EN / 中文

Diamond-Copper Thermal Management Enters Mass Production Era, Yilin Tech Focuses on AI Server Liquid Cooling Heat Spreaders

by DTbandaoti·November 17, 2026

On November 13, the opening ceremony of Hangzhou Yilin Technology Co., Ltd. (hereinafter referred to as "Yilin Technology") was held at the Hangzhou Chaoyang Industrial Park.

Yilin Technology is a high-tech thermal management manufacturing enterprise serving AI computing infrastructure, primarily engaged in the R&D, production, and sales of diamond-copper materials and liquid cooling heat spreaders. Its core indicators for diamond-copper thermal management materials are at an industry-leading level. The products have been sent to leading industry customers for sample verification, with application scenarios covering areas with high thermal dissipation demands such as AI server liquid cooling, optical communications, new energy, and high-power lasers.

I. Review of National Diamond-Copper Projects This Year

2026 is dubbed by the industry as the "Year One of Diamond Thermal Management," and domestic diamond-copper projects have begun to land intensively.
According to incomplete statistics, 18 representative projects in the domestic diamond-copper sector have landed or advanced intensively this year. The companies and institutions involved include Huifeng Diamond, Minglang Technology, Sinomach Precision, Henan Carbon True Core Materials, Rejin New Materials, Starwave Communication, Diamond Thermal Control, Xinjiang Carbon Core Technology, Botao New Materials, Shangxin Crystal Engineering, etc.

On the upstream material side, veteran superhard materials company Huifeng Diamond invested CNY 1 billion in September to build a digital-intelligent technology park, planning an annual output of 2 billion carats of high-thermal-conductivity composite heat spreaders, demonstrating the determination to rapidly scale up capacity on the material side. In terms of high-end semiconductor applications, the total investment in Ruixi Semiconductor's silicon carbide substrate and diamond-copper material project launched in September reached CNY 1.5 billion, with a planned annual output of 7 million diamond-copper heat sink wafers, directly addressing the pain points of high-end packaging thermal management. On the system and device side, Suzhou Botao has planned an annual output of 30 tons of diamond-copper composite heat dissipation components as of September.

These projects cover the entire chain from powders and composite materials to heat sink wafers and liquid cooling plates, marking that diamond-copper has crossed the "laboratory sampling" stage and entered the deep waters of industrialization where capacity arms races and customer verification proceed in tandem.

The intensive expansion of capacity is primarily driven by "validated demand" from the application side. In April this year, Sugon Data Innovation's megawatt-scale phase-change immersion liquid cooling full-rack achieved the first large-scale application of diamond-copper in the domestic liquid cooling industry, improving system thermal conductivity by 80% and chip performance by 10%. In June, the diamond-copper heat dissipation module independently developed by the Ningbo Institute of Materials Technology & Engineering, Chinese Academy of Sciences, was deployed alongside this rack at the Zhengzhou core node of the National Supercomputing Internet. From October 6–8, Suzhou Watech Electronic Technology showcased a 4000W diamond-copper flange power amplifier at EuMW 2026, achieving a mass-production breakthrough for diamond-copper (DCu) in RF power devices. On the consumer electronics side, some enterprises have already begun batch supplying to leading PC brands. Once demand is validated, project landings begin to occur intensively.

In addition to the aforementioned new projects, a number of companies have also advanced product verification or capacity preparation for diamond-copper this year. Although these do not constitute independent projects, they are necessary background for judging the competitive landscape: Huifeng Diamond, Henan Qianyuan Xinzuan under Yellow River Whirlwind, Worldia, Sf Diamond, Sirui New Materials, Dailai New Materials, Chujian New Materials/Dingli Technology, China North Industries Red Arrow/Central South Diamond, and Essence Electronics. Other manufacturers include Changsha Shenghua Microelectronics, Harbin Yisheng New Materials, Haite Xinke, Suzhou Sicui Thermal Control, Anhui Shangxin Crystal Engineering, Hunan Xinfeng, etc.

II. Diamond-Copper: Why It Matters, Where the Challenges Lie, and Its Applications

As the heat flux density of AI chips continuously exceeds the load-bearing limits of traditional materials like copper and silicon, thermal management has shifted from a supporting component to a bottleneck limiting performance release. Diamond's thermal conductivity is 4 to 5 times that of copper, and it is electrically insulating and can be directly bonded to chips, naturally pushing it to the forefront.

However, what has taken the lead in commercialization is not pure diamond, but diamond-copper. Pure diamond can achieve a thermal conductivity of over 2000 W/(m·K), but large-size fabrication is difficult, brittle, hard to process, and costly; whereas compositing diamond particles or films with a copper matrix can achieve thermal conductivities of 500–900 W/(m·K) while retaining copper's machinability, solderability, and tunable coefficient of thermal expansion (CTE). Sinomach Diamond provides an intuitive benchmark: diamond-copper's heat dissipation efficiency is about half that of pure diamond, but its price is only one-tenth. This is currently the route with the highest certainty of industrialization and the most likely to see volume growth first.

But this route is difficult, precisely because of the word "composite." Diamond and copper are naturally non-wetting—the wetting angle remains greater than 110° even at 1150°C, and no reaction occurs between them, relying solely on mechanical bonding. Interfacial pores and gaps directly increase interfacial thermal resistance; diamond conducts heat via phonons while copper conducts via electrons, causing severe interfacial scattering; moreover, their CTES differ by several times, easily leaving internal stress and micro-cracks after sintering and cooling.

There are two main solution approaches. First, adding elements such as W, Mo, Cr, and Ti into the copper matrix to allow them to react at the interface and form carbides that "bridge" the gap. Second, applying metallized coatings directly onto the diamond surface.

However, although theoretical solutions exist, engineering challenges remain: "Samples are easy to make, but mass production is extremely difficult." Currently, batch consistency and yield rates are unstable, and costs are approximately 3 to 5 times that of pure copper. Therefore, judging the quality of a company's diamond-copper products cannot rely solely on the thermal conductivity of samples; one must also look at density, performance curves after thermal shock testing, and mass production yield rates.

But once the mass production threshold is crossed, its application scenarios are very broad. Currently, diamond-copper has been implemented in key scenarios such as microchannel liquid cooling plates, heat dissipation fins, thermal interface materials, and composite heat sink wafers (such as heat sinks under chips, package lids, and electrically insulating heat dissipation substrates), directly addressing the core thermal management pain points of AI computing infrastructure.

III. Conclusion: Don't Just Focus on This Single Chain

From the intensive project landings and corporate deployments described above, it is evident that as the route with the highest current certainty of industrialization, some leading projects have already crossed the "zero-to-one" mass production threshold, while many enterprises are just entering the alternating period of sample verification and capacity ramp-up. However, this is merely the tip of the iceberg of industrial explosion.

For enterprises and investors currently deploying in this track, besides monitoring the upstream and downstream support and peer competition within the diamond-copper chain itself, they must also closely watch progress in polycrystalline diamond and single-crystal diamond. Polycrystalline, single-crystal, and diamond-copper are three product forms of the same industrial foundation: polycrystalline diamond focuses on scaling up size and reducing costs, determining whether larger-area, cheaper heat sink wafers can be used in the future; single-crystal diamond guards the upper limit of thermal conductivity and seed crystal autonomy, determining whether high-end scenarios will be choked off by overseas suppliers. In October this year, the Dongguan Optoelectronics Research Institute of Peking University produced the world's first 12-inch high-thermal-conductivity polycrystalline diamond film, and in November, Diamond Foundry announced 200mm single-crystal diamond. Any breakthrough on either side will directly rewrite the raw material sources, performance ceilings, and bargaining positions for diamond-copper.

The three chains must advance in parallel rather than waiting in line: diamond-copper is responsible for "getting started first," running through customer validation with the lowest barriers to secure orders and cash flow; polycrystalline diamond relies on MPCVD large-area deposition to move towards wafer-level thin films, responsible for "scaling up area"; single-crystal diamond starts with tackling seed crystals, crystal growth equipment, and large-size substrates, responsible for "guarding the ceiling." All three share underlying capabilities and customer verification channels such as MPCVD/HPHT crystal growth equipment, powder metallization, vacuum hot pressing, and coating bonding. Therefore, capacity planning and R&D reserves must leave interfaces for parallel development.