Lin Zong-chi, the 85-year-old founder of Taiwan's King Slide Works, has become the richest person in the country, driven by a remarkable increase in demand for components used in AI servers. Since the start of the year, the company's stock has surged nearly 280%, as reported by the Wall Street Journal.
While King Slide specializes in manufacturing furniture hinges and drawer slides, it also produces rail mechanisms for server racks, which support heavy computing equipment and allow for maintenance access without disrupting the cooling system.
According to the Wall Street Journal, King Slide commands approximately 80% of the market for high-performance server slides.
King Slide's Margin Reaches 87%
The demand for its products is increasing alongside the construction of data centers and the proliferation of more powerful AI systems. Modern server racks can cost millions of dollars, making the reliability of relatively inexpensive mechanical components critical.
“When an AI server rack costs millions, clients care more about the quality of the slides than their price,” commented Brady Van, Deputy Director at Counterpoint Research.
In the last quarter, King Slide's gross margin reached 87%, up from around 50% a few years ago. For comparison, Nvidia's margin is about 75%, while TSMC's is 68%.
Jay C. Van, Executive Vice President of King Slide, attributed this profitability to two decades of engineering advancements that have enabled the company to create specialized mechanisms for server equipment.
This business growth has positively impacted Lin's wealth, with Forbes estimating his fortune at approximately $20.3 billion. He has surpassed Foxconn founder Terry Gou, whose wealth has also increased with the development of AI infrastructure.
Source: Forbes.New Server Architectures to Sustain Demand
Analysts predict continued growth in demand for specialized slides as cloud companies transition to proprietary AI accelerators, increase equipment density, and implement liquid cooling. This necessitates redesigning server and rack structures.
The shift among hyperscalers toward their own silicon is no longer just a plan. In August, Google secured an option to purchase up to 58.97 million shares of Marvell Technology as part of an expanded agreement to develop AI processors. Concurrently, Alibaba announced plans to raise $10.2 billion to develop its stack for neural networks, encompassing computing infrastructure, proprietary chips, models, and applications.
Each accelerator varies in size, thermal output, and connectors, requiring its own rack and slides.
The density of data centers is also increasing. The reported capacity of gas power plants being constructed in the U.S. for direct energy supply to data centers has risen from 97 GW at the end of 2025 to over 189 GW by mid-2026.
Nvidia is experiencing a similar generational transition: SpaceXAI plans to scale its Vera Rubin computing platform to gigawatt capabilities.
Currently, King Slide is expanding its production capacity in Houston, Texas, to cater to North American customers. However, competition is intensifying. Daiwa Securities estimates that the company's share of slide supply for Nvidia systems may drop to about 75% next year, as the chipmaker increases its supplier base.
After the certification of new manufacturers, major data center operators will also be able to leverage competition to negotiate lower prices, noted Van. Despite this, King Slide remains optimistic about maintaining its direct dependence on investments in AI infrastructure.
It is worth mentioning that in August, Taiwanese authorities announced payments of "AI dividends" — offering 10,000 TWD (~$314) to each resident in light of the growth in AI-related industries.