Do States Have Personalities? What IR Theory Gets Right About Tech Chokepoints

A well-established strand of international relations theory holds that states develop distinct, semi-predictable “personalities” through history, geography, and social interaction. In the 1970s, Keohane and Nye challenged that state-centric view, arguing that transnational actors could bypass the state and reshape its identity from the outside. That framework now applies directly to physical infrastructure, like semiconductor fabs, and digital infrastructure, like cloud and model access. It explains a security risk most organizations are not currently measuring: a small number of firms now hold the kind of leverage over state capability that used to require a blockade or an embargo.

Action points for leaders

1. Audit vendor concentration as a geopolitical exposure, not just an operational one. Identify which cloud region, foundry node, or model API represents a single point of failure tied to another jurisdiction’s export controls or national interest.

2. Build and maintain a chokepoint map. Name the two or three suppliers whose service degradation or revocation would materially change your operating capability within a quarter. Refresh it on a set cadence, not once and shelved.

3. Anticipate state-level scrutiny of your own firm if you sit at one of these nodes. Expect regulatory attention, export control entanglement, and reputational exposure to follow the same logic states apply to each other.

The Theory

International relations has a long tradition of treating nation-states as having distinct, durable “personalities,” not as a metaphor, but as an analytical claim with real predictive power. Alexander Wendt argued that state personalities form the way an individual’s would: through relationships and repeated social interaction, not as a fixed trait handed down by geography or power alone. Strategic culture theory locates the source elsewhere, in geography, history, and formative trauma, which get embedded in military doctrine and elite thinking long after the triggering event has passed.

K. J. Holsti added a third layer: states adopt self-images (regional leader, neutral broker, protector of small states) and those self-images function as identity, shaping which actions feel congruent with “who we are.”

This view was challenged in the 1970s by Robert Keohane and Joseph Nye. They argued that transnational actors — corporations, NGOs, diaspora networks — could bypass the state entirely and reshape its identity from the outside, editing the state’s self-concept over time rather than simply lobbying its interests. Their broader claim was that transnational interdependence would erode state-centric power as the primary organizing logic of the international system.

Why This Matters for Tech

Their framework extends further than that today, directly onto physical infrastructure like semiconductor fabrication plants and digital infrastructure like cloud hyperscalers. The depth of supply chain integration and cloud dependency now produces a security risk that is structurally difficult to disentangle, because the transnational actor reshaping state capability isn’t a movement or an NGO. It’s a small number of firms sitting at physical and algorithmic chokepoints.

The Cold War version of interdependence was bilateral and visible. You could see a tariff, count an embargoed good, track a supply line. The current version is centralized around a handful of firms whose decisions function like state action without carrying state accountability.

A cloud provider suspending service, a foundry prioritizing one customer’s wafer allocation over another’s, or a model provider restricting API access are not diplomatic acts, but they carry the same coercive weight. They’re exercised by actors who were never meant to hold that kind of leverage and who answer to shareholders rather than electorates.

This is a sharper version of the Keohane–Nye thesis than they likely anticipated. They expected erosion of state power through diffusion: more actors, more channels, more complexity. What’s emerged instead is erosion through concentration.

A small number of firms sitting at physical and algorithmic chokepoints can unilaterally alter a state’s capability set the way a blockade once could.

What This Means for Leaders in Tech and Policy

Vendor risk is geopolitical risk, and it should be underwritten as such. A dependency on a single cloud region, foundry node, or model API is not just an operational single point of failure. It’s exposure to a decision made under a different state’s jurisdiction, export control regime, or national interest calculus.

Due diligence that stops at SLAs and uptime guarantees is measuring the wrong thing. Map your chokepoints before an external actor forces you to. Few organizations can name, with precision, which suppliers could be revoked or degraded and materially change their operating capability within a quarter.

That map is a prerequisite for real resilience planning, and it needs to be redrawn regularly. Firms operating at these chokepoints should expect to be treated like state actors, because they increasingly function like one. That cuts both ways.

It’s leverage in negotiations, and it’s a magnet for regulatory attention, export control entanglement, and reputational exposure when a service decision lands with geopolitical consequences the firm didn’t intend and may not be equipped to manage. The state-personality literature was built to explain why nations behave in patterned, semi-predictable ways. The more useful question for the next decade may be whether that same lens now applies to the companies sitting inside their supply chains, and whether those companies are prepared for the scrutiny that comes with it.