Companies
29 tracked · 19 public · 10 private · 8 regions
29 companies
World's largest cloud provider (AWS) and leading AI infrastructure investor. AWS CAPEX drives the majority of total company spend.
Primary investor in OpenAI with $13B+ committed. Azure AI services are the fastest-growing Microsoft segment. Announced $80B AI infrastructure plan for FY2025.
Google parent. GCP is the third-largest cloud provider. Investor in Anthropic. Extended server depreciation life from 4 to 6 years in 2023, adding ~$3B/yr to earnings.
Open-source AI leader (Llama). Highest CAPEX as % of revenue among Mag7. Building 2GW+ data center capacity. Announced 350K+ H100 GPU cluster (Grand Teton).
World's largest memory chipmaker (DRAM, NAND) and #2 foundry (Samsung Foundry / SF2). Critical AI infrastructure player: dominant supplier of HBM (High-Bandwidth Memory) alongside SK Hynix, and foundry partner for Qualcomm, NVIDIA (secondary), and Exynos. Investing heavily to close HBM yield gap with SK Hynix.
World's largest contract chipmaker — manufactures virtually all advanced AI chips (NVIDIA H/B series, Apple silicon, AMD, Qualcomm). TSMC's CAPEX is the best proxy for global AI chip supply capacity. CoWoS advanced packaging for HBM is a key AI bottleneck.
Japan's largest telecom and the world's 4th-largest datacenter operator. Developed tsuzumi — a lightweight Japanese/English LLM. Investing heavily in IOWN (Innovative Optical & Wireless Network) — photonic computing infrastructure for AI. ~$12-14B annual CAPEX makes NTT one of the largest non-hyperscaler AI infra investors.
World's #1 HBM (High-Bandwidth Memory) supplier — the memory stacked on top of every NVIDIA AI GPU. SK Hynix HBM3E is sold out through 2025/2026 to NVIDIA. The company that wins the HBM race wins AI memory. Also major DRAM and NAND supplier.
China's largest e-commerce and cloud company. Alibaba Cloud (Aliyun) is the leading Chinese cloud provider and top-4 globally. Announced massive AI cloud investment exceeding the prior 10 years combined. Qwen LLM family.
Building Dojo supercomputer for self-driving AI training. CAPEX split between Gigafactories (manufacturing) and Dojo (AI compute). AI compute a growing but minority share.
China's largest social and gaming company. Operator of WeChat. Tencent Cloud is a top-3 Chinese cloud provider. Hunyuan LLM. Rapidly accelerating CAPEX for AI infrastructure since 2023.
Lowest AI CAPEX intensity of Mag7. Apple Intelligence uses Microsoft Azure and Google Cloud rather than building own AI infra. ~$10B CAPEX focused on manufacturing and retail.
India's largest company. Jio AI-Cloud — the most ambitious AI infrastructure project in the developing world. Mukesh Ambani announced 2GW of AI compute (200 exaflop target) for India at Reliance AGM 2024. Partnered with NVIDIA for DGX Cloud and Jio AI-Cloud. The gateway to India's AI market.
Fabless GPU designer. The primary beneficiary of AI CAPEX. Does not build data centers — TSMC manufactures. R&D ($10B+) is a better AI investment signal than CAPEX.
Japanese conglomerate and world's largest technology investor. Co-lead of Project Stargate ($100B commitment). Majority owner of ARM (IPO 2023). Vision Fund has invested $100B+ in AI-adjacent companies. Masayoshi Son co-chairs Stargate with Sam Altman.
China's leading AI and search company. Creator of ERNIE Bot. Baidu AI Cloud is a top-3 Chinese cloud provider. CAPEX much lower than US peers due to scale difference and Ascend chip reliance post-export controls.
Dutch company and sole manufacturer of EUV (Extreme Ultraviolet) lithography machines — the equipment that makes advanced AI chips possible. Without ASML EUV, TSMC cannot make N3/N2 chips for NVIDIA. ASML is the ultimate AI infrastructure chokepoint. US export controls restrict ASML machine exports to China.
South Korea's dominant search and internet company. Developed HyperCLOVA X — Korea's frontier LLM. Operates Naver Cloud and is building AI datacenters. Key AI investor in Korea's sovereign AI ecosystem. Partner with Samsung, SK Telecom, and KAIST for AI research.
UK-based semiconductor IP company — designs CPU/NPU architectures used in virtually all mobile devices and increasingly in AI accelerators. ARM Neoverse powers AWS Graviton, Ampere, and datacenter CPUs. The company earns royalties, not CAPEX-intensive manufacturing. SoftBank IPO'd ARM in Sept 2023; SoftBank retains ~90% ownership.
Creator of GPT-4, ChatGPT, o1, and o3. Raised $110B at $730B valuation (Feb 2026) led by Amazon, NVIDIA, and SoftBank. Co-led Stargate with SoftBank and Oracle — $500B committed over 4 years, $100B Phase 1 in Texas.
Creator of Claude. AI safety-focused lab. Backed by Amazon ($4B) and Google ($2B+). Trains on AWS infrastructure. No owned data centers — cloud-native.
Elon Musk's AI company. Creator of Grok. Built Colossus — 100K H100 cluster in Memphis TN, operational August 2024. Targeting 1M GPUs.
Creator of TikTok and Douyin. One of the world's largest private AI compute operators. Estimated $20B+ infrastructure spend in 2024. Trains Doubao LLM. Major Nvidia customer pre-export controls.
China's largest tech company (private). Manufactures Ascend 910B AI chips — the primary domestic alternative to Nvidia GPUs. CloudMatrix cloud AI platform. Heavily invested in AI infrastructure post US export controls. No public financials.
Chinese AI lab owned by hedge fund High-Flyer Capital. Known for radical compute efficiency — R1 model trained for ~$6M, disrupting assumptions about AI infrastructure requirements. DeepSeek-V3 and R1 challenged frontier model economics.
France-based frontier AI lab and Europe's AI champion. Known for efficient open-weight models (Mistral 7B, Mixtral, Mistral Large). €600M Series B (June 2024) at €6B valuation. Partners with Microsoft and AWS.
Abu Dhabi-based AI and cloud company backed by UAE sovereign wealth. $1.5B Microsoft investment (April 2024). Partner of OpenAI and Technology Innovation Institute (Falcon LLM). Building major AI infrastructure in UAE.
Saudi Arabia's national AI company, owned by the Public Investment Fund (PIF). Announced $100B multi-year AI infrastructure investment. Partners with NVIDIA, AMD, Qualcomm, AWS, and Google Cloud. CEO: Bobby Alluwaishaan. Aims to build 500MW of AI compute in Saudi Arabia and across the region.
German AI company building sovereign European LLMs (Luminous series, PHARIA). Positioned as Europe's answer to US/China AI dominance. €500M+ raised (2024 Series B led by SAP). Partners with German government, Bundeswehr, and European public sector. Focus on explainability (AtMan) and sovereignty.