Huayuan Securities: The wave of high-power AI computing is creating opportunities for domestic substitution of thermal interface materials.

Huayuan Securities released a research report stating that the global thermal interface materials industry has strong growth potential, with AI computing power iteration constituting the core driver of the industry. Currently, the high-end TIM market remains dominated by overseas materials giants. Domestic manufacturers, leveraging the development dividends of the local computing power industry chain, have achieved technological breakthroughs in new thermal interface materials such as vertically oriented graphene and metal composites, which may accelerate the domestic substitution process. Industry chain players with independent material R&D capabilities and coordinated domestic and international market layouts are likely to benefit. Therefore, the "Outperform" rating for the communications industry is maintained. The main views of Huayuan Securities are as follows: AI computing power density continues to rise, and thermal management solutions are upgrading from server-end cooling to rack-level and chip-proximate coordination AI infrastructure is evolving from single servers to high-density racks and large-scale computing clusters, and the thermal management target is expanding from single chips and single servers to the rack and even data center level. As chip power consumption, cabinet power density, and cluster scale increase simultaneously, thermal management needs to simultaneously strengthen chip-proximate heat conduction, heat diffusion, and system-level cooling capabilities, with heat dissipation standards continuously upgrading. From the heat transfer chain perspective, chip heat is sequentially transferred through package-internal conduction, interface thermal conduction, and heat diffusion, and is ultimately discharged outward by air cooling or liquid cooling systems. TIM1, TIM1.5, and TIM2 are located at different thermal interface positions, imposing differentiated requirements on BLT, interface thermal resistance, warpage adaptation, and long-term reliability. The high-performance TIM evaluation system has shifted from a single thermal conductivity metric to a comprehensive consideration of interface thermal resistance, Z-direction heat transfer capability, mechanical adaptation, and thermal cycle stability. System-level liquid cooling penetration is accelerating, and low thermal resistance near the chip is driving continuous upgrading of high-performance TIM The theoretical thermal conductivity of graphene can reach 2000-3000 W/(mK), but the pristine sheet structure only possesses high thermal conductivity in-plane; through vertical orientation and longitudinal stacking structural design, continuous Z-direction thermal conduction channels can be constructed to fully unleash the heat transfer potential in the thickness direction. Compared with traditional polymer-based TIM, which can only achieve thermal conductivity levels of single digits to around ten-plus W/(mK), vertically oriented graphene TIM can achieve thermal conductivity performance of 680 W/(mK). Domestic manufacturer Shenzhen HFC's related products achieve a maximum thermal conductivity of 130 W/(mK), with thermal resistance of 0.04-0.06 cm2/W, while also possessing excellent rebound, compressive stress, and thermal resistance stability, capable of adapting to microscopic interface unevenness and chip warpage. Compared with traditional polymer TIM, it possesses outstanding performance advantages of high Z-direction thermal conductivity and low interface thermal resistance. AI computing power drives accelerated growth in heat dissipation demand Leading manufacturers such as NVIDIA and Google continue to increase AI data center construction and iterate high-power computing platforms, driving comprehensive growth in full-rack heat dissipation demand and driving continuous expansion of the global thermal management market. In fiscal year 2025, the total CAPEX of overseas leading cloud manufacturers including Microsoft, Google, Amazon, Meta, and Oracle was US$401.3 billion. Google's 26FYQ2 financial report further significantly raised its full-year capital expenditure forecast from the original US$180-190 billion to US$195-205 billion. AI computing power iteration is driving a significant climb in leading manufacturers' AI chip power consumption. NVIDIA's GPU single-chip power consumption has jumped from 400W for the A100 to a peak power consumption of 2300W for the Rubin platform, and Google's TPU has also achieved iteration from the hundred-watt level to the kilowatt level. Rising chip power consumption is driving continuous increases in full-rack power, and platforms such as GB200, GB300, and Rubin Ultra have successively shifted to rack-level full liquid cooling architectures, with traditional air cooling gradually reaching physical limits. The high-power trend of AI data centers has become established and will continue to drive the release of demand for thermal management products such as cold plates and high-end thermal interface materials. The full-chain heat dissipation upgrade from chip to interface to rack opens up industry growth space. Global thermal management space is broad, and domestic heat dissipation suppliers are accelerating their breakthrough MarketsandMarkets data shows that the global thermal management market size was approximately US$10.7 billion in 2022 and is expected to reach US$19.3 billion by 2028, corresponding to a CAGR of 9.7%. At the segment level, QYResearch statistics show that global thermal interface material sales in 2024 were US$2.012 billion and are expected to grow to US$4.148 billion by 2031. Overseas leading manufacturers still occupy a dominant position in the global TIM market. In 2024, the market shares of DuPont, Dow Chemical, and Henkel were 11.5%, 6.48%, and 6.05%, respectively, and companies such as Henkel, Shin-Etsu, Laird, and 3M continue to occupy the main market share of high-end TIM. The domestic market has become a place where domestic manufacturers are accelerating their breakthrough. According to Frost & Sullivan data, in 2025, Shenzhen FRD Science & Technology ranked first in the domestic TIM market with a market share of 13.1%, and local enterprises such as Jones Tech Plc, Guangdong Suqun New Material, and Shenzhen HFC have all achieved corresponding share positions and are gradually participating in market competition. Risk warnings Risks of global AI computing power capital expenditure falling short of expectations, risks of price fluctuations in bulk commodities and key raw materials, risks of high-end customer certification and introduction falling short of expectations, and risks of industry technology route iteration.