While the market’s attention is focused on NVIDIA's latest chips, OpenAI's technological iterations, and the capital expenditures of major tech giants, many investors often overlook the underlying logic supporting this AI industrial revolution —basic mineral resources.
The ultimate demand for AI goes beyond computing power and electricity—it requires a massive physical hardware infrastructure. Each data center that rises from the ground is essentially a 'super brain' built with thousands of tons of metals and rare minerals.
In a 'gold rush', the people selling shovels often profit first; in this AI arms race, the giants controlling key minerals are the 'shovel sellers' hidden behind the computing power.

1. Server motherboards and circuits: The 'neural networks' of AI and the super cycle of 'Dr. Copper'
Inside servers, dense circuit boards and connecting wires are responsible for transmitting vast amounts of data and electric currents. These components require not only excellent conductivity but also minimal signal delay.
Ubiquitous 'Dr. Copper': AI servers and data centers have an immense demand for copper in their power infrastructure. From high-voltage direct current (HVDC) transmission to internal wiring, copper is an indispensable material.
Key focus: U.S. stocks$Southern Copper (SCCO.US)$ 、 $Freeport-McMoRan (FCX.US)$ 、 $Hudbay Minerals (HBM.US)$; of the Hong Kong stock market$JIANGXI COPPER (00358.HK)$ 、 $MMG (01208.HK)$ 、 $CHINFMINING (01258.HK)$ , among others.
The Industrial Allure of Precious Metals (Gold, Silver, Platinum, Palladium): Gold and silver are widely used in high-end chip packaging and precision circuit contact points due to their excellent oxidation resistance and conductivity. The proliferation of AI devices has added strong industrial demand logic to these traditional safe-haven assets.
Key focus: U.S. stocks$Newmont (NEM.US)$ 、 $Agnico Eagle (AEM.US)$ 、 $Barrick Mining (B.US)$$Wheaton Precious Metals (WPM.US)$ 、 $Pan American Silver (PAAS.US)$ 、 $Hecla Mining (HL.US)$ 、 $Sibanye Stillwater (SBSW.US)$; of the Hong Kong stock market$ZIJIN MINING (02899.HK)$ 、 $ZIJIN GOLD INTL (02259.HK)$ 、 $SD GOLD (01787.HK)$ 、 $CHI SILVER GP (00815.HK)$ 。
2. Radiators: The 'Cooling Heroes' Under Surging Computing Power
AI chips generate significant heat during high-speed operations. If the heat is not dissipated promptly, it will not only cause a drop in computing power but may even damage the equipment. Before liquid cooling technology becomes widespread, high-performance air-cooled radiators remain mainstream.
Lightweight and Efficient Aluminum: Heat sinks require extremely high thermal conductivity and relatively light weight, making aluminum the primary material for radiators. As the number of high-density server racks increases, so does the demand for high-quality aluminum materials.
Key Focus Targets:$Alcoa (AA.US)$、 $Century Aluminum (CENX.US)$、 $Constellium (CSTM.US)$; of the Hong Kong stock market$CHINAHONGQIAO (01378.HK)$ 、 $CHALCO (02600.HK)$ 、 $RUSAL (00486.HK)$ 。
3. Semiconductors and Microchips: Indispensable 'Strategic Trace Elements'
Chips are not just made of silicon. To push advanced semiconductor processes beyond their limits, a series of special rare metals are indispensable.
Silicon and Gallium: Silicon is the cornerstone of semiconductors, while gallium is widely used in high-frequency, high-power compound semiconductors and is key to improving energy conversion efficiency.
Germanium and Tantalum: Germanium plays a crucial role in fiber-optic communications (data center data transmission) and high-frequency chips; tantalum is used to manufacture small-sized, high-capacity tantalum capacitors, which are essential components for stabilizing chip voltage.
Key focus stocks: $Dow Inc (DOW.US)$ 、 $Ferroglobe (GSM.US)$ 、 $DuPont (DD.US)$ 、 $Albemarle (ALB.US)$ 、 $Arcadium Lithium (ALTM.FTOLD.US)$ 、 $Teck Resources (TECK.US)$ 、 $CMOC (03993.HK)$ 。
In addition, almost all gallium comes from bauxite refining, therefore $Alcoa (AA.US)$、 $Century Aluminum (CENX.US)$、 $Constellium (CSTM.US)$ 、$CHINAHONGQIAO (01378.HK)$ 、 $CHALCO (02600.HK)$ 、 $RUSAL (00486.HK)$ is worth noting.
4. Storage Devices: The 'Treasure Chest' of Massive Data
AI model training relies on astronomical amounts of data, directly driving demand for large-capacity storage devices (such as enterprise HDDs and SSDs).
Rare earths and boron: High-performance mechanical hard drives (HDD) spindle motors and drive arms heavily depend on rare earth permanent magnet materials such as neodymium-iron-boron. Without these powerful magnetic materials, high-speed reading and writing of massive data would not be possible.
Key focus stocks: In the US stock market: $MP Materials (MP.US)$ 、 $USA Rare Earth (USAR.US)$ 、 $Energy Fuels (UUUU.US)$ ; In the Hong Kong stock market: $JLMAG (06680.HK)$ ; Global mining giants $Rio Tinto (RIO.US)$ 。
Summary
The construction of AI infrastructure is a protracted battle. From a fundamental perspective, the limit of computing power is electricity, and the limit of electricity and hardware ultimately lies in mineral resources in the physical world. For investors, in addition to tracking downstream SaaS applications and midstream hardware assembly, extending their focus to upstream resource sectors may allow them to capture more certain investment opportunities amid the resonance of this grand technological cycle with the commodities cycle.
Risk Disclaimer: The above content only represents the author's view. It does not represent any position or investment advice of Futu. Futu makes no representation or warranty.Read more
Comments (5)
to post a comment
72
142
