An Evidence-Based Look at the Strategic Asset Allocation Driving the AI Infrastructure Super-Cycle 

Despite headlines about de‑dollarization and capital outflows from the USA, our AI model recommends a clear overweight to U.S. assets. The signal is rooted in fundamentals: multi‑gigawatt AI data‑center build‑outs, superior access to frontier accelerators, deep capital markets, and supportive industrial policy. The United States is not only larger - its compute, power and capital stacks are compounding faster, positioning it to capture a disproportionate share of AI‑driven productivity over the next 20 -30 years. 

While public discussion swirls around the latest AI models and software applications, a more fundamental story is unfolding in the physical world.  

The artificial intelligence revolution isn't just code - it's a colossal undertaking of physical infrastructure - data centers, power grids, and advanced hardware. A new analysis of this landscape suggests that beyond the hype, the United States is building a surprising and compounding lead in the core infrastructure that powers AI. The findings point to a structural advantage rooted not just in technology, but in the sheer scale, speed, and economic alignment of its build-out. 

The U.S. Isn't Just Leading - It's Building at an Unmatched Scale

The primary U.S. advantage lies in the immense physical scale and speed of its AI data-center construction. The nation is pioneering "multi-gigawatt AI data-center buildouts," an industrial-level expansion that dwarfs efforts elsewhere. 

Flagship U.S. projects like "Stargate" illustrate this gigawatt-class scaling. While other regions are growing, they remain behind this frontier. For comparison, the European Union has a material pipeline of approximately 35-36 GW and the UK has 3.3-6.3 GW planned, but their pace is slowed by fundamental constraints like grid access and permitting. This means the U.S. isn't just building bigger; it's executing faster, widening the gap in foundational capacity. 

The Foundational Economic Shift: "Compute is the New Capital" 

This physical build-out is driven by a fundamental reframing of economic power for the 21st century. The analysis operates on a counter-intuitive but critical principle: the ability to scale computational power cheaply and quickly is the primary determinant of who will capture future wealth. 

This idea shifts the focus from traditional capital to computational capacity as the core engine of productivity and value creation.

The central argument is stated clearly: Compute is the new capital stock. Whoever scales compute the cheapest and fastest captures more future cash flows.

The U.S. Dominates the "Three Flywheels" of AI Infrastructure

The American lead is not accidental; it's the result of a virtuous cycle powered by three reinforcing advantages, or "flywheels": chips, clouds, and capital. 

U.S. leadership in frontier accelerators (chips), its established cloud computing giants (clouds), and its deep, accessible capital markets create a self-reinforcing loop. Each element strengthens the others, accelerating development and lowering costs. This cycle is further amplified by supportive government policy, with key tailwinds including CHIPS-era incentives, federal R&D, and sustained defense demand. This creates powerful network effects, where dense ecosystems of GPUs, developers, and models establish a widening competitive moat for the United States. 

A Compounding Lead for Decades to Come? 

The evidence suggests the U.S. lead in AI is not just a software story but is deeply rooted in a faster, more profound scaling of the physical inputs: compute, power, and capital. While markets in Europe and APAC hubs like Singapore and Malaysia will grow strongly, the U.S. is compounding its advantages in the foundational layers of the AI economy. 

This analysis concludes that the U.S. is positioned to "capture a disproportionate share of AI-driven productivity over the next 20–30 years." As this infrastructure gap widens, is the race for AI dominance already being decided not by algorithms, but by gigawatts and access to capital? 

Our AI model’s overweighting of U.S. assets is evidence-based. The U.S. is scaling the essential inputs to AI productivity - compute, power, and capital - faster and deeper than its peers. While Europe and APAC will grow and deserve specific exposure, the dominant alpha frontier for the AI Super-Cycle remains firmly U.S.-centric.