Futures
Access hundreds of perpetual contracts
CFD
Gold
One platform for global traditional assets
Options
Hot
Trade European-style vanilla options
Unified Account
Maximize your capital efficiency
Demo Trading
Introduction to Futures Trading
Learn the basics of futures trading
Futures Events
Join events to earn rewards
Demo Trading
Use virtual funds to practice risk-free trading
CFD
U.S. stock CFD derivatives
US Stocks
Access real US stocks and ETFs
HK Stocks
Trade quality Hong Kong-listed stocks
Korean Stocks
SK Hynix
Real Korean stocks and top assets
Stock Futures
High leverage, 24/7 trading
Tokenized Stocks
Backed by real stock assets
IPO Access
Unlock full access to global stock IPOs
GUSD
Mint GUSD for Treasury RWA yields
Stocks Activities
Trade Popular Stocks and Unlock Generous Airdrops
Launch
CandyDrop
Collect candies to earn airdrops
Launchpool
Quick staking, earn potential new tokens
HODLer Airdrop
Hold GT and get massive airdrops for free
IPO Access
Unlock full access to global stock IPOs
Alpha Points
Trade on-chain assets and earn airdrops
Futures Points
Earn futures points and claim airdrop rewards
Promotions
AI
Gate AI
Your all-in-one conversational AI partner
Gate AI Bot
Use Gate AI directly in your social App
GateClaw
Gate Blue Lobster, ready to go
Gate for AI Agent
AI infrastructure, Gate MCP, Skills, and CLI
Gate Skills Hub
10K+ Skills
From office tasks to trading, the all-in-one skill hub makes AI even more useful.
AI Competition Opens Rare Metal Super Cycle: Tin, Indium, Hafnium Facing Historic Revaluation Moment?
AI hardware investment is expanding from "buying chips" to a longer chain: server systems, high-speed network components, data center power supply, and high-density cooling facilities are all starting to consume more basic raw materials. For minor metals, the key change is not just the addition of a new concept, but that downstream usage is beginning to enter a stage where it can be calculated.
In a research report on June 20, Dongwu Securities analyst Liu Yiting stated: "Global AI capital expenditure is entering a nonlinear acceleration phase," with investment gradually covering not just single chips but also servers, high-speed networks, power infrastructure, and cooling facilities, "bringing demand dividends to upstream basic raw materials."
Among them, tin, indium, and hafnium correspond to three bottlenecks in AI hardware upgrades: tin is used in PCB electroplating and SMT soldering; indium, in the form of indium phosphide, enters high-speed optical communications; hafnium serves as a high-K gate dielectric material, enabling the continued miniaturization of advanced processes. In the projections, PCB-side tin consumption could increase by 49k tons from 2026 to 2030; the indium demand for indium phosphide in AI data centers could rise from 19 tons in 2025 to 419 tons in 2030; global hafnium demand could grow from 100 tons in 2024 to 142 tons in 2030.
The commonality among these three metals lies in the supply side: tin is affected by resource depletion, Indonesian policies, the slower-than-expected recovery of Myanmar's mine production, and shifts in trade flows; indium is constrained by zinc mining and smelting operations; hafnium is bottlenecked by zirconium-hafnium separation, environmental concerns, economic feasibility, and geopolitical disruptions. The core logic behind the upward shift in price levels is the combination of rising demand and supply constraints.
AI Hardware Spending Is No Longer Limited to GPUs
Capital expenditure is a leading indicator for this chain. In 2026, the combined capital expenditure of the four major cloud giants—Microsoft, Google, Amazon, and Meta—could reach up to $725 billion. From January to September 2025, U.S. AI-related investments contributed 39% to real GDP growth, higher than the 36% during the Internet bubble in 2000.
Hardware upgrades are concentrated in four areas: computing density, memory bandwidth, interconnect speed, and power efficiency. Chips are just one part of the equation. The number of layers in AI server PCBs has increased from the traditional 8-24 layers to typically 28-46 layers, with some projects even adopting 56-layer designs. High-speed optical modules are evolving from 800G to 1.6T and 3.2T, making the bottleneck in data center internal interconnections increasingly prominent. As advanced processes continue, traditional silicon dioxide gate dielectrics are approaching their physical limits.
Minor metals have entered the spotlight not because of their scarcity alone, but because they precisely address these upgrade points.
The New Incremental Demand for Tin Lies in PCBs, but Supply Struggles to Keep Up
Tin plays a soldering and connecting role in the electronics industry. The expansion of AI servers, high-end PCBs, and advanced packaging will all increase tin consumption.
The projection breaks down tin consumption into two parts: electroplating tin used in PCB manufacturing and SMT assembly tin. In PCB electroplating, the unit consumption of tin for HDI boards is approximately 40.19 g/m², while for multi-layer boards it is about 12.84 g/m²—meaning HDI's unit consumption is over three times that of multi-layer boards. For SMT assembly, the unit consumption of tin is about 294.22 g/m². Combined, the unit consumption of tin for PCB electroplating and SMT is approximately 318 g/m².
According to Prismark's forecasts, global PCB shipments will reach 663 million square meters by 2030, with a compound annual growth rate of about 6.7% from 2026 to 2030. Correspondingly, the projection shows that global PCB tin consumption will increase from 163k tons in 2026 to 212k tons in 2030, an increase of 49k tons over four years, with a CAGR of 6.9%. Based on the global tin consumption of 380k tons in 2025, the elasticity of PCB-side tin consumption is 12.3%.
The problem lies in supply.
Global proven tin reserves are approximately 6 million tons, with a static reserve-to-production ratio of about 20.7 years, lower than industrial metals like copper, nickel, and cobalt. From 2015 to 2025, despite a significant rise in tin prices, global tin mine production only increased from 289k tons to 290k tons—almost zero growth over a decade. China's tin mine production fell from 110k tons to 71k tons, a CAGR of -4.3%.
Indonesia is a major variable. In 2025, Indonesia's tin mine production accounted for 21% of the global total, but in recent years, frequent policy adjustments—including mining permits, illegal mining management, progressive royalties, and minimum benchmark prices—have caused large fluctuations in exports. Myanmar was once a significant supplier, producing 17% of global tin in 2018, but after resource depletion and a mining ban, its production fell to 12k tons in 2025. Even after the Wa State announced a resumption of mining in the second half of 2025, China's tin ore imports from Myanmar had only recovered to about 1,300 metal tons by April 2026, still below the pre-ban level of about 2,200 tons per month.
Trade flows in South America are also changing. Peru, Brazil, and Bolivia produced a combined 76k tons of tin in 2025, accounting for 26% of the global total. Among them, the top destination for Peru's tin ingot exports is the United States, while Bolivia's exports mainly go to the Netherlands, the United Kingdom, and the United States. The accelerated development of the U.S. tin supply chain may further absorb South American raw materials.
Overall, Dongwu Securities believes that tin metal will face both high demand growth and supply disruptions over the next three to four years, creating a strong price increase driver. On one hand, the acceleration of global AI capital expenditure and the expansion of hardware devices such as PCBs are expected to bring real incremental demand for tin. On the other hand, global tin supply is highly concentrated, unstable, and influenced by numerous factors.
Indium's Elasticity Comes from Indium Phosphide, but Production Cannot Be Easily Expanded
Indium's traditional demand is mainly for ITO targets, accounting for about 70%, used downstream in liquid crystal displays and flat screens; electronic semiconductors, solders, and alloys each account for about 12%. In 2025, global refined indium consumption was 2,316 tons, with projections of 2,510 tons in 2026 and 2,813 tons in 2027.
The new variable is optical communications. Inside AI data centers, GPUs need high-speed data exchange. In a model cluster with tens of thousands of cards, the energy consumed by data movement between chips accounts for over 90% of the system's total energy. As data transfer rates increase, the effective transmission distance of copper interconnects shrinks to just a few centimeters. Data transfer rates have upgraded from 100G/lane to 200G/lane and are continuing toward 400G/lane, making optical interconnects a more realistic direction.
The advantage of indium phosphide is clear: it is a direct bandgap semiconductor with a bandgap energy of about 1.34 eV, matching the low-loss windows of 1310nm/1550nm in fiber optic communications; its electron mobility is more than 10 times that of silicon, supporting high-frequency modulation above 100 GHz. In high-speed optical modules, indium phosphide is the core material for laser chips.
In the projection, a 4-inch indium phosphide substrate consumes approximately 32.2 grams of indium per wafer. In 2025, AI data centers are expected to demand about 600k such substrates, corresponding to 19.3 tons of indium; by 2030, indium phosphide demand could reach 13 million substrates, corresponding to 419 tons of indium—a growth of more than 22 times. Based on 2025 global indium demand, this single factor alone could bring over 20% incremental demand.
The hard constraint on supply is that indium is primarily associated with lead-zinc polymetallic deposits. About 81.2% of global indium reserves come from such deposits; primary indium mainly comes from residues of zinc processing. In other words, even if indium prices rise, one cannot simply open a standalone "indium mine" to rapidly increase production.
In recent years, zinc concentrate processing fees have declined, reducing the willingness of zinc smelters to operate. The utilization rate of refined zinc capacity has fallen to a low for the same period in recent years, constraining primary indium supply. Meanwhile, China imposed export controls on indium phosphide, trimethylindium, triethylindium, and related technical data in February 2025. Inventories are also declining: according to data from the China United Metal Trading Platform, indium inventories dropped from about 488.8 tons at the start of 2025 to 273.8 tons as of January 28, 2026.
As of June 11, 2026, domestic refined indium prices were 4.7 million yuan per ton, up 58% from the beginning of the year.
Hafnium's Value Lies in Advanced Processes, with Challenges in Separation and Expansion Economics
Hafnium's traditional demand is concentrated in nuclear energy and superalloys. In the consumption structure, nuclear energy accounts for 45%, superalloys/aerospace 35%, and semiconductors/electronics 10%.
The change on the semiconductor side comes from process miniaturization. At nodes of 65nm and below, the excessive thinness of traditional silicon dioxide gate dielectrics leads to quantum tunneling effects, causing increased gate leakage current and putting chip power consumption and reliability at risk. Hafnium oxide has a dielectric constant of about 18-25, far higher than silicon dioxide's 3.9. This allows for an increase in physical thickness while maintaining the equivalent oxide thickness, thereby reducing leakage.
After Intel introduced hafnium-based high-K materials to replace silicon dioxide gate dielectrics in its 45nm process, the gate leakage current for NMOS was reduced by more than 25 times, and for PMOS by more than 1,000 times. As 3nm and 2nm nodes transition from FinFET to GAA architectures, the demand for high-K dielectrics will continue to rise.
In the demand path, global hafnium demand is expected to increase from 100 tons in 2024 to 142 tons in 2030. Semiconductor demand will grow from 40 tons to 64 tons, contributing nearly half of the increase; superalloy demand from 45 tons to 60 tons; and nuclear energy demand from 15 tons to 18 tons.
Hafnium's supply is more problematic than its demand. Hafnium is mainly a byproduct of the production of nuclear-grade sponge zirconium. Global capacity for nuclear-grade sponge zirconium exceeds 10k tons per year, with actual annual production of 6,000-7,000 tons, corresponding to about 100 tons of sponge hafnium. Major producers include the United States, France, Russia, and China.
Zirconium-hafnium separation is highly difficult. The two elements have very similar physical and chemical properties, and hafnium in nature and in zirconium chemicals typically accounts for only 1% to 3% of the total zirconium-hafnium content. Existing processes involve toxic solvents or high-concentration acids, posing environmental and equipment corrosion challenges. Expansion is also uneconomical: two U.S. producers could theoretically increase hafnium output by about 100%, but each company would generate an additional approximately 2,000 tons per year of de-hafniated zirconium, which, without customers to absorb it, makes the expansion difficult to close.
Geopolitical disruptions have further pushed up prices. After the Russia-Ukraine conflict in 2022, Russian sponge hafnium supplies were cut off, causing international hafnium prices to surge from $1,200-1,400/kg to $4,500-5,000/kg. At the end of 2024, China included hafnium in its dual-use items management. In 2025, exports of unwrought hafnium, hafnium waste and scrap, and hafnium powder totaled 20.2 tons, a year-on-year decrease of 22%.
Domestic prices of 4N-grade hafnium oxide have also risen sharply. From about 4.5 million yuan per ton in early 2022, it rose to 9.5 million yuan per ton as of June 16, 2026, an increase of 111%.
Risk Warning and Disclaimer