Over the past decade, global financial innovation has revolved almost entirely around "digitization": internet platforms, ETFs, stablecoins, and today's RWA (Real World Asset) tokenization. However, with the rapid development of AI, a deeper question is beginning to emerge: what will truly underpin the AI era?
The answer may not be the code itself, but the real world behind the code.
AI is Redefining the Asset Layer Structure
For a long time, AI was often imagined as a "dematerializing" technological revolution, but the reality is precisely the opposite. AI has not weakened dependence on the real world; instead, it is strengthening that dependence. Every model training, inference, and deployment requires vast real-world infrastructure support, including data centers, power grids, cooling facilities, advanced hardware, and industrial resources. In other words, what appears to be a digital system is essentially an industrial system.
In this system, the truly irreplaceable elements are metals and real-world resources. Materials like copper, silver, and gold collectively determine conductivity, durability, and infrastructure performance—capabilities that software cannot replace.
S&P Global expects that copper demand from data centers alone will grow from 1.1 million metric tons in 2025 to 2.5 million metric tons by 2040. Concurrently, the market expects the global refined copper deficit to reach 304,000 metric tons in 2025 and could expand to 6 million metric tons by 2035. As AI infrastructure continues to expand, demand for these real-world resources is growing rapidly, while supply remains constrained by structural factors.
A growing number of industry observers are beginning to see this not as a short-term cyclical issue, but as a long-term structural change. What may truly limit AI expansion is no longer just computing power itself, but the "Physical Layer" constituted by energy, metals, and real-world infrastructure. This layer is also developing its own logic of scarcity, pricing, and asset system.
A New "Asset Stack" Structure is Forming
In this context, the market is beginning to re-understand the relationship between the physical layer, the financial layer, and the digital layer:
-
Physical Layer: Metals, energy, real-world resources
-
Financial Layer: Government bonds, ETFs, structured products
Digital Layer: Tokenization infrastructure, programmable assets
The digital layer is built upon the financial layer, which ultimately depends on the real-world physical layer. Over the past few decades, the market has long highly rewarded "upper-layer assets," including stocks, ETFs, internet platforms, and digital financial infrastructure. But now, AI is pulling market attention back to the foundational real-world resources themselves.
Tokenization Does Not Create Value Out of Thin Air
This also explains why most RWA projects haven't truly succeeded. The problem isn't entirely in the technology itself, but in asset selection.
Tokenization does not create value out of thin air; it merely reconnects assets the market already trusts. For an asset to be genuinely tokenized, it typically needs mature demand, deep liquidity, and institutional consensus. Otherwise, tokenization often brings complexity, not value.
From this perspective, the current development path of tokenization is quite logical. The first to be tokenized was sovereign debt because it has the world's most mature liquidity and credit system. Next came gold, which has centuries of global consensus. After that came silver, possessing both reserve attributes and industrial demand. And the direction for future expansion is likely the industrial materials that the real economy truly relies on.
Notably, the order of tokenization does not entirely depend on these assets' importance to AI infrastructure. The importance of copper and industrial metals is no less than that of gold. What truly determines the order is where market consensus is first established, with each step inheriting the credibility accumulated by the previous one.
This is also the core logic behind the RWA tokenization platform Matrixdock: start with assets where the market has established long-term trust, including sovereign debt, gold, and silver. Currently, Matrixdock manages over $200 million in on-chain assets and serves institutional clients who need both the stability of real-world assets and the programmability of on-chain infrastructure.
Gold ETFs and Gold Tokens are Heading in Different Directions
In the gold sector, a new change is also emerging.
Gold ETFs were one of the most successful financial innovations of the past two decades. They solved the issues of physical gold being difficult to trade, lacking liquidity, and having high holding costs, allowing gold to be easily bought and sold by ordinary investors like stocks for the first time.
But the core logic of ETFs is essentially to give investors "exposure to gold," not to truly bring gold into the financial system. The gold in ETFs essentially remains within the traditional financial holding system, difficult to achieve programmable settlement, native collateral, or cross-system interaction like on-chain assets.
With the development of programmable finance and on-chain finance, the market is beginning to ask a new question: besides "holding," can gold truly participate in financial activities? For example, can it achieve instant settlement, cross-border collateral, and flow without requiring a custodial intermediary?
In a sense, this is the fundamental difference between gold tokens and gold ETFs. Gold ETFs solve the problem of making gold "investable"; gold tokens, however, are exploring broader functions for gold within the digital financial system.
Matrixdock's gold token, XAUm, operates precisely on this logic. Currently, XAUm has approximately $74 million in assets under management (AUM) and cumulative trading volume exceeding $100 million. Its goal is not to simply replicate an ETF but to begin bringing gold into the on-chain financial system.
From "Store of Value" to "Functional Asset"
And gold may just be the starting point.
As AI infrastructure continues to expand, more and more industrial materials are shifting from "commodities" to "strategic resources." Silver is used for conductivity, copper underpins energy and connectivity infrastructure, and industrial metals are becoming the real-world physical substrate behind AI infrastructure.
Silver, in particular, has already begun to see changes in its supply-demand structure. It has faced a structural supply deficit for five consecutive years, and the deficit is projected to widen to 46.3 million ounces by 2026. Industrial demand from solar power, electric vehicles, and AI infrastructure is continuously pushing up consumption, while growth in mining supply struggles to keep pace.
If gold represents a "Store of Value," then industrial metals are more like a "Store of Function." However, the tokenization of industrial metals will not completely replicate the path of gold. This is because industrial metals are consumed; their focus is not solely on reserve attributes but on how to establish operational and circulatory connections between the real commodity system and digital infrastructure.
Matrixdock's silver token, XAGm, is the first step in this direction. Its positioning is to connect the reserve logic of precious metals with the functional demand of industrial metals. As the roadmap delves further into the "Physical Layer," the direction is becoming increasingly clear: those industrial metals upon which AI infrastructure heavily relies may become an important component of the next stage of the on-chain asset system.
In a sense, the asset layer is evolving towards a direction that is more firmly grounded in the real, physical world, more strategic, and simultaneously more programmable. And the assets truly worth tokenizing in the future may not just be those "easiest to digitize," but those important assets upon which the real economy has long depended.
