are upstream cost pass-throughs inside this revenue, not separate add-ons.

Examples
TSMSK HynixSamsungMUCoWoS

Layer boundaries prevent double-counting: L2 routes L1 demand, L3 allocates the same pool, and L4 is supplier pass-through inside L3 revenue.

L1. AI CAPEX demand pool
CompaniesRoleAI rangeConfidence
Microsoft
MSFT
US hyperscaler$150-170Bpublic guidance
Alphabet
GOOGL
US hyperscaler$140-160Bpublic guidance
Amazon
AMZN
US hyperscaler$130-155Bpublic guidance
Meta
META
US hyperscaler$105-130Bpublic guidance
OpenAI / Stargate
Private
AI-native lab$80-120Bcompute commitment
Anthropic
Private
AI-native lab$45-80Bcloud commitment
Oracle
ORCL
AI cloud capacity$45-58Bguidance + estimate
xAI
Private
AI-native lab$30-55Bcluster estimate
ByteDance
Private
China AI cloud / apps$24-32Banalyst estimate
Alibaba
BABA / 9988.HK
China cloud$15-22Bmulti-year plan
Tencent
0700.HK
China cloud / internal AI$11-17Brun-rate estimate
Tesla
TSLA
AI / autonomy$8-13Bpublic guidance
Baidu
BIDU / 9888.HK
China AI cloud$5-9Banalyst estimate
DeepSeek
Private
China AI lab$2-8Bopaque estimate
Mistral AI
Private
Europe AI lab$2-6Bopaque estimate
MiniMax
Private
China AI lab$1-4Bopaque estimate
Moonshot AI
Private
China AI lab$1-4Bopaque estimate
Zhipu AI
Private
China AI lab$1-4Bopaque estimate
L2. Capacity operators
routing layer, not extra spend

L2 answers who converts L1 AI CAPEX demand into real purchase orders. It is an execution path, not incremental spend.

L3. AI compute-center resource allocation β†’ L4. Upstream pass-through

L3 is a parallel allocation of the L1 AI CAPEX pool. These categories add up to L1; L4 upstream pass-through is shown separately and is not additive.

Upstream pass-through, not additive
TSM / HBM / CoWoS / OSAT / materials
not counted again
Who captures the spend?
Flow of capital
Orders
Capacity operator
AI systems
Facility stack
Upstream pass-through
Interactive drilldown
AI CAPEX source estimates
Direct AI / cloud infrastructure pool
$795-1047B
L1. AI CAPEX demand pool
$795-1047B
L1 β†’ L2 β†’ L3 allocation path
Allocation logic
L1. AI CAPEX demand pool$795-1047B
Direct AI / cloud infra estimate$795-1047B
L2. Capacity operator: who places orders$795-1047B
L3. Direct payee categories$795-1047B
L2: capacity operator

routing layer, not extra spend

L3: direct payee categories
Selected node
Direct AI / cloud infra estimateHyperscaler cloudAI accelerators / systems
Hyperscaler cloud
self-build + customer cloud
L2. Capacity operator: who places orders
$534B
58%
routing layer, not extra spend

The largest path. In many cases the demand owner and platform owner are the same company: Meta, Google, Amazon, and Microsoft turn internal AI and customer demand into their own data-center, chip, network, and power orders.

US cloud
$374B
AzureAWSGoogle CloudOCI
Internal AI infra
$117B
Meta AIGoogle TPUAWS Trainium
Enterprise cloud AI
$43B
CopilotVertex AIBedrock
AI accelerators / systems
Direct AI / cloud infra estimate
$442B
48%

Direct spend paid to accelerator and AI server vendors. Foundries, HBM, and advanced packaging are upstream cost pass-throughs inside this revenue, not separate add-ons.

NVIDIA platform
$327B
GB200/GB300HGXNVL racksInfiniBand/NVL
AMD / merchant accelerators
$44B
MI300/MI350Instinctaccelerator cards
Custom ASIC systems
$71B
TPUTrainiumMaiaMTIA
L4: direct payees / upstream pass-throughs
Direct payees
MSFTAMZNGOOGLMETAORCLNVDAAMDAVGODELLSMCI
Upstream pass-through, not additive
TSMSK HynixSamsungMUCoWoS

Disclaimer: All research generated by AlphaVector is produced by AI agents and is intended solely for educational and informational purposes. It does not constitute financial advice, investment advice, a recommendation to buy or sell any security, or any form of personalized financial guidance. Past performance is not indicative of future results. AI-generated analysis may be incomplete, incorrect, or outdated. Always conduct independent research and consult a qualified financial advisor before making any investment decision.

Terms of ServicePrivacy PolicyRefund Policy