Know what you’re betting on.

NVDA

NVIDIA

AI compute systems — GPUs, the CUDA software stack, and NVLink rack-scale interconnect sold as one machine.

TSM

TSMC

Leading-edge logic wafer manufacturing and CoWoS advanced packaging for nearly every AI chip.

AVGO

Broadcom

AI-cluster fabric — network switching, optical transceivers, and hyperscaler custom-accelerator co-design.

000660.KS

SK Hynix

High-Bandwidth Memory (HBM) design and fabrication — the current schedule and quality leader.

MU

Micron

HBM and enterprise DRAM supply — the only US-headquartered frontier DRAM maker.

005930.KS

Samsung Electronics

DRAM and HBM supply plus a logic foundry — the swing vendor of the AI supply chain.

LITE

Lumentum

Indium-phosphide lasers and co-packaged-optics components — the start and end of every optical link.

COHR

Coherent

Frontier transceiver laser arrays and optical materials — the second indium-phosphide vendor.

TSEM

Tower Semiconductor

Specialty foundry for silicon-photonics manufacturing and transceiver chips.

WOLF

Wolfspeed

Silicon-carbide (SiC) power semiconductors and solid-state grid-conversion hardware.

STX

Seagate

Hard-disk-drive systems and HAMR technology for petabyte-scale AI cold storage.

WDC

Western Digital

Enterprise HDD and NAND flash systems — the second supply path for AI cold storage.

GOOGL

Google

The full-stack AI player: Gemini models, DeepMind research, custom TPUs, and the distribution of Search, Android, YouTube, and Cloud.

META

Meta

Social apps at billions of users, the Llama open-weight ecosystem, Meta Superintelligence Labs, and a wearable bet on glasses.

AMZN

Amazon

AWS hyperscale cloud, custom Trainium and Inferentia silicon, plus retail logistics and warehouse robotics.

AAPL

Apple

The device layer where personal AI should live: on-device inference, Private Cloud Compute, and Siri, across a billion-plus install base.

NFLX

Netflix

Global streaming at scale, using AI for recommendation, localization, ad targeting, and creative previsualization.

EPYC server CPUs and Instinct MI-series GPUs, the credible second source to NVIDIA in AI accelerators.

EUV and High-NA EUV lithography scanners, the sole supplier of the tool every leading-edge chip needs.

High-performance Ethernet switching that scales GPUs into a training fabric, the open-networking alternative for AI clusters.

Hyperscaler custom-accelerator co-design, electro-optics, and data-center networking, the second custom-silicon toll road.

Azure cloud, the OpenAI partnership, and Copilot across Windows, Office, and GitHub, AI sold through the channels enterprises already buy.

Data-center power distribution, UPS, busways, and the liquid cooling that high-density AI racks now require.

Electrical switchgear, busways, and power-management gear for data centers, plus electrification across industry, aerospace, and vehicles.

Gas turbines, grid equipment, and wind, the generation and transmission side of the AI power crunch.

A specialized GPU cloud renting NVIDIA capacity (Hopper, Blackwell) to AI labs and enterprises.

A European GPU cloud (spun out of former Yandex assets) renting NVIDIA capacity to AI customers.

A data-center operator (rooted in bitcoin mining) converting owned, energized sites into AI and HPC GPU-cloud capacity.

PCIe and CXL retimers, smart cable modules, and fabric switches that move data inside and between AI servers.

An outsourced assembly and test (OSAT) provider, including the advanced packaging AI accelerators depend on.

OCI cloud infrastructure scaling into AI compute, on top of a database, ERP, and applications (Fusion, NetSuite) franchise.