Chokepoints & Workarounds

The world is busy rerouting. From oil flowing around the Strait of Hormuz to bits of neural networks escaping walled gardens, actors large and small are trying to turn single points of failure into a messy, resilient web. The result looks less like tidy deterrence and more like a global version of urban planning after every bridge has been claimed by a hostile neighbor.

This week's dispatch stitches together three flavors of strategic improvisation. Riyadh and Abu Dhabi are betting that new pipelines can blunt Tehran's maritime leverage; militaries are betting that cheap, mass-producible defences can blunt cheap, mass-produced drones; and technologists (and policy wonks) are arguing the U.S. should embrace open-weight AI rather than try to put a moat around compute and models. Each choice is practical, but none are morally neutral: they change who holds power, how quickly wars spread, and who profits from the new plumbing.

Beneath the tactical moves is the same logic: when a choke point becomes weaponized, everyone rushes to multiply routes, reduce single points of leverage, and turn concentrated threats into diffuse problems. It sounds prudent—until you remember diffusion also creates new fault lines, new markets for violence, and new reputational hazards for firms, states, and ops officers who thought redundancy equals safety.

The Strait of Hormuz has been weaponized; so its neighbors are rewriting the map. Iran's campaign of drone and missile attacks on civilian shipping has taken a centuries-old sea corridor and turned it into a political liability. Where once roughly 23 million barrels per day transited with the air of a neutral highway, Tehran's actions made the waterway a bargaining chip and a target. See the reporting in CNN for the strategic fallout: CNN.

The regional answer is not guns but pipes. Saudi Arabia aims to expand its east–west Red Sea route by 1–2 mb/d, pushing bypass capacity toward 9 mb/d by 2029; the UAE will double its desert‑to‑coast capacity to 3.6 mb/d by 2027; Iraq is chasing Mediterranean links through Syria, Turkey, and Jordan. Goldman Sachs estimates these corridors could carry roughly 60% of Hormuz's prewar volumes by around 2028. The goal—explicitly framed by U.S. and regional officials—is to render Hormuz “an afterthought” rather than a choke point.

This is geopolitics by infrastructure: make your adversary's leverage irrelevant by building redundant options. It works, in theory. In practice the projects face political friction, financing headaches, and obvious counter‑moves: sabotage, escalation, and the simple fact that pipelines take years, not weeks. Still, if completed they will redistribute economic and strategic weight across the region, turning Tehran's maritime coercion into an expensive, long‑term strategic failure—provided the projects themselves don't become targets.

War is getting modular and affordable. The U.S. military says it used the Israeli‑designed Merops counter‑drone system to shoot down an Iranian Shahed—a first combat kill for the system in U.S. hands after a long run of successes in Ukraine. The details read like a Silicon Valley pitch for low-cost attrition: radar, a launcher, a ground‑control station, and a $15,000 interceptor that can hit over 175 mph and be flown with an off‑the‑shelf Xbox controller. (Read the Business Insider account here: Business Insider).

This matters for three reasons. First, it signals a shift from expensive, centralized air defence to distributed, relatively disposable layers—defences you can proliferate among allies and partners. Second, mass‑production economics matter: Army officials expect per‑unit costs to fall below $10,000 with scale, which makes stooping costs compatible with attrition warfare. Third, there's a doctrine change: make drone strikes cheaper and more deniable, and opponents will meet them with cheap, scalable counters.

The arithmetic is sobering. Cheap attackers plus cheap defenders doesn't necessarily reduce escalation; it democratizes the battlefield. More actors can throw low‑cost, high‑impact tools into crises. The hope is that cost curves favor defence enough to blunt swarms; the risk is that lowering the price of violence lowers the political cost of using it.

Meanwhile, in the digital realm, openness is the new strategic terrain. A persuasive thread on the open‑weight movement argues that weights—raw model parameters released publicly—are becoming the substrate for the next AI ecosystem. The analogy on offer is Kubernetes: weights are portable artifacts that let developers run, tweak, and stitch models into new services. The original essay (circulated on Hacker News) urges the U.S. to compete inside that open ecosystem rather than try to wall it off: Hacker News.

Key datapoints make the point practical: Hugging Face hosts >2 million public models; competing open models like GLM‑5.2 score competitively against closed‑box incumbents; and 41% of model downloads on Hugging Face come from Chinese models, which complicates any blunt export ban. The prescription is tidy: release useful American open weights under usable licenses, use procurement to seed interoperable stacks, and fund the rest of the ecosystem—serving, tooling, runtimes, safety standards—rather than pretend APIs are permanent moats.

Why this belongs in the same newsletter as pipelines and Merops is less about code and more about leverage. Open weights decentralize capability the way pipelines decentralize oil transit: they reduce single‑vendor choke points and force strategic competition to move to standards, tooling, and service quality. Build the rails and you cede the edge to whoever builds better products on top. Stick behind fences and the world will knit its economy around the open substrate—with or without you.

If you preferred your geopolitical crises concentrated, you missed the memo. The week saw another near‑collision at Scarborough Shoal where, according to CNN, a Chinese flotilla used water cannon against a Philippine vessel carrying a CNN crew while warning them to "Leave immediately … Otherwise, we will take firm action in accordance with the law" (CNN). Simultaneously, Iran declared bases used by the U.S. in Britain as "legitimate targets," prompting the UK to say it was “ready 24/7 to defend itself” as tensions climbed (reporting in BBC here: BBC).

These incidents are not unrelated curiosities. The South China Sea confrontation is about maritime control and the coercive use of civilian modalities (fishing boats, coastguard vessels, and water cannon) to achieve strategic aims. Britain's public posture—hosting U.S. assets and then having them labeled "legitimate targets" by Tehran—shows how allies become geographic multipliers for distant conflicts. Both stories illustrate a shared dynamic: states using ostensibly civilian infrastructure and ambiguous means to assert influence without crossing classical thresholds of war.

That ambiguity is dangerous because it lowers the bar for miscalculation. A water cannon comes too close and a ship takes on water; a drone swarm seeks a crude oil tanker and an interceptor misses; a foreign missile hits a base in Britain and Washington feels compelled to answer. In an era of many cheap tools and many routes—literal and digital—escalation management becomes the rarest commodity.

Here's the throughline: weaponized chokepoints drive redundancy; redundancy creates new markets and new vectors for contestation. Pipelines bypass Hormuz; open weights bypass API gatekeepers; Merops and similar systems proliferate defense. The clever part is that each workaround is itself programmable and therefore contestable.

Quick takeaways:

• Pipelines and alternative sea routes reduce Tehran's leverage but present new targets and financing bottlenecks (CNN on pipelines).
• Cheap interceptors lower the cost of defense and, perversely, of attack, by making drone campaigns more feasible (Business Insider on Merops).
• Open weights democratize AI capability in the same way transit redundancy democratizes energy flows; a policy of isolation risks ceding standards to others (Hacker News on open weights).

The practical policy implication is humbling: resilience doesn't remove politics. It reallocates it. You don't eliminate leverage—you change its location from one chokepoint to many. That is progress if the rule‑set governing those new locations is stable, transparent, and enforceable. It is a mistake if the new architecture simply fragments power into a thousand small fights.

Final thought: We're living through a period when strategic thinking is being outsourced to engineers and traders. Pipelines are the new diplomacy; model weights are the new arsenals; $15,000 interceptors are the new patriot missiles. That decentralization has virtues—it reduces single‑point pressure—but it also turns global strategy into a thousand small competitions, each with its own winners, losers, and accidental wars.

Plan for redundancy, then plan the politics that will govern it. Expect smarter adversaries to target the seams. And when you describe the future to a friend, say this: modern power is less about holding a chokepoint and more about defining the rules for the thousands of ways people will try to get around it.

Sources: CNN (Hormuz), Hacker News (Open Weights), Business Insider (Merops), BBC (UK readiness), CNN (Scarborough).