Executive Summary
According to the available reporting, historian Stephen Kotkin argued at Stanford University that computing power, rather than broad economic sanctions, is the most meaningful geopolitical lever the United States currently holds over China. In that framing, the strategic importance of Nvidia extends beyond corporate execution: the company sits near the center of how advanced AI compute is designed, allocated, and controlled.
That argument matters because it shifts attention away from trade-policy instruments alone and toward the harder question of who controls access to frontier computing capacity. For technology and capital markets, this is a more structural lens on US-China competition. It brings semiconductors, AI infrastructure, and export controls into a single strategic frame.
For Asia, the implications are particularly relevant. The region remains deeply embedded in semiconductor manufacturing, packaging, assembly, and the buildout of AI-related infrastructure. If compute is becoming a primary instrument of state power, then Asia is not just adjacent to the story. It is one of the main operational arenas where this competition is expressed.
Watch the Short Brief
Watch this short visual briefing for the key strategic implications behind the story.
Key Developments
According to the report, Kotkin’s core argument is that computing power now matters more than broad sanctions as a source of US leverage over China. The significance of that view is not merely rhetorical. It suggests that access to advanced AI chips and related infrastructure may carry more strategic weight than wider trade pressure.
The same reporting places Nvidia at the center of that discussion. The company’s importance comes from its role in advanced AI computing hardware, which makes its commercial decisions relevant to a much wider geopolitical debate. The available information does not identify a specific Nvidia product decision, export outcome, or policy intervention tied to Kotkin’s remarks, so the strategic point should be understood as a framing rather than a report of a discrete corporate action.
The draft also referenced a claim, attributed to the cited discussion, that the United States controls roughly 75% to 80% of global computing power. That figure is notable but should be treated cautiously. The available information does not independently establish the methodology, timeframe, or exact definition of “computing power” behind the estimate. Even so, the claim helps clarify the broader point being made: current AI capability may be highly concentrated.
Geographically, the issue centers on the United States and China, but the consequences reach into Asia’s semiconductor and infrastructure base. Any shift toward compute as a strategic lever naturally raises the importance of chip production capacity, packaging, supply-chain resilience, and regional access to high-end AI systems.
Strategic Analysis
The real value of Kotkin’s argument is that it reframes the mechanism of power in the US-China technology contest. Traditional sanctions and tariffs work by increasing friction. They can raise costs, slow activity, and create political pressure, but they are often negotiated around, rerouted, or partly absorbed over time. Compute concentration is different. It is tied to physical capacity, technical know-how, and control over scarce systems that are difficult to replicate quickly.
That distinction matters in semiconductors. Advanced AI hardware is not interchangeable with general industrial output. It depends on highly specialized design capability, advanced manufacturing, packaging, software ecosystems, and deployment at scale in data-center environments. When those elements are concentrated, the resulting leverage may be more durable than conventional trade restrictions.
In that context, Nvidia’s importance is not simply that it is a large chip company. It is that advanced AI compute increasingly looks like strategic infrastructure. If access to the most capable AI systems depends on a narrow set of chips and architectures, then suppliers of those systems are operating in a space where commercial strategy and state strategy can overlap. That does not make Nvidia a policymaker, but it does mean its product positioning, market access, and compliance posture can carry geopolitical significance.
This is where Asia becomes central. The region hosts many of the most important operational links in the semiconductor value chain, including advanced manufacturing, packaging, testing, and growing AI infrastructure deployment. If policy and markets begin to treat compute as a core strategic asset, then Asia’s role becomes even more consequential. The issue is no longer only whether components move across borders. It is whether countries and firms can secure reliable access to the computing capacity needed for AI development and deployment.
One implication is that the language of “compute sovereignty” may become more relevant across the region. That does not necessarily mean full self-sufficiency, which remains difficult in semiconductors. More realistically, it could mean a greater push for diversified sourcing, domestic capability development, closer coordination between industrial policy and data-center investment, and more attention to where AI infrastructure is physically located.
The reported 75% to 80% US share of global computing power, if it were substantiated by broader evidence, would reinforce this view of concentration. But even without treating that number as established fact, the underlying issue still stands: frontier AI capacity appears to be unevenly distributed, and that asymmetry has strategic consequences. For investors and operators, the exact number matters less than whether the concentration remains durable.
That durability should not be assumed. Compute leadership can persist for long periods, but it is not permanent by definition. Capacity can expand. Alternate architectures can mature. Competitors can channel large amounts of capital into domestic capability. Policy itself can also reshape market structure, either by tightening constraints or by encouraging workarounds and substitution. In other words, compute may be a powerful lever today, but its strength depends on whether the current bottlenecks remain intact.
There is also a broader analytical caution. A compute-centric view explains a great deal about the present moment, especially in AI. But it should not be mistaken for a complete description of US-China competition. Manufacturing depth, software ecosystems, energy availability, talent concentration, and capital intensity all still matter. Compute is best understood as a central strategic node within a larger system, not as the only variable.
Investor Takeaway
For investors, the most useful way to read this development is as a structural lens rather than a narrow company story. According to the report, Kotkin’s argument suggests that control over advanced computing capacity may be more important than broad sanctions in shaping the next phase of US-China rivalry. If that framing gains traction, it could influence policy design, corporate strategy, and capital allocation across the semiconductor stack.
Several areas merit close attention.
First, investors should monitor whether export controls become more tightly focused on computing capability rather than broader categories of trade. That would be consistent with a world in which policymakers see frontier compute as a core strategic asset.
Second, Nvidia remains an important barometer. Not because the available reporting confirms a specific new move by the company, but because its market position makes its compliance posture, geographic exposure, and product segmentation strategically relevant indicators.
Third, Asia’s response may be as important as Washington’s policy direction. If governments and major enterprises across the region place greater emphasis on AI infrastructure, chip supply security, and local or diversified compute access, that could signal that the compute-centric framework is shaping real investment behavior.
Finally, the key question is whether concentration in advanced compute persists. If it does, the strategic premium attached to leading AI hardware and related infrastructure may remain elevated. If it begins to erode through new capacity, substitution, or regional diversification, the geopolitical balance described in the report could become less stable and less one-sided.
The broader takeaway is not a simple call on one company. It is that semiconductors, AI infrastructure, and geopolitics are becoming more tightly linked, and Asia sits at the heart of that linkage. In that environment, investors should pay close attention not only to chip demand, but also to who controls the systems, supply chains, and physical capacity that make advanced AI possible.