AI Infrastructure Stocks
AI infrastructure stocks, often called AI data center stocks, are the listed companies that make or buy the equipment inside AI data centers, from GPU chips, memory and servers to the transformers, switchgear and chillers that power and cool every site. Stockatomics links each one to the components it makes or buys, and every component carries its own monthly benchmark price, so you can follow the AI buildout through the physical equipment behind each stock. As of 26 October 2026, the list holds 64 listed companies trading on BIT, EPA, ETR, HKG, KRX, NASDAQ, NYSE, SHA, SHE, SWX, TPE and TYO, from the makers of GPUs and transformers to the cloud providers building AI data centers.
- Listed companies
- 64
- Producers
- 47
- Buyers
- 23
- AI components
- 15
List of AI infrastructure stocks
A company is linked to a component as a producer when it makes it and as a buyer when its business depends on it. Share prices, exposure and stock vs benchmark performance for every company are in the Stockatomics Terminal.
What counts as an AI infrastructure stock
Every AI model runs on physical equipment, and AI infrastructure stocks are the listed companies on either side of it. On one side sit the producers, the chipmakers, memory makers, server builders as well as power and cooling suppliers that make the equipment. On the other sit the buyers, the cloud providers and technology groups pouring capital into AI data centers.
A company joins this list when Stockatomics links it to at least one component in the AI Infrastructure Index, as a producer when it makes that component and as a buyer when its business depends on it. That keeps the list tied to the equipment AI data centers are built from, rather than to a theme or a label.
Many of these companies sell far more than AI equipment. A transformer maker also supplies utilities and factories, and a memory maker sells into phones and PCs, so a place on this list shows where a company meets the AI buildout. How much of its business rides on it varies widely from one company to the next.
AI infrastructure stocks by layer
AI data centers are built in four layers, compute, networking, power and cooling, and each layer has its own group of listed producers. The buyers that pay for all four form a fifth group.
AI chip, memory and server stocks
Compute is where most of the money in an AI data center goes. GPU chips, cards and boards train and run the models, DRAM memory keeps them fed with data, and servers and AI racks bring it all together on printed circuit boards packed with capacitors. The producers range from Nvidia and TSMC to memory makers such as SK hynix and server builders such as Dell.
These stocks tend to move with the price and volume of what they ship. Servers and GPUs carry the most weight in the AI Infrastructure Index, so price moves here shape the cost of the whole AI buildout.
Stockatomics links AMD, Broadcom, Dell Technologies, Hon Hai Precision Industry, HP, Ibiden, Kyocera, Micron Technology, Murata Manufacturing, NVIDIA, Samsung Electro-Mechanics, Samsung Electronics, SK hynix, TDK, TSMC, TTM Technologies, Unimicron Technology, Vishay Intertechnology and Yageo to AI compute components.
AI networking and optical stocks
Networking turns thousands of separate GPUs into a single AI cluster. Optical transceivers turn electrical signals into light, optical fibre cable carries that light between racks and buildings, and high-speed connectors join the parts inside each server.
Each step up in GPU speed needs faster, costlier optics, so networking makers tend to move with every new generation of AI systems.
Stockatomics links Coherent, Corning, Eoptolink Technology, Lumentum, Prysmian, YOFC and Zhongji Innolight to AI networking components.
Data center power stocks
Electricity, rather than chips, has become the main limit on how fast new AI data centers can open. Large power transformers and switchgear connect each site to the grid, UPS systems and power converters keep the supply steady, copper cable carries it to every rack, while gas turbines increasingly generate it on site.
Producers such as GE Vernova, Eaton and Schneider Electric have faced long lead times as AI data centers and utilities order the same equipment at once. Some of the new power is set to come from nuclear plants as well, which links the AI buildout to uranium stocks.
Stockatomics links ABB, Baoshan Iron & Steel, Cleveland-Cliffs, Delta Electronics, Eaton, GE Vernova, HD Hyundai Electric, Hitachi, Mitsubishi Heavy Industries, Nexans, Nippon Steel, POSCO, Prysmian, Schneider Electric, Siemens, Siemens Energy, thyssenkrupp and Vertiv to AI power components.
Data center cooling stocks
Almost every watt that reaches an AI rack leaves it again as heat. Chillers and air-conditioning units carry it out of conventional data halls, while the densest AI racks increasingly rely on liquid cooling and heat exchangers that draw it straight off the chips.
AI racks pack far more power into the same space than ordinary servers, so they need far more cooling, which brings makers such as Vertiv into the heart of the AI buildout.
Stockatomics links Carrier, Daikin Industries, Johnson Controls, Trane Technologies and Vertiv to AI cooling components.
Companies building AI data centers
On the other side of every order sit the buyers. The largest technology companies, often called the hyperscalers, along with cloud providers and AI labs, are spending on GPUs, servers, power and cooling at a pace few industries have ever seen, driven by ever larger models and the race towards superintelligence.
Microsoft, Amazon, Alphabet, Meta, Oracle and CoreWeave are among the companies Stockatomics links to these components as buyers. For them, the same rising price that signals pricing power for a producer is a growing cost, which is why buyers and producers on this list can move in opposite directions.
Stockatomics links Alphabet, Amazon, Apple, Arista Networks, Cisco Systems, Constellation Energy, CoreWeave, Dell Technologies, Digital Realty, Duke Energy, Equinix, GE Vernova, Hitachi, HP, Meta Platforms, Microsoft, NextEra Energy, NVIDIA, Oracle, Quanta Services, Siemens Energy, Tesla and Vistra to AI infrastructure as buyers.
What moves AI infrastructure stocks
Capital spending by the largest buyers sets the pace. When the hyperscalers raise their spending plans, orders flow down to the chip, memory, server, power and cooling makers on this list, and when they hold back, the whole group tends to feel it.
Below that, each layer runs on its own cycle. Memory swings between shortage and glut more than almost any other component, GPUs move with each new generation, and power equipment is held back by factory capacity and long lead times.
Copper prices, tariffs and export rules on advanced chips add further swings. That is why the price of the components behind a company often tells a more specific story than the AI theme as a whole.
Why component prices matter for AI infrastructure stocks
Share prices react to news, earnings and sentiment. Component prices record what is happening in the physical market underneath, what producers are charging for GPUs, memory and transformers, as well as what buyers are paying to fit out their AI data centers.
Every component on Stockatomics has its own monthly benchmark price. A sustained rise in something a company makes can point to tight supply and pricing power, while for the companies that buy it, the same rise is a growing cost. A fall points the other way, towards easing supply or slowing demand.
Because these prices move month by month, shifts in the AI buildout often show up in them before they reach quarterly results or capital spending guidance. The Stockatomics Terminal sets each stock against the benchmarks of the components it depends on, so you can see whether the physical market is moving with its share price or against it.
AI infrastructure stocks and the AI Infrastructure Index
The Stockatomics AI Infrastructure Index tracks what it costs to build AI data centers, month by month. It combines the benchmark prices of the GPUs, servers, memory, optics, power and cooling equipment inside them, weighted by how much is spent on each. Every company on this page is linked to at least one of those components.
The index tracks equipment prices, not share prices, which is what makes it useful beside them. A rising index points to tight supply and pricing power for the producers on this list and rising costs for the buyers, while a falling index points to easing supply. Read together, the two show which side of the AI buildout is gaining ground.
From AI data centers back to the mines
The equipment inside AI data centers starts out as critical minerals. Copper runs through every cable and transformer winding, while gallium and germanium go into chips, power electronics and optics.
Stockatomics follows that chain all the way down, with benchmark prices for each mineral and the listed companies that produce them. For the mining side of the AI buildout, see copper stocks and uranium stocks, the gallium and germanium price pages, or every mineral and component AI depends on in critical minerals and components for AI.
AI infrastructure stocks FAQ
What are AI infrastructure stocks?
Shares in listed companies that make or buy the equipment AI data centers are built from. They include the makers of GPU chips, memory and servers, the optical networking suppliers that link them, the power and cooling equipment makers that keep each site running, as well as the hyperscalers and cloud providers buying it all.
Which companies make AI data center equipment?
Stockatomics links 47 listed companies to AI data center components as producers. They are ABB, AMD, Baoshan Iron & Steel, Broadcom, Carrier, Cleveland-Cliffs, Coherent, Corning, Daikin Industries, Dell Technologies, Delta Electronics, Eaton, Eoptolink Technology, GE Vernova, HD Hyundai Electric, Hitachi, Hon Hai Precision Industry, HP, Ibiden, Johnson Controls, Kyocera, Lumentum, Micron Technology, Mitsubishi Heavy Industries, Murata Manufacturing, Nexans, Nippon Steel, NVIDIA, POSCO, Prysmian, Samsung Electro-Mechanics, Samsung Electronics, Schneider Electric, Siemens, Siemens Energy, SK hynix, TDK, thyssenkrupp, Trane Technologies, TSMC, TTM Technologies, Unimicron Technology, Vertiv, Vishay Intertechnology, Yageo, YOFC and Zhongji Innolight, and each one’s page shows the components it makes and their monthly prices.
Which companies are building AI data centers?
The buyers on this list, led by the hyperscalers and cloud providers. Stockatomics links Alphabet, Amazon, Apple, Arista Networks, Cisco Systems, Constellation Energy, CoreWeave, Dell Technologies, Digital Realty, Duke Energy, Equinix, GE Vernova, Hitachi, HP, Meta Platforms, Microsoft, NextEra Energy, NVIDIA, Oracle, Quanta Services, Siemens Energy, Tesla and Vistra to AI data center components as buyers.
Where are AI infrastructure stocks listed?
On BIT, EPA, ETR, HKG, KRX, NASDAQ, NYSE, SHA, SHE, SWX, TPE and TYO. Each company’s page shows its exchange and ticker.
What are the best AI infrastructure stocks?
Stockatomics doesn’t pick stocks or make recommendations. It shows the physical side of each company, the components it makes or buys and their monthly prices, so you can see where supply is tightening and whose costs are rising, and weigh that alongside the share price.
What is the difference between AI stocks and AI infrastructure stocks?
AI stocks is a broad label that can cover any company tied to artificial intelligence, including software and apps. AI infrastructure stocks are narrower, the companies that make or buy the physical equipment AI runs on, from GPUs and memory to transformers and cooling.
Are data center stocks the same as AI infrastructure stocks?
They overlap. Data center stocks often means the companies that own and run data centers, while AI infrastructure stocks reach back to the makers of the chips, servers, power and cooling equipment inside them, as well as forward to the hyperscalers paying for it.
What are data center cooling stocks?
Shares in companies that make the equipment that removes heat from data centers, such as chillers, air-conditioning units, liquid cooling systems and heat exchangers. AI racks produce far more heat than ordinary servers, which has made cooling a central part of every AI data center build. The Cooling filter on this page shows the ones Stockatomics links to AI data centers.
Which AI infrastructure stocks make chips?
The compute producers on this list, such as Nvidia for GPUs, TSMC, which manufactures advanced chips, and SK hynix for memory. Filter the list by Compute to see every one, with the components each makes.
Why are power stocks part of AI infrastructure?
Because electricity has become the main limit on how fast AI data centers can open. Every site needs transformers, switchgear, UPS systems, cable and often its own gas turbines, and the makers of that equipment have faced long lead times as AI data centers and utilities order at the same time.
Are hyperscalers AI infrastructure stocks?
Yes, as buyers. Microsoft, Amazon, Alphabet, Meta and Oracle spend heavily on GPUs, servers, power and cooling for their AI data centers, so Stockatomics links them to those components as buyers. For them, component prices are a cost rather than a source of revenue.
How do component prices affect AI infrastructure stocks?
For producers, rising prices for what they make can point to tight supply and pricing power, while falling prices can signal oversupply. For buyers the same components are a cost, so rising prices can squeeze returns unless they are passed on. Because Stockatomics prices each component monthly, these shifts can show up before they appear in company results.
How is this list different from an AI ETF?
An ETF is a fund holding a basket of shares chosen and weighted by its provider. This page lists individual companies, each linked to the AI data center components it makes or buys, with a monthly price for every component, so you can see what sits behind each stock.
How do critical minerals connect to AI infrastructure stocks?
The equipment inside AI data centers is made from them. Copper runs through cable and transformers, while gallium and germanium go into chips and optics. Stockatomics links each mineral to the companies that produce it, from copper stocks to uranium stocks for nuclear power.
How often is the list updated?
Companies are added as Stockatomics links new ones to AI data center components, and every component price updates monthly. The page always shows the current list.
Where can I see share prices and exposure for these stocks?
In the Stockatomics Terminal, which shows each company’s share price, its exposure to every mineral and component it is linked to and how its stock has moved against the benchmarks of the components it depends on.