How we pick the list
There is no official index of AI stocks. This list is an editorial choice: US-listed companies whose revenue or valuation is materially driven by the AI build-out, grouped by where they sit in the stack. Membership is reviewed when a company's business changes, not on a schedule.
Inclusion criteria
A company is on the list when all of the following hold:
- It is listed on a US exchange, as common stock or an ADR, and files with the SEC (10-K/10-Q or 20-F/6-K).
- Its own reporting ties a material part of revenue, orders, backlog or capital spending to AI infrastructure or AI products: a named segment (Data Center, Cloud, AI semiconductors), a stated AI revenue figure, or AI-driven capital spending. Narrative mentions of AI in marketing do not qualify.
- It occupies a distinct place in the stack that the category pages explain: compute, cloud, memory, networking, power, software, or facilities.
Excluded: funds and ETFs, private companies, companies listed only outside the US, and companies whose AI link is indirect (general industrials, retailers, banks). Additions and removals are recorded here with the date and reason.
Chips & Compute
The processors that train and run AI models, and the factories that make them. Demand for AI compute starts here; everything else in the stack exists to feed, connect, power and house these chips.
NVDA · AMD · AVGO · TSM · ARM · MRVL · INTC · QCOM · also tagged: MSFT · AMZN · GOOGL · META · SMCI
Hyperscalers & Cloud
The cloud platforms and internet giants that buy most of the AI hardware, build the largest data centers, and rent capacity back out. Their capital spending is the single biggest demand signal for the rest of the list.
MSFT · AMZN · GOOGL · META · ORCL · also tagged: CRWV
Memory & Storage
Every accelerator is bottlenecked by how fast it can be fed data. High-bandwidth memory sits next to the GPU; DRAM, flash and hard drives hold the training data and model outputs.
MU · SNDK · WDC · STX
Networking
AI training runs across thousands of chips at once. The switches, optical transceivers and cabling that link them into one machine decide how much of that compute is actually usable.
ANET · CSCO · CIEN · COHR · LITE · also tagged: NVDA · AVGO · MRVL
Power & Cooling
AI data centers draw far more electricity per rack than the previous generation and run too hot for air alone. Generators, grid equipment and liquid-cooling suppliers are the physical limit on how fast capacity can be added.
VST · CEG · GEV · ETN · VRT
Software
The companies selling AI-driven software to businesses: data platforms that feed models, applications that embed them, and the tools that monitor them in production. Exposure here is usually a share of revenue, not the whole business.
PLTR · NOW · CRM · SNOW · DDOG · also tagged: MSFT · GOOGL · ORCL · CSCO
Data Centers
The buildings, the server makers that assemble chips into racks, and the specialist clouds that rent GPUs by the hour. Closest to the physical build-out; most sensitive to its pace.
EQIX · DLR · SMCI · DELL · CRWV · also tagged: ETN · VRT
Category tags
Each company has one primary category, which decides where it appears in the tables, and may carry secondary tags where its business straddles groups (a chip designer that also sells networking silicon, a cloud provider that designs its own accelerators).
AI exposure labels
Each company page carries one of three labels. They are editorial judgments, reviewed when the company reports.
- Core: AI demand is the main driver of the business today.
- Significant: A large and growing share of revenue or spending is AI-related.
- Partial: One line of business, or an indirect beneficiary.
Reported versus estimated
On company pages, figures the company itself published (with the period and a link to the filing or release) are shown separately from estimates by analysts or this site. Anything without a source is not shown.
Data sources
Quotes, charts, fundamentals and headlines: TradingView widgets, delayed. Filings and insider transactions: SEC EDGAR, fetched no more than every 15 minutes. Insider transaction rows are parsed from the raw Form 4 XML; the linked filing is the record.