It made me think of 1951, when France and Germany took the chokepoints of their age, coal and steel, and put them under joint governance. It wasn’t a defensive arrangement. It was a positive project: make war between them materially impossible and build something together instead.
Today every major power seems to have a chokepoint strategy, and none a project like that.
What worries me most is AI. The White House’s 2025 AI plan calls “unquestioned and unchallenged global technological dominance” a national security imperative.
Peter Thiel taught a generation of founders that “competition is for losers” and that “monopoly is the condition of every successful business.” In search or payments, you can debate it. But that logic has now moved from companies to states, and it has been aimed at the most powerful general-purpose technology in history.
The difference is what happens to the loser. A company that loses a market closes or gets bought. A state that fears losing a race for power has another option: force.
"Chokepoint dominance" is but the latest evolution in monopolistic practices, albeit on an international scale, with ample assistance from governments. It would be nice to imagine Sherman or Clayton Acts governing international commerce in critical goods and materials, but I'm afraid that would be in the category of "AI will benefit societies" tosh, and never see realization, tant pis.
"As the late Robert Jervis argued decades ago, each power’s individual efforts to make itself safer may tighten a collective trap in which all of them end up worse off."
On Saturday, I chased down the source of an intermittent misfire on one cylinder of my 54-year-old Triumph motorcycle. It was a cold solder connection of the primary lead from the right-side set of points to its coil. I cut off the offending terminal, stripped and thoroughly cleaned the 18 gauge lead, and re-soldered and crimped the terminal shank. No more misfire. 50 mpg motorcycle back in daily service. Try that with ECU boards controlling the average Luxe-mobile.
My point here is that this ever-more hyper-complex pile of 'systems' leaves us ever more vulnerable to the skeevier operators, eager to capitalize on choke points that they can create in the name of 'efficiency', or hijack like literal pirates in the narrow points of trade.
No thanks. I hope to put my 52-year old Swedish wagon back into analog service this year, too.
Why are we painting ourselves into this corner? That's the question.
Really insightful piece here, Mr. Farrell. Thank you.
I’d suggest the U.S’ supranational and threateningly coercive approach to ASML isn’t simply a security strategy, but is just another manifestation of a by now deeply and insidiously embedded malignancy in the body of the hegemony, symptomatic of the broader cultural and systemic failure which will be added by history to the ever extending list of such and also ultimately will fail to mask its terminal decay.
I was hoping to learn from the New Yorker how, precisely, the U.S. prohibition on any sales of EUV machines to China is implemented. Is there a specific statute or regulation or does the USA simply tell the Dutch “don’t do that.” Furthermore, the amazing lasers that shoot the droplets of tin to produce EUV light — where did these lasers come from? Sounds like “high-energy physics” to me. Does that mean the National Labs were involved? Who owns the relevant IP? Some supply chains seem not to have end points….
But there's a significant qualitative difference between what China is doing, and what the USA is doing. China's actions generally are defensive in nature, and it only used the rare earth restrictions as a counter to the US weaponizing tariffs. Maybe that will change with China, but there's no sign of this - and I think the only reason that there is a differing perception in the west, is because the west analyzes China from the perspective of _what they would do_.
What changed between now and 20 years ago is that US global dominance made highly fragile export chains seem like a non-brainer. Now the downsides (to the US) of such a strategy have come into focus, along with China's rise as a plausible challenger and rapid US decline - the merits of such a strategy look very different _to the west_. It was always fragile, required a lot of coercion and so wasn't terrible stable. It's fall shouldn't be that surprising in retrospect.
If advanced chip production requires highly globalized supply/manufacturing chains then it may not survive. Or it may be replaced by something sturdier and far more centralized (probably in SE Asia, and dominated by China due to its sheer size). Add to that likely declines in certain forms of essential energy production (diesel primarily), and the world is likely to look very different over the coming decades. Smaller probably.
“a globally interdependent industry that pools the best each participant has to offer provides the best path to the future” may be poignant... but it is correct relative to the alternatives. The era of "decoupling"/"derisking" produced multifront major power war and global democratic decline almost immediately.
The Jervis security dilemma is *also* a double-edged sword. If there weren't so many chokepoints (including some in places like Netherlands and Taiwan) the global security regime would almost certainly be in a worse place. Having chokepoints in "open cities" provides a useful buffer. It spreads 'infrastructural power' out, and promotes niche proliferation, both of which build robustness into the system.
So (IMO) the EU isn't behind the times to resist this frame: "Now, global interdependence is threat rather than promise". It was always both, a version of "keep your friends close and your enemies closer". E.g., the original rationale for NATO: "Keep the Russians out, the Americans in, and the Germans down". The logic being that the middle of those is least desired but necessary for the other two. But there was always a threat that came along with this, as many European leaders noted both then and since. In fact, it is easy to argue that EU institutions themselves emerged in response to *this* form of interdependence, which was weaponized in the form of the US military building permanent bases all over Europe.
The USA switched sides politically, consequently the previous "interdependence" is now "asymmetric dependence". Any defection is going to hurt the people being defected against, but not having been interdependent previously is unlikely to have left countries in better strategic position. On the contrary, previous patterns of interdependence creates opportunities for rapid network revision in constructive ways, through k-adic closure and other mechanisms of 'community-building'.
That is why these pathetic "blockades" keep failing: with niche proliferation comes path proliferation.
As some leaders are finally noticing, having been indirectly connected (with a 'hegemon' as bridge) provides a social environment through which direct connections can be formed, connections that are *not* mediated by the 'hegemon'. Having an AMSL or TSMC buys time for those connections to form, provides an incentive for them to form, and the pre-existing structure provides clear pathways for them to form. *Too* much focus on static chokepoints can miss the ways in which the network is dynamic. Recent weaponization of chokepoints have proven to provide (at best) a short-run tactical advantage, which is only gained through sacrifice of long-run structural position (and thus not worth it for any leader with a time horizon longer than a week).
Which suggests to me -- perhaps controversially, as is my habit -- that the problem our various "nationalisms" are reacting against is not too much interdependence but too little: national leaders do not have the governance mechanisms to manage global complexity (such as that described in Gideon's excellent article). I know that's out of step with most current analysis, but to adopt Keohanian language for a second: the "demand for regimes" is clearly higher than the supply, as (almost) everyone fights over control of a system that is no longer operable.
The EU was never in control of this system and never wanted to be. That has been a disadvantage at points in the past, but could be an advantage now, should Europe decide to leverage it.
Isn't "the security dilemma" presented as a logically inevitable consequence of combining inter-state anarchy with weapons? Once you "weaponize" X, it does't matter whether X is nuclear fusion, anti-missile technology, anti-satellite technology, banking, internet connections, chip technology, or rare earths, all that matters is A) the effectiveness of the weapon and B) that there is no way of preventing weaponization or forcing disarmament. It always starts with "we will only use this weapon for good" or "we good guys must develop this weapon before the bad guys do" and it always ends with everyone armed and insecure.
Great article, and I love the "security dilemma" angle. You and your readers probably don't need a reminder, but it is always useful to recall that the US Dollar system is Washington's uber-checkpoint, and that its weaponization basically launched the ongoing descent into checkpoint warfare. Much of my work as a US diplomat involved sanctions enforcement, which was just a form of exploiting and weaponizing interdependence. Finally, it looks as though the New Yorker has put up the paywall on the Gideon Lewis-Kraus article.
Thank you for sharing this fascinating text as well as the Foreign Affairs article, which offered me a perspective that now seems obvious to me but that I hadn’t been aware of before.
I’m interested in ports, which (along with straits) are the most emblematic geographical features of chokepoint effects. In this field, the prevailing literature focuses primarily on China’s Belt and Road Initiative strategy and its effects, without mentioning how it was developed in response to the U.S.’s use of chokepoint effects.
Chokepoints are costly to exploit. For example, China's use of its monopoly on "rare earth" magnets is leading Europe to develop its own industry. Even with.a 30 per cent cost premium, lots of buyers will prefer to source their magnets their rather than submit to Chinese control.
What is the answer to this problem, though? Should the US not try to build up rare earths security? Should China not attempt to indigenize its semiconductor supply chain? I struggle to find solutions that don't ignore the increasingly adversarial relationship between the U.S. and China today. This isn't 1951 France and Germany.
This may be a naive and simplistic comment. We live on one planet, with natural resources spread around the globe. These have been exploited to different degrees to benefit either those in close proximity, by nature not man, or developed by those lucky enough to be born in there proximity. Yet, the development and use of natural resources to some degrees helps the world in general, if not used foolishly, and if often best done by local specialization.
This whole strategy of twisting arms with chokepoints is totally blind and a recipe for disaster. When everyone on the planet decides to grab each other by the throat, the entire world ends up suffocating and nobody can breathe
Great piece! I’d love to read your thoughts on what Europe should do here (or indeed other middle powers like Japan, Canada, Australia, India etc).
Is this a classic prisoner’s dilemma, whereby it would be optimal not to waste time and energy on building choke points, but once the others do it we have to do it too? Or maybe all we need is one or two (like ASML) to let us “carry a big stick” in stand-offs, but mostly aim to foster a more interconnected world?
Thank for this, and the link to your Foreign Affairs article as well.
In the latter you and your co-writers mention historical examples - I'd add the US cutting off oil to Japan in 1941, and pushed Japan to invade Indonesia, Borneo and Brunei for oil and others for strategic purposes.
Interesting piece. That said, our global supply chains (wheat, oil, etc) have shown remarkable resilience recently given COVID, Ukraine and the Iran War.
It made me think of "game theory" and A Beautiful Mind: If everyone makes a play for the same woman, no one meets the woman, so no one "wins" and if everyone uses chokepoints, then ultimately no one wins...
Thank you for laying this out so clearly.
It made me think of 1951, when France and Germany took the chokepoints of their age, coal and steel, and put them under joint governance. It wasn’t a defensive arrangement. It was a positive project: make war between them materially impossible and build something together instead.
Today every major power seems to have a chokepoint strategy, and none a project like that.
What worries me most is AI. The White House’s 2025 AI plan calls “unquestioned and unchallenged global technological dominance” a national security imperative.
Peter Thiel taught a generation of founders that “competition is for losers” and that “monopoly is the condition of every successful business.” In search or payments, you can debate it. But that logic has now moved from companies to states, and it has been aimed at the most powerful general-purpose technology in history.
The difference is what happens to the loser. A company that loses a market closes or gets bought. A state that fears losing a race for power has another option: force.
I think AI needs coordination on shared principles, not a race to dominate. I wrote about this in my first essay: https://thinkuphill.substack.com/p/we-didnt-just-lose-the-peace-we-lost
If chokepoints only buy each power more insecurity, what do you think it would take to start talking about shared principles again?
"Chokepoint dominance" is but the latest evolution in monopolistic practices, albeit on an international scale, with ample assistance from governments. It would be nice to imagine Sherman or Clayton Acts governing international commerce in critical goods and materials, but I'm afraid that would be in the category of "AI will benefit societies" tosh, and never see realization, tant pis.
"As the late Robert Jervis argued decades ago, each power’s individual efforts to make itself safer may tighten a collective trap in which all of them end up worse off."
On Saturday, I chased down the source of an intermittent misfire on one cylinder of my 54-year-old Triumph motorcycle. It was a cold solder connection of the primary lead from the right-side set of points to its coil. I cut off the offending terminal, stripped and thoroughly cleaned the 18 gauge lead, and re-soldered and crimped the terminal shank. No more misfire. 50 mpg motorcycle back in daily service. Try that with ECU boards controlling the average Luxe-mobile.
My point here is that this ever-more hyper-complex pile of 'systems' leaves us ever more vulnerable to the skeevier operators, eager to capitalize on choke points that they can create in the name of 'efficiency', or hijack like literal pirates in the narrow points of trade.
No thanks. I hope to put my 52-year old Swedish wagon back into analog service this year, too.
Why are we painting ourselves into this corner? That's the question.
Really insightful piece here, Mr. Farrell. Thank you.
Tim Long, Just Up the Hill from Lock 15
I’d suggest the U.S’ supranational and threateningly coercive approach to ASML isn’t simply a security strategy, but is just another manifestation of a by now deeply and insidiously embedded malignancy in the body of the hegemony, symptomatic of the broader cultural and systemic failure which will be added by history to the ever extending list of such and also ultimately will fail to mask its terminal decay.
I was hoping to learn from the New Yorker how, precisely, the U.S. prohibition on any sales of EUV machines to China is implemented. Is there a specific statute or regulation or does the USA simply tell the Dutch “don’t do that.” Furthermore, the amazing lasers that shoot the droplets of tin to produce EUV light — where did these lasers come from? Sounds like “high-energy physics” to me. Does that mean the National Labs were involved? Who owns the relevant IP? Some supply chains seem not to have end points….
But there's a significant qualitative difference between what China is doing, and what the USA is doing. China's actions generally are defensive in nature, and it only used the rare earth restrictions as a counter to the US weaponizing tariffs. Maybe that will change with China, but there's no sign of this - and I think the only reason that there is a differing perception in the west, is because the west analyzes China from the perspective of _what they would do_.
What changed between now and 20 years ago is that US global dominance made highly fragile export chains seem like a non-brainer. Now the downsides (to the US) of such a strategy have come into focus, along with China's rise as a plausible challenger and rapid US decline - the merits of such a strategy look very different _to the west_. It was always fragile, required a lot of coercion and so wasn't terrible stable. It's fall shouldn't be that surprising in retrospect.
If advanced chip production requires highly globalized supply/manufacturing chains then it may not survive. Or it may be replaced by something sturdier and far more centralized (probably in SE Asia, and dominated by China due to its sheer size). Add to that likely declines in certain forms of essential energy production (diesel primarily), and the world is likely to look very different over the coming decades. Smaller probably.
“a globally interdependent industry that pools the best each participant has to offer provides the best path to the future” may be poignant... but it is correct relative to the alternatives. The era of "decoupling"/"derisking" produced multifront major power war and global democratic decline almost immediately.
The Jervis security dilemma is *also* a double-edged sword. If there weren't so many chokepoints (including some in places like Netherlands and Taiwan) the global security regime would almost certainly be in a worse place. Having chokepoints in "open cities" provides a useful buffer. It spreads 'infrastructural power' out, and promotes niche proliferation, both of which build robustness into the system.
So (IMO) the EU isn't behind the times to resist this frame: "Now, global interdependence is threat rather than promise". It was always both, a version of "keep your friends close and your enemies closer". E.g., the original rationale for NATO: "Keep the Russians out, the Americans in, and the Germans down". The logic being that the middle of those is least desired but necessary for the other two. But there was always a threat that came along with this, as many European leaders noted both then and since. In fact, it is easy to argue that EU institutions themselves emerged in response to *this* form of interdependence, which was weaponized in the form of the US military building permanent bases all over Europe.
The USA switched sides politically, consequently the previous "interdependence" is now "asymmetric dependence". Any defection is going to hurt the people being defected against, but not having been interdependent previously is unlikely to have left countries in better strategic position. On the contrary, previous patterns of interdependence creates opportunities for rapid network revision in constructive ways, through k-adic closure and other mechanisms of 'community-building'.
That is why these pathetic "blockades" keep failing: with niche proliferation comes path proliferation.
As some leaders are finally noticing, having been indirectly connected (with a 'hegemon' as bridge) provides a social environment through which direct connections can be formed, connections that are *not* mediated by the 'hegemon'. Having an AMSL or TSMC buys time for those connections to form, provides an incentive for them to form, and the pre-existing structure provides clear pathways for them to form. *Too* much focus on static chokepoints can miss the ways in which the network is dynamic. Recent weaponization of chokepoints have proven to provide (at best) a short-run tactical advantage, which is only gained through sacrifice of long-run structural position (and thus not worth it for any leader with a time horizon longer than a week).
Which suggests to me -- perhaps controversially, as is my habit -- that the problem our various "nationalisms" are reacting against is not too much interdependence but too little: national leaders do not have the governance mechanisms to manage global complexity (such as that described in Gideon's excellent article). I know that's out of step with most current analysis, but to adopt Keohanian language for a second: the "demand for regimes" is clearly higher than the supply, as (almost) everyone fights over control of a system that is no longer operable.
The EU was never in control of this system and never wanted to be. That has been a disadvantage at points in the past, but could be an advantage now, should Europe decide to leverage it.
Isn't "the security dilemma" presented as a logically inevitable consequence of combining inter-state anarchy with weapons? Once you "weaponize" X, it does't matter whether X is nuclear fusion, anti-missile technology, anti-satellite technology, banking, internet connections, chip technology, or rare earths, all that matters is A) the effectiveness of the weapon and B) that there is no way of preventing weaponization or forcing disarmament. It always starts with "we will only use this weapon for good" or "we good guys must develop this weapon before the bad guys do" and it always ends with everyone armed and insecure.
Great article, and I love the "security dilemma" angle. You and your readers probably don't need a reminder, but it is always useful to recall that the US Dollar system is Washington's uber-checkpoint, and that its weaponization basically launched the ongoing descent into checkpoint warfare. Much of my work as a US diplomat involved sanctions enforcement, which was just a form of exploiting and weaponizing interdependence. Finally, it looks as though the New Yorker has put up the paywall on the Gideon Lewis-Kraus article.
Thank you for sharing this fascinating text as well as the Foreign Affairs article, which offered me a perspective that now seems obvious to me but that I hadn’t been aware of before.
I’m interested in ports, which (along with straits) are the most emblematic geographical features of chokepoint effects. In this field, the prevailing literature focuses primarily on China’s Belt and Road Initiative strategy and its effects, without mentioning how it was developed in response to the U.S.’s use of chokepoint effects.
Chokepoints are costly to exploit. For example, China's use of its monopoly on "rare earth" magnets is leading Europe to develop its own industry. Even with.a 30 per cent cost premium, lots of buyers will prefer to source their magnets their rather than submit to Chinese control.
What is the answer to this problem, though? Should the US not try to build up rare earths security? Should China not attempt to indigenize its semiconductor supply chain? I struggle to find solutions that don't ignore the increasingly adversarial relationship between the U.S. and China today. This isn't 1951 France and Germany.
This may be a naive and simplistic comment. We live on one planet, with natural resources spread around the globe. These have been exploited to different degrees to benefit either those in close proximity, by nature not man, or developed by those lucky enough to be born in there proximity. Yet, the development and use of natural resources to some degrees helps the world in general, if not used foolishly, and if often best done by local specialization.
This whole strategy of twisting arms with chokepoints is totally blind and a recipe for disaster. When everyone on the planet decides to grab each other by the throat, the entire world ends up suffocating and nobody can breathe
Great piece! I’d love to read your thoughts on what Europe should do here (or indeed other middle powers like Japan, Canada, Australia, India etc).
Is this a classic prisoner’s dilemma, whereby it would be optimal not to waste time and energy on building choke points, but once the others do it we have to do it too? Or maybe all we need is one or two (like ASML) to let us “carry a big stick” in stand-offs, but mostly aim to foster a more interconnected world?
Thank for this, and the link to your Foreign Affairs article as well.
In the latter you and your co-writers mention historical examples - I'd add the US cutting off oil to Japan in 1941, and pushed Japan to invade Indonesia, Borneo and Brunei for oil and others for strategic purposes.
Interesting piece. That said, our global supply chains (wheat, oil, etc) have shown remarkable resilience recently given COVID, Ukraine and the Iran War.
It made me think of "game theory" and A Beautiful Mind: If everyone makes a play for the same woman, no one meets the woman, so no one "wins" and if everyone uses chokepoints, then ultimately no one wins...