"AI Resources - Military Industry" Triad of Survival -- Global Asset Revaluation Under the New Debt Cycle

The market often compares this AI boom to the 2000 dot-com bubble, focusing on whether valuations are high and if a bubble will burst. Dongwu Securities’ strategic in-depth report deliberately avoids that line of discussion and instead asks a deeper question: under rising global debt pressures, increasing geopolitical frictions, and narrowing central bank policy boundaries, how should global assets be priced?

Chen Meng, a strategy analyst at Dongwu Securities, provides a compelling framework in a report dated March 2: “The world is entering a new cycle centered on ‘AI-Resources-Military Industry,’ forming a ‘survival trinity.’ AI is the ‘engine,’ resources are the ‘fuel,’ and the military industry is the ‘chassis.’” Debt pressures are forcing technological leaps; the implementation of these technologies will inevitably consume and compete for resources, and the contest over resource sovereignty will ultimately bring security and military industries to the forefront.

The starting point is the “new debt cycle.” The report emphasizes two differences: the interest rate environment and the debt foundation have changed. The 2000 dot-com bubble relied on low interest rates and fiscal space; this AI cycle began in 2022 with policy rates soaring from near 0% to over 5%, while global public debt has increased more than fourfold since 2000, and global debt has doubled compared to 2000. By the end of 2024, U.S. government debt as a percentage of GDP exceeds 120%.

The core contradiction is “r > g”—the long-term interest payments on sovereign debt exceed the natural growth rate of the economy. With demographic dividends fading (global population growth slowing) and marginal returns on capital declining, fiscal austerity or simple deleveraging cannot solve the problem. The only way out is to enlarge the denominator: use the AI revolution to boost TFP (Total Factor Productivity), so that future incremental wealth can cover the accumulated debt.

This also explains why governments worldwide are elevating AI to a national strategic level rather than leaving it solely to private enterprise. In the first half of 2025, AI investment contributed 1.42 percentage points to U.S. GDP, surpassing private consumption’s 1.06 points for the first time; in Q1 2025, AI’s contribution to GDP reached 1.3 percentage points, breaking the record set during the 1999 internet bubble. AI’s impact on the economy has shifted from “narrative” to “concrete data.”

Hard constraints like electricity, copper, and uranium will be priced first. As resource sovereignty rises, the military industry will resemble a “security tax.”

The key point is: demand for computing power is growing exponentially, transforming AI into a heavy-asset system—beyond chips, power supply, cooling, metals, and transmission capacity are all bottlenecks. Dongwu Securities believes AI competition will spill over from “algorithm contests” to a “computing power—physical infrastructure” game, with strategic importance of resource and energy commodities passively rising.

The resource contest is depicted as an ongoing action:

Mergers and acquisitions are accelerating, extending from metal minerals to oil and gas “energy bases.” Transaction volumes in metals and mineral industries peaked at $86 billion by the end of 2024; tech giants are shifting from “power purchase agreements” to “owning power generation assets,” aiming to lock in long-term stable electricity.

The report also places U.S. policy actions within the same logic: early 2026, the Trump administration launched the “Project Vault,” attempting to lock down $12 billion in strategic resources; meanwhile, in November 2025, USGS expanded the list of critical minerals from 50 to 60, adding copper, silver, silicon, uranium, and other materials, and plans to establish national strategic reserves. The conclusion is: U.S. definitions of “strategic resources” are being rewritten along the entire AI industry chain, elevating resource possession to “resource hegemony.”

The report elevates the military industry from a “theme rotation” to the endpoint of the survival trinity: resource sovereignty requires security backing, and security needs will, in turn, drive technological spillovers, changing warfare and national security boundaries.

On the data front, it supports this view with the global increase in military spending: in 2024, global military expenditure reached $2.72 trillion, a 9.4% real increase—the highest ever; military spending as a share of global GDP rose to 2.5%, and as a share of government expenditure to 7.1%. Regionally, from 2022 to 2024, U.S. military spending as a percentage of GDP increased from 3.31% to 3.42%, Europe from 1.52% to 1.90%, Japan from 1.01% to 1.37%; South Korea remains high at around 2.6% (2024: 2.56%).

Historical context: validation of the “trinity” across three industrial revolutions

The report reviews the three industrial revolutions to validate the “survival trinity” framework historically.

The First Industrial Revolution (1800s): “Debt-driven production”—Britain relied on high debt, steam engines, and coal to boost productivity, expanding through naval power and global trade networks to accumulate wealth.

The Second Industrial Revolution (1900s): “Production reshaping the world”—internal combustion engines and oil transformed industries, but resource competition eventually led to large-scale wars, with debt issues resolved through violent means.

The Third Industrial Revolution (1980s): “Networked world”—after the Bretton Woods system collapsed, the U.S. failed to rapidly boost TFP through technology, instead relying on the “Petrodollar” system and financial mechanisms to transfer risk.

These historical episodes reveal three paths to resolve massive debt: internal TFP efficiency breakthroughs, external geopolitical violent resolutions, and the financial cost-shifting by currency hegemony nations.

“Donroe Doctrine”: The market’s underestimated expectation gap

The report identifies the most significant variable for 2026 as the strategic leap of the “Donroe Doctrine.” This concept first appeared in American right-wing media in 2025, as a contemporary rewrite of the Monroe Doctrine—focusing on security and resource dominance in the Western Hemisphere as the “anchor,” and extending this to reshape global trade, capital, and industry rules.

It points out that markets generally misjudge Trump’s strategy as isolationist, when in fact it is an upgraded unilateralism: in 2025, the focus was on using tariffs and industrial policies to reverse trade deficits with an emphasis on “soft interventions”; by 2026, geopolitical actions will play a more prominent role, with core demands shifting from “trade balance” to “strategic resource control” in a “hard contest.”

There are two major pricing expectation gaps in the market: First, the “low-volatility illusion” of geopolitical risks versus the “high-volatility reality” of great power intervention—misjudging the Donroe Doctrine as isolationism will severely underestimate the upward potential of geopolitical risk premiums; second, the “long-term bearish” view on the dollar versus the “mid-term strength” anchored by resources—by 2026, with mid-term elections and inflation stabilization, a strong dollar path is more likely, serving as a defensive tool to curb inflation, attract capital, and suppress rivals.

Winning strategies: TFP leap or stagflation spiral

Using a debt sustainability model (DSA), the report quantifies the critical TFP contribution needed for AI, constructing a “interest rate–TFP” sensitivity model:

• Optimistic scenario (interest rate 3%): If the Fed can effectively lower nominal rates to 3%, AI needs only a 0.07% TFP increase to balance debt dynamics;

• Neutral scenario (interest rate 3.4%): Maintaining current rates requires at least a 0.5% TFP contribution from AI;

• Pessimistic scenario (interest rate 4%): If long-term rates stay above 4%, AI must contribute over 2.8% to TFP, demanding an exponential increase in technological revolution intensity.

Scenario A (Efficiency breakthrough): If “re-industrialization” can deeply integrate with AI and robotics, spreading through corporate organization and supply systems, the macro key will be whether TFP growth can outpace debt compounding.

Scenario B (Hardball and stagflation): If the U.S. over-relies on “Donroe” unilateral tools, wielding “commodity blockades” and “financial sanctions,” it may force resource countries to accelerate building settlement systems, leading to structural high inflation, major holders dumping U.S. debt, soaring financing costs, a collapse in dollar reserve credibility, and a sharp revaluation of global assets.

Global asset allocation: Focus on “physical bottlenecks”

The report’s straightforward approach is to focus on physical bottlenecks and scarcity points within the “survival trinity” system, rather than traditional developed/emerging or growth/value classifications.

Key focus areas include:

• AI: core chips and optical modules; critical cooling systems; copper connections; AI PCs/phones, humanoid robots, and embodied intelligence.
• Resources: bottlenecks in power grid transmission (transformers, high-voltage cables); “industrial blood” copper; nuclear and uranium resources; natural gas power generation and turbines; rare earths and small metals like tungsten, antimony, gallium.
• Military industry: ammunition materials and shipbuilding/maintenance (capacity scarcity, consumables); commercial space and low-earth orbit satellites (central to new domain warfare and major arteries).

The report also lists four risk categories: AI productivity shortfalls, tech giants’ capex peaking and hardware valuations collapsing, over-crowded core ecosystems leading to expectation exhaustion, and geopolitical hardball spiraling into more severe stagflation. For this framework, the most dangerous scenario is not short-term sector fluctuations but the failure of TFP to accelerate while conflicts intensify.

Risk warnings and disclaimers

Market risks are present; investments should be cautious. This article does not constitute personal investment advice and does not consider individual users’ specific investment goals, financial situations, or needs. Users should evaluate whether any opinions, viewpoints, or conclusions herein are suitable for their circumstances. Investment is at their own risk.