"AI-Resources-Military Industry" Survival Trinity -- Global Asset Revaluation in the New Debt Cycle
```
The market always likes to compare this wave of AI enthusiasm to the internet bubble of 2000, often discussing whether the valuations are too high or if the bubble will burst. This in-depth strategy report from Dongwu Securities deliberately avoids that line and instead asks a more fundamental question: In the context of rising global debt pressure, increased geopolitical friction, and narrowing central bank policy boundaries, what logic should really be used to price global assets?
Dongwu Securities strategy analyst Chen Meng presents a strong framework in the report dated March 2: The world is currently entering a new cycle centered on "AI-resources-defense," a "survival trinity," where AI is the "engine," resources are the "fuel," and defense is the "chassis". Debt pressure forces technological leaps, technology implementation inevitably consumes and competes for resources, and resource sovereignty competition ultimately brings security and defense to the forefront.
The starting point of the report is the "new debt cycle." It emphasizes that this AI boom is not growing in a low interest rate environment, but is unfolding in a pressure test of a Federal Reserve rate above 5% and global debt more than doubled since 2000. The reason why countries are betting on AI is because it is seen as the "technological cure" to dilute debt—by expanding total factor productivity (TFP), growing the denominator, rather than relying on traditional "quantity expansion" methods to extend life.
Following this logic, asset allocation is also forced to shift frameworks: no longer the old classification of "developed vs emerging" or "growth vs value," but focusing on who occupies irreplaceable ecological positions within the "AI-resources-defense" system. The report also puts a critical variable for 2026 on "Donroe Doctrine": Trump’s ambitions shifting from trade balance to strategic resource control, and the market might be underestimating the aggressiveness of the geopolitical and dollar path—which could in turn change the way risk premiums and pricing for commodities, defense spending, and supply chains are set.
AI as the "macro must-have" in a high debt era
The report first addresses the structural dilemma of macro debt, revealing the deep logic of this cycle.
It emphasizes two differences: the interest rate environment and the debt base have changed. The 2000 internet bubble benefited from low rates and fiscal space; this AI boom began with policy rates rising sharply from near 0% in 2022 to above 5%, while global public debt has increased more than fourfold since 2000, and total global debt has doubled. By the end of 2024, US government debt will account for more than 120% of GDP.
The core contradiction is "r > g"—the interest expense on sovereign debt long exceeding economic natural growth rate. With the decline of the demographic dividend (global population growth is slowing) and diminishing marginal capital returns, fiscal tightening or simple deleveraging cannot solve the problem. The only way out is to expand the denominator: that is, to enhance TFP through the AI revolution, using future incremental wealth to cover historical stock debt.
This explains why governments worldwide have elevated AI to a national strategy and are fully advancing it, rather than leaving it to corporate voluntary action. In the first half of 2025, AI investment contributed 1.42 percentage points to US GDP growth, surpassing private consumption’s 1.06 points for the first time; in Q1 2025, AI investment’s GDP contribution rate reached 1.3 percentage points, breaking the record set in the internet bubble of 1999. The economic boosting effect of AI is shifting from "narrative" to "manifest data."
Electricity, copper, uranium and other "hard constraints" will be priced first; after resource sovereignty climbs, defense will act as a "security tax"
A key point in the report is: exponential growth in computing demand turns AI into a heavy-asset system—beyond chips, power supply, cooling, metals, and transmission capacity all become bottlenecks. Dongwu Securities judges that AI competition will spill over from "algorithm contests" to a game of "computing power — physical base," passively elevating the strategic status of resources and energy.
Resource competition is described as an ongoing action:
M&A is accelerating, extending from metal mining to oil and gas "energy base." By the end of 2024, transaction volume in metals and mining reached $86 billion, a cyclical peak; tech giants are moving from "power purchase agreements" to "owning power assets," aiming to lock in stable long-term electricity.
The report also incorporates US policy moves: early 2026, Trump’s administration launches "Project Vault," attempting to lock up $12 billion in strategic resources; at the same time, USGS expands its list of critical minerals from 50 to 60 by November 2025, newly including copper, silver, silicon, uranium, and plans to build a national strategic reserve. The conclusion is: The US definition of "strategic resources" is being rewritten in line with the entire AI industry chain, and resource possession is being upgraded to "resource hegemony".
The report takes defense out of the usual "theme rotation" and portrays it as the endpoint of the survival trinity: resource sovereignty demands a security guarantee, and security needs in turn drive technology spillover, changing the form of warfare and national security boundaries.
On the data side, it uses the global rise in defense spending to support the argument: in 2024, total global defense spending hit $2.72 trillion, with a real growth rate of 9.4%, a new historical high; defense spending as a proportion of global GDP rose to 2.5%, and as a share of government spending to 7.1%. Regionally, from 2022 to 2024, US defense/GDP went from 3.31% up to 3.42%, Europe from 1.52% to 1.90%, Japan from 1.01% to 1.37%; Korea fluctuates at a high level near 2.6% (2.56% in 2024).
Historical rhythm: "Trinity" validation in three industrial revolutions
The report reviews three industrial revolutions to underpin the survival trinity framework with historical validity.
The First Industrial Revolution (1800s) "debt to production": Britain, under high debt pressure, relied on the steam engine and coal to drive productivity leaps, and achieved wealth accumulation through naval expansion and global trade networks.
The Second Industrial Revolution (1900s) "production splitting the world": Internal combustion engines and oil reshaped industry, but resource competition eventually evolved into large-scale war, and debt problems were solved violently.
The Third Industrial Revolution (1980s) "rule by networks": After the Bretton Woods system collapsed, the US did not achieve a rapid TFP leap through technology, but shifted risk via the "petrodollar" system and financial tools.
These three episodes reveal three ways to resolve super-debt: internal TFP efficiency breakthrough, external geopolitical violent clearing, and financial cost transfer by currency hegemon nations.
"Donroe Doctrine": the biggest underestimated market variable
The report identifies the most unavoidable variable for 2026 as the strategic leap of the "Donroe Doctrine." First appearing in US right-wing media in 2025, it is Trump’s contemporary rewrite of the "Monroe Doctrine"—using the security and resource dominance of the Western Hemisphere as "anchors," and spilling over into reshaping global trade, capital, and industry rules.
The report points out that the market generally misjudges Trump's strategy as isolationism, when in fact it is an upgraded version of unilateralism: in 2025, focusing on tariffs and industrial policy to reverse trade deficits, preferring "soft intervention;" whereas in 2026, the weight of geopolitical action rises sharply, with core demands shifting from "trade balance" to a "hard game" of strategic resource control worldwide.
The market has two major mispricing differences: First, the "low-vol illusion" of geopolitical risk vs. the "high-vol reality" of power intervention—misjudging "Donroe Doctrine" as isolationism will seriously underestimate the upward elasticity of geopolitical risk premium; Second, "long-term bearish" on dollar credit vs. "medium-term strength" anchored by resources—as the midterm election constraints intensify and demand for stable inflation grows in 2026, a strong dollar path is more likely to return, becoming a defensive choice to suppress inflation, attract funds, and pressure adversaries.
Decisive scenarios: TFP leap or stagflation spiral?
The report, based on Debt Sustainability Analysis (DSA), quantitatively estimates the critical TFP value AI must deliver, constructing an "interest rate-TFP" sensitivity model:
- Optimistic scenario (rate 3%): If the Fed brings nominal rates down to 3%, AI needs only a 0.07% TFP increment to achieve dynamic debt balance;
- Neutral scenario (rate 3.4%): Maintaining current rates, AI must deliver at least 0.5% TFP growth;
- Pessimistic scenario (rate 4%): If long-term rates stay above 4%, AI needs to contribute more than 2.8% to TFP increment, demanding a much stronger technological revolution.
Scenario A (efficiency breakthrough): If "reindustrialization" deeply integrates AI and robotics, and diffuses through organizational and supply systems, macro success hinges on whether TFP increment outpaces debt compounding.
Scenario B (hard game & stagflation): If the US excessively relies on unilateral "Donroe Doctrine" tools and wields both "commodity blockade" and "financial sanctions," it could force resource countries to accelerate settlement systems, causing structural high inflation in the US, major holders selling US debt, skyrocketing financing costs, dollar reserve credibility collapse, and violent repricing of global assets.
Global asset allocation: Go long "physical bottlenecks"
The report’s allocation plan is direct: track physical bottlenecks and scarce links within the survival trinity, not dividing along developed/emerging, growth/value lines.
Areas of focus are listed in three areas:
AI: core chips and optical modules; liquid cooling as a "key safeguard system"; copper connections; terminal/embodied intelligence like AI PC/Phone and humanoid robots.Resources: bottlenecks in power grid transmission (transformers, high-voltage cables); copper—the "blood of industry"; nuclear power and uranium resources; gas power generation & turbine equipment; rare earths and small metals like tungsten/antimony/gallium, crossing tech and defense logic.Defense: ammunition materials & shipbuilding repair (capacity shortage, consumable nature); commercial spaceflight & low-orbit satellites (center and main artery of new domain warfare).
The report also lists four types of risks: AI productivity leap not meeting expectations; tech giants’ capital expenditures peaking early and causing hardware valuations to collapse; excessive crowding in key ecological positions leading to expectation overshoot; and uncontrolled hard geopolitical games triggering severe stagflation. For this framework, what’s most fatal is not short-term sector ups and downs, but “TFP fails to materialize, yet conflicts accelerate first.”
Risk warning and disclaimerThe market is risky and investment requires caution. This article does not constitute personal investment advice and does not consider individual users' specific investment goals, financial circumstances, or needs. Users should consider whether any opinions, views, or conclusions herein fit their particular circumstances. Acting on this is at your own risk. ```