Featured Stocks

Designs the GPUs and full-rack systems that power nearly every frontier AI model. Its CUDA software and NVLink networking keep customers locked in.

Still the default choice for AI compute — data-center revenue keeps setting records.

Nvidia's chief GPU rival. Instinct accelerators and EPYC server CPUs give data centers a second source for AI compute.

A 6-gigawatt Instinct GPU partnership with OpenAI re-rated the stock within a year.

Builds custom AI chips (XPUs) for hyperscalers like Google and Meta. Also supplies the Ethernet silicon that links giant AI clusters.

The main alternative to Nvidia at hyperscale, with an enormous custom-chip backlog.

Designs custom AI accelerators for hyperscalers. Also makes the optical and networking chips that move data inside AI clusters.

A core custom-silicon supplier for AI data centers — big upside, big swings.

The world's largest contract chip maker. Nearly every advanced AI processor is physically manufactured in its fabs.

The indispensable backbone of the AI boom — leading-edge capacity is effectively sold out.

Dutch maker of the EUV lithography machines needed to print the most advanced chips. A true monopoly at the leading edge.

Every leading-edge fab depends on ASML — the ultimate choke point of the AI-chip supply chain.

Sells AI three ways: Azure cloud, frontier models via OpenAI, and Copilot inside Office. AI flows through apps businesses already pay for.

Azure AI demand keeps outrunning the data-center capacity Microsoft can build.

Runs Google Cloud, designs its own TPU chips, and builds the Gemini models. A fully integrated AI stack from silicon to software.

One of the few players that owns every layer, and its capital spending keeps climbing.

Operates AWS, the world's largest cloud platform. Its own Trainium and Inferentia chips cut its reliance on Nvidia.

AWS is the single biggest renter of AI compute on the planet.

One of the largest builders of hyperscale AI infrastructure. Its data centers and custom silicon power the Llama models and its ad engine.

Among the world's biggest AI spenders — the payoff shows up in sharper ad targeting.

Oracle Cloud Infrastructure has become a surprise AI-cloud winner. It signs huge multiyear deals to host AI training workloads.

Landed some of the largest cloud contracts ever signed, though the stock has pulled back from its highs.

The leading "neocloud" — a pure-play AI hyperscaler. It rents Nvidia GPU capacity at scale to AI labs and enterprises.

A pure bet on AI-compute rental demand: explosive growth, high volatility.

Builds compact Aurora fast reactors sold as power-as-a-service to data centers. Sam Altman took it public in 2024.

Won a key U.S. safety approval for its first Idaho reactor and locked in fuel for a Meta data-center project.

The only small-modular-reactor design approved by the U.S. nuclear regulator. Its 77 MW factory-built modules stack into full power plants.

Construction on its 6 GW Tennessee Valley Authority program could begin later this year.

Operates the largest nuclear fleet in the United States. The default partner when a hyperscaler wants firm, carbon-free power.

Signed a 20-year deal with Microsoft to restart Three Mile Island purely to feed AI data centers.

One of the world's largest uranium producers and co-owner of Westinghouse. A picks-and-shovels play on any nuclear build-out.

Every new reactor needs uranium — Cameco supplies the front end of the fuel cycle.

The only company enriching uranium on U.S. soil. Also the first licensed Western producer of HALEU, the fuel most next-gen SMRs need.

HALEU supply is the choke point for the whole SMR industry, and Centrus is the first mover.

One of the largest independent U.S. power producers, pairing nuclear with gas and batteries. Increasingly contracted straight to data centers.

A go-to merchant supplier of firm, carbon-free power for round-the-clock AI demand.

The only U.S.-based maker of DRAM and NAND memory. A leading supplier of the high-bandwidth memory (HBM) beside every AI accelerator.

Record results and unprecedented multiyear contracts as DRAM and NAND stay in acute shortage.

Pure-play NAND flash maker spun out of Western Digital in 2025. Supplies the flash inside SSDs, devices and AI data centers.

The best-performing large-cap of early 2026, riding a multi-year NAND shortage.

Now the pure hard-disk-drive leader after spinning off SanDisk. Its high-capacity drives store data-center and AI-training data.

Bulk hard drives are the cheapest home for AI's exabytes, and hyperscalers are buying every drive.

The other hard-disk-drive giant. Its HAMR technology packs ever more capacity onto each drive for hyperscale storage.

HAMR drives command premium pricing, pushing Seagate to record highs.

Designs the controller chips that make NAND flash usable in SSDs and memory cards. A picks-and-shovels supplier to every flash maker.

Rising NAND demand lifts controller volumes across the whole SSD market.

Sells enterprise all-flash storage arrays and hybrid-cloud data management. The systems layer that turns raw drives into AI-ready infrastructure.

All-flash and AI workloads keep its storage systems in demand.

Sells AI operating platforms — AIP, Foundry and Gotham. They put large language models to work inside governments and Fortune-500 workflows.

One of the purest software plays on enterprise AI adoption.

The largest customer-relationship-management platform. Now pushing Agentforce, autonomous AI agents inside its enterprise apps.

Betting its next leg of growth on AI agents sold on top of the customer data it already hosts.

Runs the workflow platform big enterprises use for IT, HR and operations. AI is layered across every module.

A leading pick for AI agents embedded in day-to-day enterprise systems.

The cloud platform where enterprises store and query their data. Now adding AI and analytics tools on top.

Enterprise AI needs data in one place, and Snowflake is where much of it lives.

The leading cloud-native cybersecurity platform. Uses AI to stop threats across endpoints, cloud and identity.

AI powers new attacks too, keeping AI-driven security in durable demand.

The observability platform that monitors cloud apps and infrastructure. Increasingly the tool teams use to watch AI systems in production.

As AI apps go live, monitoring those workloads becomes its own growth driver.

Supplies the power, cooling and thermal systems that keep dense AI racks running. A direct picks-and-shovels play on the data-center build-out.

Liquid cooling for AI racks is a fast-growing product line.

An AI "neocloud" spun out of Yandex. Builds and rents GPU cloud capacity to AI developers.

A pure-play bet on renting AI compute, backed by an Nvidia investment.

A power-management giant. Supplies the switchgear and backup systems that feed power into data centers.

Its data-center order book has swelled with the AI build-out.

Builds the high-speed Ethernet switches that link thousands of GPUs into one AI cluster. The networking backbone of large training systems.

The hyperscaler favorite for solving AI's networking bottleneck.

Assembles AI servers and full racks, including liquid-cooled systems. Ships Nvidia and AMD silicon to data centers at speed.

A high-beta way to play AI server demand — the most volatile name in this group.

A data-center REIT that owns and leases the physical buildings. The real estate AI compute actually sits in.

Landlord to the AI boom: demand for space is outrunning supply.