How are global supply chains restructuring post-COVID — nearshoring, friendshoring, and the end of just-in-time
Why Your Stuff Stopped Arriving on Time — and Why Fixing It Is Harder Than It Looks
Based on analysis of a 112-node, 410-edge knowledge graph mapping post-COVID global supply chain restructuring.
The Lego Problem
Imagine you built an enormous Lego spaceship. Every single piece came from one factory in one town. The factory was incredibly efficient — it made exactly the right piece, in exactly the right quantity, and shipped it to you the day before you needed it. No storage costs, no waste, no extra pieces lying around. Perfect.
Then the factory closed for a year.
That is roughly what happened to global supply chains during COVID. For decades, companies had built systems designed around one idea: order what you need, exactly when you need it, from wherever it is cheapest to make. Usually, that was China. The system worked beautifully — right up until it didn’t.
What the knowledge graph maps is what happened next: every attempt to fix the problem, and every reason those fixes are harder than they appear.
The Old System: Just Enough, Just in Time
The old way of running supply chains is called “just-in-time” manufacturing. The idea is simple: don’t keep a warehouse full of parts. Instead, have your suppliers deliver exactly what you need, right when you need it. Less storage, less cost, more efficiency.
This worked for about forty years. Companies got very good at it. They also got very good at finding the cheapest possible place to make each component — which often meant China, because of its combination of scale, infrastructure, cost, and manufacturing skill.
The hidden cost was fragility. Just-in-time only works if every link in the chain delivers on schedule. One late shipment, one closed port, one factory shutdown, and everything grinds to a halt. There was no buffer. The system had no slack.
When COVID hit, there was no slack to absorb the shock.
The Obvious Fix and Why It Creates New Problems
The obvious response was: stop relying on just-in-time. Keep more stuff in stock. Build in a buffer. This is called “just-in-case” inventory — instead of ordering what you need today, you order more than you need and keep the extra on hand in case something goes wrong.
Companies did this. And then something interesting happened.
When everyone started stockpiling at the same time, orders went through the roof. Factories ramped up production to meet demand. Then, about a year later, all those stockpiles arrived simultaneously. Companies suddenly had warehouses full of stuff they didn’t need for months. Orders collapsed. Factories slowed down. Prices swung wildly.
This is called the “bullwhip effect” — a small wobble at the consumer end of the chain creates a big whipping motion at the production end. Just-in-case inventory is supposed to reduce this problem, but when everyone adopts it at once, it temporarily makes it worse.
The graph shows something subtler still: keeping large stockpiles costs money. Companies have to borrow to finance all that inventory. That borrowing creates financial stress. Financial stress makes companies cut back — including cutting their stockpiles. Which leaves them exposed again. The fix contains the seed of the next problem.
Moving the Factory: The Sounds-Simple Plan
If the problem is depending too much on one country, the solution seems obvious: move some of the manufacturing somewhere else. Or build new factories closer to home — “nearshoring” — or in countries that are considered political allies — “friendshoring.”
The United States passed major laws (including the IRA and CHIPS Act) to encourage companies to build factories in America or in allied countries. Mexico became a major target for nearshoring, since it is close to the US and shares a trade agreement.
But the graph reveals a structural trap that keeps reasserting itself: moving the factory does not automatically move the supply chain.
Here is the concrete version. Suppose an electronics company moves its final assembly line from China to Vietnam. The finished product is now “Made in Vietnam.” But the circuit boards, the specialized chips, the battery cells, the fasteners, the cables — a significant share of those components still come from China. Vietnam assembles; China supplies.
The graph names this the “China Plus One Dependency Paradox.” It is not a prediction — it is a description of what has actually been happening. The data on trade flows shows that countries like Vietnam, Mexico, and India have indeed increased their share of finished goods exports to the US, but their imports from China have also increased in parallel. The assembly moved. The supply chain did not.
The Chokepoint Nobody Talks About
Deep inside supply chains, before you get to assembly, before you get to components, there is a layer called raw materials processing. This is where mined ore gets turned into refined materials that manufacturers can actually use.
For a surprisingly large number of critical materials — the things that go into electric vehicle batteries, semiconductors, solar panels, and defense systems — China controls the processing step. Not the mining, necessarily, but the refining and processing.
This matters because policy responses designed to reduce supply chain concentration keep running into this constraint. You can move assembly to Mexico. You can build a new semiconductor factory in Arizona. But if the specialized chemicals used in chipmaking, or the refined lithium for batteries, still flow primarily through Chinese processing facilities, the dependency has not been eliminated — it has been moved one step upstream and made less visible.
The graph shows this node — “Critical Minerals China Processing Monopoly” — sitting upstream of nearly every major Western policy response. It constrains them without those policies being able to easily constrain it back.
There is also a non-obvious wrinkle involving Europe’s carbon border tax. Europe has created a policy that charges extra for imports made with high-carbon industrial processes. The intent is to push manufacturers toward cleaner methods. The unintended effect, the graph suggests, is that alternatives to Chinese minerals processing are currently — at their smaller scale — often more carbon-intensive than the Chinese facilities they would replace. The policy designed to reduce concentration may inadvertently penalize the alternatives it needs to succeed.
The Announcement Gap
When a government announces a major factory investment — a new semiconductor fab, a battery plant, a pharmaceutical facility — the announcement is real. The completion is not guaranteed, and not always timely.
The graph captures this as a feedback loop. The laws that created incentives for reshoring also created the measurement frameworks to track whether reshoring was actually happening. Those measurements kept showing gaps between announcements and operational facilities. Those gaps then became political ammunition against the very laws creating the incentives.
Simultaneously, the tariffs meant to make domestic production more competitive also raised costs throughout the economy, including the costs of building the factories those tariffs were supposed to encourage. Higher construction costs, higher equipment costs, higher borrowing costs — all of which made the economics of reshoring harder, which widened the gap between announcements and reality.
The Defense-Automation Wrinkle
One of the less obvious connections in the graph involves defense manufacturing. Decades of offshoring hollowed out the industrial base that makes specialized military components. That hollowing created a problem for national security — but it also, paradoxically, created a captive market for automated domestic manufacturing.
Because defense procurement has requirements that cannot be satisfied by foreign suppliers, it generates a reliable, high-margin demand signal that justifies investing in expensive automation. Automation is expensive to deploy, but it narrows the wage gap between high-cost and low-cost countries — which is one of the things that makes reshoring economically viable.
The same cause — defense industrial hollowing — produces both the security vulnerability and one partial path toward solving it.
Why Nothing Fully Resolves
The graph does not show a system moving toward a stable new equilibrium. It shows a system with multiple self-reinforcing loops running simultaneously.
The JIT-to-JIC transition generates the financial conditions that eventually push companies back toward JIT. The China Plus One strategy generates the trade infrastructure that allows Chinese goods to flow through third countries, partially defeating its own purpose. The reshoring policies generate the political friction with allies that those policies depend on for success. The carbon border mechanism strengthens the concentration it is trying to reduce.
These are not design failures in individual policies. They are structural features of a complex, interconnected system. Each intervention has second-order effects that partially offset it. No single node in the graph resolves the core tension. The paradoxes are reinforced from multiple directions simultaneously.
Bottom Line
The knowledge graph describes a system under genuine stress, attempting to reorganize, and partially succeeding — but constrained by several structural features that do not yield easily to policy action.
Assembly is moving, but input sourcing is not moving at the same rate. The financial cost of resilience creates the conditions for the next fragility. Critical minerals processing is the deepest chokepoint, sitting upstream of most policy responses. The reshoring effort is real but systematically slower than announced. And the policy architecture that funds allied supply chain integration simultaneously creates friction among the allies it depends on.
The clearest structural finding is not that any single policy is wrong. It is that the system contains self-undermining loops at multiple levels, and those loops were largely invisible when the system was running smoothly. The knowledge graph is, in part, a map of what was not visible before — and a reminder that the next disruption is more likely to come from a feedback loop that is already present than from a shock that arrives from outside.