The AI trade has spread well beyond Silicon Valley.
While Nvidia and the Magnificent Seven dominate headlines, many of this year’s biggest stock winners are the companies supplying the hardware that powers artificial intelligence — and many of them are based in Asia.
“The current phase of AI development is overwhelmingly infrastructural,” Allianz Research said in a recent report, as hyperscalers, governments, and corporations race to build AI infrastructure and expand computing capacity.
Many of those companies occupy critical positions in the AI supply chain. They make the components that allow advanced chips to store data, process information, and communicate at high speed.
The AI infrastructure boom has spilled over into equity markets. Despite volatility driven by tariffs, geopolitical tensions, and recession fears, investors have piled into manufacturers of memory chips, electronic components, and circuit board materials. Those moves have propelled stock markets in South Korea, Japan, and Taiwan to fresh record highs, outpacing gains in the US market.
These stock markets have been volatile in recent weeks, shaving gains off the three booming markets, but analysts remain upbeat.
“We think the semiconductor memory supercycle is still not fully priced in the North Asian markets of Korea and Taiwan,” Goldman Sachs analysts wrote in a note on emerging markets, which doesn’t include Japan, this week.
The bank expects South Korea and Taiwan to post the strongest earnings growth through 2027, though retail trading and shifting sentiment around AI could drive further market volatility.
These were the five best-performing non-US stocks tracked by MSCI’s All Country World Investable Market Index during the first half of 2026:
1. Samsung Electro-mechanics
Unlike Samsung Electronics, which is best known for smartphones and memory chips, sister company Samsung Electro-Mechanics specializes in the behind-the-scenes components used in AI hardware.
The company produces semiconductor substrates that connect advanced chips to circuit boards and multilayer ceramic capacitors that help deliver stable power to high-performance servers.
As demand for AI accelerators has surged, so has demand for these components, helping lift the company’s shares 660% in the first half of 2026.
2. Kioxia Holdings, Japan
Shares of Japanese memory-chip maker Kioxia climbed about 631% in the first half of 2026 as booming demand for AI data centers drove record earnings.
Spun off from Toshiba in 2017, the memory business was sold the following year to a Bain Capital-led consortium for about $13 billion before being rebranded as Kioxia in 2019.
The company makes NAND flash memory and storage products used in AI servers and cloud data centers.
As AI training and inference required ever more high-speed storage, investors bet demand would continue to outpace supply, helping Kioxia become Japan’s most valuable company by market capitalization, around $300 billion.
3. Kingboard Laminates
Hong Kong-listed Kingboard Laminates’ shares jumped about 535% in the first half of 2026 as investors bet on rising demand for AI hardware.
The company supplies materials used to make printed circuit boards, which allow chips and other electronic components to communicate with one another.
As AI data centers require more powerful servers that process and move larger amounts of data, demand has grown for higher-performance circuit boards and the materials used to make them.
4. Yageo Corporation
Shares of Taiwan-listed Yageo climbed about 357% in the first half of 2026 as investors bet on rising demand for AI hardware.
Earlier this year, the company completed its acquisition of Japan’s Shibaura Electronics, adding temperature sensors that can help monitor heat in AI servers and data centers.
Yageo is also one of the world’s largest manufacturers of capacitors and resistors, components found throughout AI hardware.
5. Unimicron Technology
Taiwan-listed Unimicron Technology’s shares climbed about 345% in the first half of 2026 as demand for advanced AI chips surged.
The company makes printed circuit boards and IC substrates, which connect powerful processors to the rest of an AI server.
As chipmakers develop more powerful AI processors, demand has risen for these increasingly sophisticated components.
