The Next Strait of Hormuz Isn’t a Strait: AI, Invisible Risk, and the Future of Global Power

For decades, global risk has been understood through a geographic lens. Power was mapped through physical corridors — straits, pipelines, ports, and borders — where disruption could be anticipated, monitored, and, to some degree, controlled. Few places have symbolized this reality more clearly than the Strait of Hormuz, a narrow maritime passage whose stability has long been tied to the stability of global energy markets. That mental map, however, is no longer sufficient.

While physical chokepoints continue to matter, the architecture of global systems has fundamentally shifted. Today, the most consequential vulnerabilities are not only geographic; they are embedded in the digital infrastructures that increasingly govern how energy flows, how supply chains function, and how decisions are made. These infrastructures are algorithmic, distributed, and often invisible to those who depend on them. The greatest risks to global systems are no longer physical chokepoints but algorithmic, systemically embedded decision infrastructures.

This shift does not eliminate the importance of geography but instead it transforms it. Oil still moves through straits, goods still pass through ports, and energy still depends on physical infrastructure. But the coordination of these systems, the allocation of resources, the timing of flows, the pricing of commodities, and the response to disruption is now mediated by digital platforms and increasingly by artificial intelligence. In this sense, geography has not disappeared. It has been layered with a new, less visible geography of code, data, and decision logic.

In energy markets, for example, algorithmic trading systems now interpret signals and execute transactions at speeds far beyond human capability. In logistics, routing decisions are dynamically optimized through real-time data streams, adjusting to disruptions as they occur. In financial systems, automated responses to market movements can amplify volatility within seconds. Across these domains, decisions are no longer sequential and human-driven. They are simultaneous, automated, and deeply interconnected.

This transformation is well underway, but it is still in its early stages. The systems described here will continue to evolve as technologies become more sophisticated and are embedded more deeply across sectors. The implications for risk, power, and resilience will therefore expand in the years ahead.

From Physical Chokepoints to Invisible Systems

This transformation introduces a new kind of chokepoint — one that cannot be easily mapped or secured. Unlike a strait or a pipeline, algorithmic chokepoints are distributed across systems. They reside in models, in data architectures, and in the interactions between autonomous processes. They are not controlled by a single actor, nor are they always fully understood by those who rely on them. Power, as a result, is being redefined.

Historically, control over resources and territory has conferred strategic advantage. Today, advantage increasingly lies in the ability to design, deploy, and govern intelligent systems. Data becomes a strategic asset not only because of its volume, but because of how it informs decision-making. Algorithms become instruments of power because they shape how reality is interpreted and how action is taken. The capacity to influence these systems, whether through innovation, regulation, or disruption becomes as significant as control over physical assets. Yet this shift also introduces new forms of fragility.

As systems become more intelligent and more optimized, they also become more sensitive to error. A flawed model, a biased dataset, or an adversarial input can propagate through interconnected systems, producing outcomes that are difficult to predict and even harder to contain. In highly optimized environments, small perturbations can trigger disproportionate effects, particularly when systems are tightly coupled and operating at high speed.

The energy transition illustrates this dynamic. As energy systems become more distributed and reliant on digital coordination through smart grids, decentralized generation, and real-time balancing, the role of algorithmic decision-making expands. This increases efficiency and responsiveness, but it also introduces dependencies on systems that may be vulnerable to failure, manipulation, or misalignment. The resilience of the system no longer depends solely on physical redundancy; it depends on the robustness of the underlying decision infrastructure.

An even more complex dimension of risk emerges when these systems interact. We are entering a world in which multiple autonomous systems operate simultaneously across sectors and borders. These systems are often developed independently, governed by different standards, and optimized for different objectives. Yet they interact continuously, responding to shared signals and influencing one another in real time.

In such an environment, disruption is less likely to result from a single point of failure and more likely to arise from the interaction of systems, or lack thereof. Competing algorithmic trading platforms may react to the same market signal in ways that amplify volatility. Logistics systems may reroute simultaneously, creating congestion rather than alleviating it. Security or defense systems may interpret ambiguous data and escalate responses before human decision-makers have the opportunity to intervene.

These dynamics introduce a new category of risk: fast-moving, non-linear, and often difficult to attribute. Events may unfold across multiple systems, leaving no clear origin and no single point of control. The speed of these interactions compresses decision timelines, reducing the ability of human actors to observe, interpret, and respond. This creates a new category of invisible risk. 

Digital systems are often perceived as tools that enhance control and predictability. In reality, they may produce a different kind of uncertainty—one that is less visible but more pervasive. AI-driven systems operate probabilistically, not deterministically. They identify patterns, generate predictions, and recommend actions based on data, but they do not guarantee outcomes. As these systems take on a greater share of decision-making, leaders must navigate environments in which not all decisions are transparent and not all outcomes are explainable. These dynamics create a leadership challenge that cannot be solved through faster reaction alone. They require a new capacity to govern the invisible systems through which risk is produced, interpreted, amplified, or contained.

The Rise of Algorithmic Power

The metaphor of the Strait of Hormuz remains useful, but it must be reinterpreted. The next global disruption is unlikely to originate from a single, visible chokepoint. It is more likely to emerge from the convergence of digital dependency, systemic complexity, and accelerated decision-making. It may begin with a misaligned algorithm, an unexpected interaction between systems, or a cascade triggered by incomplete or misleading data.

When such an event occurs, it may not be immediately clear where it began, who is responsible, or how it can be contained. The disruption may propagate through networks faster than traditional response mechanisms can operate. In this sense, the next Strait of Hormuz is not a place. It is a condition. It is a state in which global systems are tightly interconnected and digitally mediated. They are also operating at a speed that challenges human comprehension.

The task, therefore, is not to eliminate uncertainty, but to build the capacity to absorb, interpret, and respond to disruption as it moves across systems. This reframes algorithmic power as a governance challenge, not only a technical one. 

It also requires collaboration across sectors and borders. Algorithmic systems do not operate in isolation, and their risks cannot be managed by any single organization or nation. Shared standards, transparency mechanisms, and cooperative governance will become increasingly important as the density of these systems grows.

At the same time, technology will not independently determine how this landscape evolves. Geopolitics, public policy, and cultural contexts will shape how algorithmic systems are developed, deployed, and constrained. Different societies will make different choices about autonomy, oversight, data use, and risk tolerance, leading to divergent models of governance and, potentially, new forms of fragmentation or competition across digital infrastructures.

The transition from oil routes to algorithms is not merely a technological shift. It is a reconfiguration of global power. Those who understand and shape this new geography who can see the invisible infrastructures that underpin modern systems will be better positioned to navigate disruption and to create value in an era defined by complexity.

Those who do not may find themselves reacting to events that seem to come from nowhere, yet are deeply embedded in the systems they depend upon. The challenge is not only to adapt to a changing world but to learn how to see it differently.

Leadership in the Age of Invisible Risk

The rise of algorithmic power changes not only where risk resides, but what leadership requires. When disruption emerges from systems that are distributed, automated, and difficult to see, leaders can no longer rely only on traditional models of oversight. The task is not simply to respond more quickly when disruption occurs. It is to understand and govern the systems that determine how disruption is detected, interpreted, amplified, or contained.

This requires a shift from reactive leadership to systems leadership. Leaders must move beyond asking, “What happened?” and begin asking, “What conditions allowed this to happen, and how might this system behave under stress?” In an algorithmically mediated world, leadership depends on the ability to see dependencies that are not visible on an organizational chart, a balance sheet, or a map. The most important vulnerabilities may be embedded in data flows, model assumptions, vendor relationships, cloud infrastructure, automated decision rules, or the interaction of systems optimized for different goals.

The first leadership responsibility is to make invisible dependencies visible. Leaders do not need to become technologists, but they do need to understand the architectures on which their organizations and markets depend. A system may appear efficient while becoming more fragile. A supply chain may appear optimized while losing the slack needed to absorb disruption. A model may appear accurate while reinforcing assumptions that fail under stress. To lead in this environment, leaders must learn to see not only assets and outcomes, but also interdependencies, feedback loops, and hidden points of concentration.

The second responsibility is to question the decision logic embedded in algorithmic systems. Algorithms do not simply process information. They classify, prioritize, rank, predict, route, price, and recommend. In doing so, they encode judgments about what matters, what counts as risk, what should be optimized, and what can be ignored. Leaders must therefore ask not only whether a system works, but what it is optimizing for, whose judgment it reflects, where its data comes from, and how it might behave when conditions change. The deeper question is not whether an algorithm is efficient, but whether it is aligned with the organization’s purpose, obligations, and tolerance for systemic risk.

The third responsibility is to govern speed and autonomy. In digital systems, speed is often treated as an unquestioned advantage. Faster transactions, faster logistics, faster intelligence, and faster responses all appear to create value. Yet in tightly connected systems, speed can also magnify error. When automated systems react to one another in real time, a local misinterpretation can become a market signal, a routing failure can become a supply chain disruption, and a defensive response can become an escalation. Leaders must know when to accelerate and when to slow the system down. The ability to pause, override, audit, or reintroduce human judgment may become as important as the ability to automate.

The fourth responsibility is to design resilience before disruption occurs. Invisible risk cannot be managed only through crisis response. It requires architectures that can absorb shock, governance protocols that define when humans intervene, and stress tests that examine how systems interact rather than how they perform in isolation. Leaders should ask not only, “What happens if this system fails?” but also, “What other systems will respond, and how might those responses interact?” Resilience in an algorithmic environment depends less on preventing every failure than on preventing failures from cascading across interconnected systems.

Finally, leadership in this environment requires a broader sense of accountability. Algorithmic systems do not operate within the boundaries of a single organization, sector, or nation. They depend on platforms, vendors, data centers, regulators, standards bodies, and multinational networks of coordination. As a result, fiduciary responsibility must expand beyond financial oversight and operational compliance. Leaders must become stewards of the decision infrastructure on which their organizations and stakeholders increasingly depend.

Leadership in the age of invisible risk is therefore not simply the management of assets or the response to disruption. It is the discipline of seeing hidden dependencies, questioning embedded assumptions, governing the boundary between human judgment and machine autonomy, and building systems capable of absorbing shocks that cannot be fully predicted. The leaders best prepared for this environment will be those who can govern not only what is visible, but the invisible systems that increasingly shape global power.

Redrawing the Power Map

For generations, leaders have been trained to see risk where it is visible. They see it along coastlines, across borders, and within the physical infrastructures. This analysis suggests that the map is no longer sufficient. The most consequential disruptions of the future will not always originate in places we can point to on a chart. They will emerge from systems we have built but do not fully see or govern.

The transition from oil routes to algorithms represents more than a technological evolution. It marks a fundamental redefinition of where power resides and how it is exercised. Control over physical chokepoints once conferred advantage because it enabled the disruption or protection of flows. Today, advantage increasingly lies in shaping the systems that determine how those flows are interpreted, optimized, and executed in real time. The new geography of power is not bounded by territory but is embedded in code, data, and decision architectures.

This shift demands a new mindset. Leaders must move beyond managing visible risk to recognizing the invisible dependencies that now shape organizational resilience, market stability, and geopolitical power. The task is not only to adapt to disruption, but to build the capacity to see, question, and govern the systems through which disruption can spread.

Equally important, it requires collective action. Algorithmic systems do not respect national boundaries, and their risks cannot be mitigated through unilateral strategies. The governance of these systems, which includes how they interact, how they are constrained, and how accountability is established, will define the stability of global markets and the security of nations. In this sense, the future of geopolitics will be shaped as much by cooperation in digital and algorithmic domains as by traditional diplomacy. Crucially, these outcomes will reflect ongoing negotiations between states, institutions, and societies, each bringing different priorities, values, and constraints to how these systems are governed and trusted.

The next Strait of Hormuz will not be a narrow passage between landmasses. It will be a convergence of systems, signals, and decisions that exposes the fragility of an interconnected world. Whether that moment becomes a crisis or an inflection point will depend on how well leaders have learned to see what was previously invisible. The imperative now is not simply to adapt, but to recognize that the terrain of global power has shifted, and to lead accordingly.

Diplomacy is a human time-scale instrument. It is deliberate, negotiated, and slow, as we are seeing in the latest diplomatic dialogue between the U.S. and Iran in Pakistan. On the contrary, markets and stocks, which are heavily reliant on algorithms, respond swiftly. When this war started and Hormuz was closed off, Asian markets and oil stocks started to respond immediately. Under such an unpredictable swirl, international traders and businesses need to make split-second decisions.

As discussed, AI and human-machine interaction can aid diplomatic solutions in such crises by enhancing crisis forecasting, facilitating faster negotiations, and providing impartial data analysis to de-escalate tensions. AI tools help identify trends from massive datasets, offering strategic advantages for early warning and conflict mitigation. Humans are capable of sound, rational decisions, but that capacity is slowed by fatigue, cognitive load, and lapses in attention. Even at our best, we are bound by biology. A 2013 study in Scientific Reports published in Nature found over 18,000 financial events that happened too fast for any human to see, let alone stop. The machines are not smarter than us, they are just operating on a clock we cannot reach.  

Yet this same capability introduces new risks — specifically governance risks. The acceleration of insights and actions does not necessarily produce stability. These systems require informed governance. Across recent industry surveys, organizations report adopting AI at roughly twice the rate at which they govern it. The following table shows the breakdown:

Figure 1

AI Readiness Gap: Organizations are using AI faster than they are governing it

Note. Table compiled by the author(s). Data sources: Deloitte 2026 Global Human Capital Trends (https://www.deloitte.com/us/en/insights/topics/talent/human-capital-trends/2026/decision-making-with-ai.html); Deloitte 2024 board survey on AI oversight (https://www.deloitte.com/us/en/insights/topics/leadership/successful-ai-oversight-may-require-more-engagement-in-the-boardroom.html); Deloitte High Impact Decision Intelligence research, 2023 (https://action.deloitte.com/insight/3316/decision-intelligence-the-time-is-now); Oracle Global Decision Dilemma Study, 2023 (https://www.prnewswire.com/news-releases/global-study-70-of-business-leaders-would-prefer-a-robot-to-make-their-decisions-301799591.html).

Governing AI in business and trade does not begin with slowing systems down. Rather, it starts with deciding, in advance, where speed should be absolute, where it should be constrained, and where it should be forbidden altogether. In practice, three factors need to be accounted for.

First, firms must audit the assumptions embedded in the models they deploy, as no algorithm is impartial, and every algorithm is crystallized by its designers. Auditing is therefore crucial. This includes understanding what data trained the models, what outcomes they were optimized for, and whose judgment shaped those choices. Some initiatives have been taken by AI developers and companies. A recent example is Anthropic AI’s decision not to make its Claude Mythos model publicly available, as they feared it was too powerful for public release and could lead to remote code execution vulnerabilities, exploitation, and high-severity cyber risks in the absence of robust public AI governance. Under "Project Glasswing," Anthropic instead decided to give restricted access to over 50 tech organizations, including Microsoft, Nvidia, and Cisco, which have the security capabilities for defensive and patching efforts.

Second, organizations need to create intentional friction in high-stakes decisions. This includes establishing points where human judgment is necessary, pathways that are tested rather than assumed, and clear rules for when systems need to step back, even if it means sacrificing efficiency.

Third, governance cannot just be a quarterly topic in the boardroom. It must move at the same pace as the systems it oversees, using ongoing monitoring, real-time escalation channels, and shared standards across companies, industries, and countries. One firm managing its own models effectively is not enough if its algorithms interact with others it cannot control.

This brings us back to our original arguments on Hormuz. The strait has remained governable for more than half a century for oil, LNG, and other critical goods transshipment, not because it is physical, but because the speed of the threat has matched the speed of the response. A closure unfolds over hours or days, with lethal strikes on civilian vessels, something not seen before, long enough for diplomacy to convene, for markets to price the signal, and for naval forces to reposition. That symmetry is what makes the chokepoint manageable. Today, when Hormuz closes, the physical disruption still moves at the speed of diplomacy, but the economic disruption moves at the speed of code. Oil futures, shipping insurance, currency pairs, and equity indices all respond within milliseconds to the same signal, and they respond to each other. The diplomatic channel in Pakistan operates on one clock. The algorithmic cascade across Asian markets operates on another. Governance in this environment is not about choosing between them. It is about ensuring that the slower clock, where legitimacy, accountability, and human judgment still reside, does not become irrelevant to the faster one. The next Hormuz-scale event will not be a failure of diplomacy or a failure of technology. It will be a failure to have built the institutions that let the two clocks speak to each other in time. While we are talking about AI software for key decision-making, AI-powered robots or humanoids can also become key assistants in human diplomacy. A recent article published by the University of Southern California’s Center on Public Diplomacy asserts that:

“Robots might mediate disputes over resources or climate policy with impartial accuracy, while humans interpret the emotional undercurrents, the trust, hesitation, and hope that no algorithm can fully quantify.”

The report highlights Sophia, a humanoid robot, for gaining popularity in asking questions and making decisions based on ethics, personhood, and AI governance. Saudi Arabia has already granted it citizenship. These AI-powered robots and technologies are not going to replace humans; rather, they can assist human negotiators by providing faster and more accurate information, translating languages, analyzing historical patterns, and generating multiple decision-outcome scenarios in real time in negotiation settings. While efforts are underway to integrate AI into diplomacy, AI governance needs to be reconsidered to address human bias in how systems are designed and to ensure human-supervised code correction and continuous refinement.