AI Stocks: The Full AI Compute Supply Chain

Every public ticker that builds, ships, powers or sells AI compute. Click any node in the map below to open its quote, AI score and chart on Fintwit.

🔴 = capacity-constrained (single-source, 18+ month lead time, or sold out into 2027+). Percentages are dollar-share approximations within each stage. 133 US-listed tickers mapped across 8 stages.

Hover any node for ticker, stage, and constraint reason. Multi-listing rows (e.g. AAPL/AMZN/MSFT in-house Arm) open a picker. Foreign-only listings (TPE:, KRX:, HKG:) are shown for context but are not clickable on Fintwit. Updated quarterly.

AI stocks by stage

The AI compute supply chain has eight discrete stages. Each stage contains a handful of dominant public companies — most of which are bottlenecks, not commodities. The list below names every US- tradable ticker in each stage with the company description and stage-share weight, so you can quickly find AI infrastructure stocks, AI semiconductor stocks, hyperscaler stocks, data-center REIT stocks, power-generation AI stocks, and more.

1. Atoms

Polysilicon / quartz
18% of stage
Electronic-grade is tight; commodity poly is not.
Gallium & Germanium
🔴 constrained
5N+ (VNP)
4% of stage
China export controls Aug 2023.
InP substrates
🔴 constrained
6% of stage
Photonics laser bottleneck.
Rare earths
🔴 constrained
3% of stage
Magnets for cooling fans, motors.
Copper
12% of stage
Interconnect, busbars, transformers.
High-purity quartz
🔴 constrained
2% of stage
Spruce Pine, NC nearly sole-sourced.
Photoresist & specialty chemicals
🔴 constrained
22% of stage
EUV photoresist near-monopoly.
Industrial / specialty gases
15% of stage
Ne, Xe, NF3, WF6.
Si wafers (300mm)
🔴 constrained
18% of stage
300mm wafer slots tight.

2. Equipment

EUV lithography
🔴 constrained
28% of stage
Single-source EUV; High-NA limited.
DUV lithography
6% of stage
Etch & deposition
24% of stage
Process control / metrology
9% of stage
Onto on advanced packaging.
Ion implant
3% of stage
CMP / cleaning
4% of stage
Test equipment (ATE)
🔴 constrained
7% of stage
HBM test slots tight.
Wafer-handling / subfab
8% of stage
Adv. packaging tools
🔴 constrained
6% of stage
The CoWoS bottleneck.
Photomask / blanks
5% of stage

3. Foundries

TSMC (N3, N2, CoWoS)
🔴 constrained
62% of stage
Sole-source for AI accelerators.
Samsung Foundry
KRX:005930
11% of stage
Intel Foundry (18A, 14A)
7% of stage
Ramping; not yet at AI scale.
GlobalFoundries
5% of stage
Mature, RF, SiPh.
SMIC (China)
HKG:0981
6% of stage
Restricted from leading nodes.
UMC
4% of stage
Tower
2% of stage
Analog, RF, SiPh.
Vanguard / PSMC
TPE:5347·6770
3% of stage

4. Memory

SK Hynix (HBM3E, HBM4)
🔴 constrained
KRX:000660
50% of stage
Sole HBM3E supplier to Nvidia for periods.
Samsung Memory
🔴 constrained
KRX:005930
30% of stage
Catching up on HBM3E qualification.
Micron (HBM3E, HBM4)
🔴 constrained
18% of stage
Sold out 2025/2026.
Kioxia (NAND for AI)
TYO:285A
2% of stage

5. Packaging

TSMC CoWoS
🔴 constrained
55% of stage
Single tightest bottleneck in AI.
Amkor
12% of stage
Arizona facility for Apple/AMD.
ASE / SPIL
14% of stage
OSAT giant.
Powertech
TPE:6239
3% of stage
Memory packaging.
JCET / Tongfu / Hua Tian
600584·002156·002185
6% of stage
China OSAT trio.
ABF substrates
🔴 constrained
8% of stage
Capacity tight for AI sockets.
Hybrid bonding tools
2% of stage
Sold through OSATs.

6. Chip design

Accelerators

Nvidia (Hopper, Blackwell, Rubin)
🔴 constrained
78% of stage
Allocation-rationed.
AMD (MI300, MI325, MI350)
7% of stage
Broadcom (Google TPU, Meta MTIA)
🔴 constrained
8% of stage
TPU v5/v6 design partner.
Marvell (Amazon Trainium ASIC)
3% of stage
Intel (Gaudi 3)
1% of stage
Tenstorrent / Cerebras / Groq / SambaNova
(private)
3% of stage

CPU

Intel Xeon
35% of stage
AMD EPYC
35% of stage
Gaining share fast in AI servers.
Arm (IP)
20% of stage
Powers Grace, Graviton, Cobalt.
Ampere
(private)
2% of stage
Being acquired by SoftBank.
Hyperscaler in-house Arm
8% of stage
Apple, Graviton, Cobalt.

Networking

Nvidia (NVLink, Spectrum-X)
🔴 constrained
42% of stage
InfiniBand + scale-up.
Broadcom (Tomahawk, Jericho, Bailly)
🔴 constrained
38% of stage
Ethernet AI fabric king.
Marvell (Teralynx)
9% of stage
Optical DSPs.
Astera Labs
🔴 constrained
3% of stage
PCIe / CXL / Ethernet retimers.
Credo
4% of stage
AECs, retimers.
Cisco Silicon One
4% of stage

Photonics

Coherent (transceivers, SiPh)
🔴 constrained
32% of stage
NVIDIA $2B March 2026.
Lumentum (transceivers, EML)
🔴 constrained
28% of stage
NVIDIA $2B March 2026.
Marvell (optical DSPs, Polariton)
7% of stage
Ciena
11% of stage
Coherent transport.
Innolight / Eoptolink
300502·300308
14% of stage
Applied Optoelectronics
3% of stage
Ayar / Lightmatter / POET
(private/POET)
1% of stage
In-package optics, photonic compute.

Power ICs

Monolithic Power
🔴 constrained
22% of stage
Lost some Blackwell sockets to Infineon.
Infineon
ETR:IFX
25% of stage
Won Nvidia Blackwell Vcore.
Texas Instruments
18% of stage
Analog Devices
14% of stage
Vicor
🔴 constrained
4% of stage
Vertical-power-delivery niche.
Renesas
TYO:6723
9% of stage
ON Semi
8% of stage

7. System integration

Server / ODM

Hon Hai / Foxconn
TPE:2317
30% of stage
Largest GB200 rack assembler.
Quanta
TPE:2382
18% of stage
Wistron / Wiwynn
TPE:3231·6669
12% of stage
Inventec
TPE:2356
6% of stage
Super Micro
9% of stage
Dell (PowerEdge XE, AI factories)
8% of stage
HPE (Cray, Blade)
5% of stage
Lenovo
HKG:0992
4% of stage
Cisco UCS
2% of stage
Pegatron / Compal / Asus / MiTAC
4938·2324·2357·3706
6% of stage

Cooling

Vertiv
🔴 constrained
28% of stage
Direct-to-chip leader.
Schneider Electric
EPA:SU
18% of stage
Eaton
9% of stage
nVent
6% of stage
Boyd / Cooler Master / Asetek / CoolIT / JetCool / Accelsius
(mostly private)
20% of stage
Modine
4% of stage
CDU / HVAC.
AVC / Auras
🔴 constrained
TPE:3017·3324
15% of stage
Thermal modules for GB200.

Connectors

Amphenol
🔴 constrained
22% of stage
NVLink cable supplier.
TE Connectivity
18% of stage
Lotes
🔴 constrained
TPE:3533
9% of stage
NVIDIA socket connectors.
Molex
(Koch)
12% of stage
Aspeed (BMC)
TPE:5274
3% of stage
Bizlink
TPE:3665
4% of stage

8. Data center

Hyperscalers

Microsoft (Azure / OpenAI workload)
22% of stage
Alphabet (Google Cloud + DeepMind)
18% of stage
Amazon (AWS + Anthropic)
20% of stage
Meta
14% of stage
Oracle (OCI)
6% of stage
CoreWeave
6% of stage
Nebius
2% of stage
xAI (Colossus)
(private)
4% of stage
Tesla (Dojo)
2% of stage
Apple Intelligence
3% of stage
Tencent / Alibaba / ByteDance / Baidu
3% of stage
China demand is separate supply chain.

REITs

Equinix
🔴 constrained
22% of stage
Powered shells.
Digital Realty
🔴 constrained
18% of stage
Iron Mountain DC
4% of stage
Applied Digital
3% of stage
GDS Holdings (China)
4% of stage
Switch / CyrusOne / QTS / Compass
(private)
14% of stage
All taken private.
Vantage / Stack / EdgeConneX
(private)
16% of stage
Aligned / DataBank / Cologix
(private)
9% of stage

Power gen

Constellation
🔴 constrained
14% of stage
Nuclear baseload.
Vistra
🔴 constrained
12% of stage
Comanche Peak nuclear + gas.
Talen Energy
🔴 constrained
6% of stage
Susquehanna nuclear.
NextEra
11% of stage
Renewable + Duane Arnold restart.
Brookfield Renewable
5% of stage
AES
3% of stage
Regulated utilities (DUK·D·SO·ETR·DTE·XEL·EVRG)
Bloom Energy
🔴 constrained
3% of stage
Oracle, AEP, Brookfield deals.
Oklo / NuScale / BWXT (SMRs)
1% of stage
Future.
GE Vernova (turbines, transformers)
🔴 constrained
10% of stage
80GW gas turbine backlog into 2029.

EPC

Eaton
🔴 constrained
18% of stage
9-yr data center backlog.
Vertiv
🔴 constrained
14% of stage
GE Vernova (transformers, HVDC)
🔴 constrained
12% of stage
Schneider Electric
🔴 constrained
EPA:SU
11% of stage
Siemens Energy
🔴 constrained
ETR:ENR
9% of stage
Hitachi Energy
🔴 constrained
TYO:6501
7% of stage
HVDC.
Hubbell
3% of stage
nVent
3% of stage
Powell Industries
🔴 constrained
2% of stage
Switchgear.
Hammond Power Solutions
🔴 constrained
TSX:HPS.A
1% of stage
Dry-type transformers.
Quanta Services
6% of stage
Grid build EPC.
MasTec
3% of stage
MYR Group
2% of stage
Argan (Gemma EPC)
🔴 constrained
2% of stage
Gas-plant EPC.
Comfort Systems USA
4% of stage
IES Holdings
2% of stage
Modular MEP.
Generac / Cummins / Caterpillar (gensets)
3% of stage
Backup.

Frequently asked questions

What are the best AI stocks to buy right now?

The most popular AI infrastructure names are NVIDIA (NVDA) for accelerators, TSMC (TSM) for foundry capacity, Broadcom (AVGO) for custom ASICs and networking, AMD for CPUs and MI-series accelerators, Micron (MU) and SK Hynix for HBM memory, ASML for EUV lithography, and the hyperscalers (MSFT, GOOGL, AMZN, META, ORCL). Power and cooling exposure runs through Vertiv (VRT), Eaton (ETN), GE Vernova (GEV), Constellation Energy (CEG) and Vistra (VST). The map above shows how every one of them connects to the rest of the stack and which ones are capacity-constrained.

What is the AI compute supply chain?

The AI compute supply chain is the end-to-end path that turns raw materials into a running AI workload: raw materials → cap- equipment makers → foundries → memory → advanced packaging → chip designers → system integrators → data centers. A bottleneck at any stage limits how fast the industry can scale. Today the tightest stages are EUV lithography (ASML), CoWoS advanced packaging (TSMC, BESI), HBM memory (SK Hynix, Micron) and grid power (CEG, VST, GEV).

How often is this list updated?

Quarterly. Stage-share weights, capacity-constraint flags, and ticker assignments are refreshed each quarter based on the companies’ latest filings, earnings transcripts, and verified industry sources.