ARTICLE AD BOX
The AI wave has spared no one. For those who have been highly involved in it from the beginning have seen them become rising stars and darlings of the stock market. These companies have invested capital, years of R&D, and developed an ecosystem of manufacturing capabilities that have seen buyers bidding through their noses for their outcomes.
With an explosive demand for AI, developing the infrastructure for the same has seen these frontrunner companies leapfrog established players to exert a newfound influence on the global economy. These are the companies that have been riding the AI wave.
Samsung Electronics
Samsung’s investments in high-bandwidth memory (HBM) and advanced semiconductors have positioned it as a critical supplier for AI infrastructure worldwideChange in share price since Jan 2025: Approximately 574% gain
AI infrastructure role: Samsung is a primary producer for AI-related memory, especially high-bandwidth memory (HBM), generally used for training large language models (LLMs). To keep up with demand, Samsung is expanding its HBM and DRAM production as well as its foundry capacity, as it remains a rather unknown cog in the AI hardware infrastructure ecosystem.
SoftBank
SoftBank’s aggressive investments in AI startups and large-scale data center projects are strengthening its role in the global AI ecosystemChange in share price since Jan 2025: Approximately 272% gain
AI infrastructure role: Masayoshi Son-led SoftBank has been one of the leading backers of the AI ecosystem, especially as an early investor in OpenAI. It further plans to invest Euro 45 billion in France to build 5 GW of AI data center capacity by 2031 to meet surging demand.
Nvidia
Nvidia’s powerful GPUs have become the backbone of generative AI, powering everything from ChatGPT to hyperscale AI data centersChange in share price since Jan 2025: Approximately 74.3% gain
AI infrastructure role: Nvidia’s GPUs (especially the H100, H200 and Blackwell) are widely used as standards for training and serving large AI models in hyperscale data centers. These are designed to run trillion-parameter models at up to 25x lower cost and energy, with Google Cloud, AWS and Oracle Cloud infrastructure depending on Nvidia GPUs for their next generation of AI capacity.
TSMC
TSMC manufactures the advanced AI chips designed by companies like Nvidia and Apple, making it one of the world’s most important semiconductor foundries.Change in share price since Jan 2025: Approximately 103.4% gain
AI infrastructure role: TSMC has been at the center of the booming demand for AI chips, and it is looking to spend about $75 billion to expand capacity to build AI accelerators and advanced packaging, making it the key manufacturing partner for Nvidia, Apple, and other AI chip designers.
Broadcom
Broadcom’s custom AI accelerators and networking solutions help connect and scale massive AI computing clusters across global data centersChange in share price since Jan 2025: Approximately 104.9% gain.
AI infrastructure role: Broadcom designs custom AI accelerators and high-speed networking equipment that help wire up hyperscale AI clusters. The TPU design plus rack-level networking and powers AI-based components, making Broadcom’s custom accelerators invaluable for data center compute.
SK hynix
SK hynix has emerged as a dominant supplier of high-bandwidth memory, a key component required for training and running advanced AI modelsChange in share price since Jan 2025: Approximately 914.8% gain
AI infrastructure role: SK Hynix today has secured about 2/3rds of Nvidia’s HBM supply contracts for its 2026 Vera Rubin AI platform. This has helped the ultra-high bandwidth memory maker a critical chokepoint supplier for AI GPUs, directly constraining or enabling how fast hyperscalers can bring new AI capacity online.
Cisco Systems
Cisco’s next-generation networking infrastructure enables thousands of GPUs to communicate efficiently inside modern AI supercomputersChange in share price since Jan 2025: Approximately 118.6% gain
AI infrastructure role: Cisco has developed the Silicon One G300 switching chip specifically for gigawatt-scale AI clusters. These improve GPU job completion times by about 28% and enable massive 800Gbps/1.6 Tbps data center fabrics. Cisco’s high-bandwidth, low-latency back-end networks link tens of thousands of GPUs in modern AI data centers.
English (US) ·