Skip to main content

Why Supply Chains Are Fragile: The Concentration Problem

Modern supply chains are simultaneously marvels of efficiency and monuments to fragility. They deliver goods across continents in weeks. They allow companies to operate with minimal inventory. They've enabled globalization and lower consumer prices.

But they fail catastrophically when anything goes wrong. A fire at a single factory in Japan shuts down auto plants worldwide. A virus forces port closures and freezes shipping for months. A geopolitical conflict threatens chip supplies for an entire continent. A shipping container gets stuck sideways in a canal and paralyzes global trade for days.

The fragility isn't a bug—it's a feature. Supply chains are optimized for efficiency, not resilience. But the optimization created several distinct failure modes that make even small disruptions cascade into crises. Understanding these failure modes is essential to understanding why the 2020-2022 supply-chain crisis happened and why future crises are inevitable.

Quick definition: Supply chain fragility is the result of concentration risk: when a few suppliers control critical components, when production is geographically concentrated, or when lead times are long enough that disruptions can't be quickly overcome, the entire system becomes vulnerable to small shocks.

Key takeaways​

  • Concentration creates single points of failure: When 80% of a critical component comes from one supplier or one country, the entire system depends on that supplier not failing
  • Geographic concentration amplifies natural disasters: Factory clusters in Japan, semiconductor production in Taiwan, shipping through the Suez Canal—all create geographic risk
  • Long lead times force inventory decisions before demand is known: If a component takes 90 days to arrive, you must order 90 days in advance, based on forecasts that are often wrong
  • JIT systems eliminate buffers, making disruptions immediate: With just-in-time, a 1-week supplier delay stops production within 1-2 weeks
  • Information opacity creates cascading failures: Downstream players don't see upstream disruptions until orders stop arriving
  • Fragility is cheaper than resilience: Diversifying suppliers, holding buffer inventory, and building geographic redundancy all cost money—more than consolidating around the cheapest supplier
  • Fragility is invisible until it fails: Companies seem to operate smoothly, then suddenly everything stops
  • Global supply chains are longer, making fragility worse: A 3-week disruption matters less in a 2-week supply chain than a 12-week supply chain

The Economics of Concentration: Why Fragility Makes Sense​

To understand why supply chains are fragile, you first need to understand why companies consolidate suppliers and concentrate production geographically. The answer is straightforward: it's cheaper.

A Numerical Example: The Cost of Resilience​

Imagine a manufacturer needs a specific electronic component. Two options:

Option A: Single Supplier (Cheap)

  • One supplier in Taiwan produces the component
  • Cost: $10 per unit
  • Lead time: 12 weeks
  • Supplier capacity: 100,000 units per month
  • Uptime: 99.5% (industry average)
  • Annual cost for 1 million units: $10 million
  • Concentration risk: If supplier fails, you have nothing

Option B: Dual Suppliers (Resilient)

  • Two suppliers: one in Taiwan ($10) and one in South Korea ($11)
  • Average cost: $10.50 per unit
  • Lead time: 12 weeks (both)
  • Each supplier has 50,000 units/month capacity (redundancy)
  • Uptime: 99.5% each
  • Annual cost for 1 million units: $10.5 million
  • Resilience: If one fails, the other covers 50% immediately, then ramps up

Option C: Three Suppliers, Geographically Diverse (Highly Resilient)

  • Taiwan ($10), South Korea ($11), Vietnam ($12)
  • Average cost: $11 per unit
  • Lead time: 12 weeks (Taiwan), 10 weeks (South Korea), 14 weeks (Vietnam)
  • Combined capacity: 150,000 units/month (50% redundancy)
  • Uptime: 99.5% each, but geographically separated (unlikely all fail together)
  • Annual cost for 1 million units: $11 million
  • Resilience: Multiple independent failure points, diversified lead times, geographic spread

The cost of resilience: $1 million per year, or 10% of total input costs. For a company with 3% profit margins, that's a 33% reduction in profit.

For a typical manufacturer, resilience is expensive. That's why most companies choose Option A: single supplier, maximum efficiency.

When Does a Company Choose Resilience?​

Companies choose resilience only when:

  1. The component is critical (production stops without it)
  2. The disruption cost exceeds the resilience cost (loss of production + reputation damage > cost of dual suppliers)
  3. Regulatory requirements demand it (government mandates supplier diversity)
  4. The market is competitive and tight (lost sales due to stockouts cost more than higher input costs)

For most commodity components, resilience isn't worth the cost. But for critical, specialized, or single-source components, companies should maintain redundancy. Often, they don't, until a crisis forces the issue.

The Four Key Failure Modes​

Supply chain fragility manifests in four main ways.

1. Geographic Concentration — When One Region Controls the World​

Semiconductors in Taiwan: Taiwan produces 92% of the world's advanced microchips (those below 16 nanometers). This concentration exists because:

  • TSMC achieved economies of scale (costs drop with volume)
  • Semiconductor fabrication requires specialized infrastructure (power, water, cooling, precision)
  • Taiwan invested heavily in the industry for decades
  • High barriers to entry (need billions in capital)

This concentration is extraordinarily fragile. If Taiwan faced geopolitical conflict, a major earthquake, or a typhoon that damaged multiple fabs, the world would have no advanced chips. Phones, cars, and data centers would grind to a halt.

The 2021 drought in Taiwan (reducing water for chip cooling) threatened chip supplies for the entire world. A drought.

Auto Manufacturing in Japan: Japan produces 10% of the world's cars, but supplies 30% of the world's auto components. Japanese suppliers are concentrated near Tokyo and Nagoya. The 2011 Fukushima earthquake disrupted Japanese supplier networks for months. U.S. auto plants shut down because they couldn't get Japanese parts.

Container Shipping through the Suez Canal: 12% of global trade flows through the Suez Canal (the shortcut between Europe and Asia). In March 2021, the Ever Given container ship got stuck crosswise in the canal, blocking all traffic for 6 days. For 6 days, hundreds of ships waited in queues. Supply chains everywhere experienced delays.

This is geographic concentration in transportation infrastructure, not production. But it has the same effect.

2. Supplier Concentration — When One Company Controls the Bottleneck​

TSMC's Dominance in Advanced Chips: Not only is chip production concentrated in Taiwan, but TSMC owns ~54% of the world's chip manufacturing capacity. If TSMC shuts down, 50% of the world's advanced chips stop being made.

Nokia's Dominance in Telecommunications (Historically): In the 1990s, Nokia was the world's dominant cellphone maker. It bought components from specialized suppliers. When one supplier had a fire, Nokia was crippled. Other phone makers could survive because they had multiple sources.

Neon Gas for Semiconductor Production: Advanced chip production requires ultra-pure neon gas. During the early 2020s, 70% of neon came from Russia and Ukraine. When Russia invaded Ukraine in February 2022, neon supplies tightened. Chip makers suddenly faced a shortage of a gas that was formerly abundant.

3. Lead-Time Vulnerability — When Orders Must Precede Demand Knowledge​

Some components have such long lead times that companies must order before knowing actual demand. This creates two problems:

Over-ordering when demand surges: During the 2020 pandemic surge in demand for electronics, manufacturers saw orders jump. They over-ordered components with 12-week lead times, thinking the surge was permanent. By the time components arrived (12 weeks later), demand had normalized. They were stuck with excess inventory.

Under-ordering when demand accelerates unpredictably: Conversely, if demand surges unexpectedly, companies can't instantly access more supply. They're locked into previous orders. When auto demand recovered unexpectedly in 2021 (after the 2020 crash), manufacturers couldn't get chips because they'd under-ordered 3 months earlier.

Long lead times force forecasting, and forecasts are wrong. The longer the lead time, the worse the forecast error compounds.

4. Information Opacity — The Cascade Invisible Until It's Too Late​

In complex supply chains, downstream companies don't see upstream disruptions until the disruption has traveled all the way down.

Example: The 2021 Texas Freeze In February 2021, extreme cold disabled natural gas production in Texas. This affected:

  • Petrochemical plants (which need gas for production and steam)
  • Plastic resin producers (which feed from petrochemicals)
  • Automotive plastic component suppliers (which need resin)
  • Auto manufacturers (which need components)

But the disruption was invisible at the auto-manufacturer level for weeks. Auto makers knew they had parts in inventory. They didn't know that upstream, the entire supply pipeline was stopped. By the time they realized the disruption, 6 weeks had passed and parts were nowhere to be found.

Companies can't see disruptions upstream because they don't have visibility into supplier facilities. They only see: "Our orders are arriving on time" or "Our orders are late." By the time orders are late, the disruption has already propagated for weeks.

How Concentration Creates Single Points of Failure​

A single point of failure (SPOF) is a component or supplier whose failure stops the entire system.

Example: Advanced Lithium-Ion Battery Production​

Modern electric vehicles require advanced lithium-ion batteries. As of 2023:

  • CATL (Chinese company) produces ~35% of the world's EV batteries
  • BYD (Chinese company) produces ~20%
  • LG Energy Solution (South Korean) produces ~15%
  • Remaining suppliers split ~30%

If CATL faced a major disruption, EV production worldwide would drop 35%. If CATL and BYD both failed, EV production would drop 55%.

This concentration exists because:

  • Battery manufacturing requires massive upfront capital
  • Economies of scale are extreme (bigger factories = lower cost per unit)
  • Learning curves are steep (early-mover advantages)
  • Raw materials (lithium, cobalt) are concentrated geographically

So the industry consolidates around a few low-cost producers. The industry becomes more efficient. And fragile.

Historical Precedent: The Yogurt Shortage​

In 2014, a bacterial contamination was discovered in yogurt produced by Straus Family Creamery (a major U.S. organic yogurt producer). The contamination was limited to a specific strain and posed minimal health risk, but the company faced lawsuits and reputational damage.

Straus had to recall products and temporarily cut production. Retailers experienced yogurt shortages because Straus was a major supplier of organic yogurt to natural/organic chains.

This wasn't a supply-chain crisis, but it illustrates the point: if one supplier controls a significant share, their failure creates immediate shortages.

How Long Lead Times Enable Disruptions to Cascade​

Lead times are the time between ordering and receiving goods. In global supply chains, lead times are often 8–16 weeks.

The Problem with Long Lead Times​

If you order a component today, it arrives in 12 weeks. During those 12 weeks:

  • Demand might change
  • A supplier might experience a disruption
  • A geopolitical crisis might emerge
  • Your own demand forecast might prove wrong

But you can't cancel the order or change the quantity because the components are already in transit or manufacturing.

Concrete example: The 2021 Semiconductor Shortage

Demand for chips surged in late 2020 (pandemic demand spike). Manufacturers looked at their lead times (12 weeks) and orders in the pipeline (old orders based on normal demand). They realized they were under-ordered. They placed huge new orders with TSMC and other foundries.

But TSMC was already at full capacity. The new orders weren't fulfilled for 16-20 weeks (longer than the normal 12 weeks because the fab was overbooked). During those 16 weeks:

  • Demand would change
  • Supply constraints would ease or tighten
  • New disruptions would emerge

By the time the chips arrived, some of the demand had shifted to different products. Some buyers had shifted to different suppliers. Some manufacturers had over-committed and now had excess chip inventory.

The long lead time meant manufacturers had to guess what demand would be months in the future. They guessed wrong.

The Invisibility Problem: Why Crises Seem to Appear Suddenly​

One of the most frustrating aspects of supply-chain fragility is that disruptions seem to appear out of nowhere.

Company executives say things like: "We had no warning" or "It hit us suddenly." In reality, the disruption was propagating for weeks, but they couldn't see it because they lack upstream visibility.

The Information Cascade Delay​

When a supplier shuts down:

  • Day 1: Supplier closes. Upstream players (the raw-materials suppliers to that supplier) notice immediately.
  • Days 1-7: The supplier's immediate customers (the company's distributors) know there's a problem, but they have 1-2 weeks of inventory, so they're not panicked.
  • Days 7-14: The supplier's customers start drawing down inventory and requesting expedited shipments, but many don't realize the root problem (they think it's a temporary delay).
  • Days 14-21: The company receives late shipment notifications from its suppliers. It now knows there's a problem.
  • Days 21-28: The company's own production starts to slow as it exhausts inventory.
  • Days 28-42: The company's customers (retailers, other manufacturers) notice delays.

From the company's perspective, the disruption seems to appear on Day 21. In reality, it started on Day 1, but information took weeks to cascade through the supply chain.

By the time the company realizes there's a problem, the shortage is already baked in, and there's nothing they can do to recover.

Real-World Examples of Supply-Chain Fragility​

The Fukushima Earthquake (2011)​

Japan experienced a 9.0-magnitude earthquake followed by a tsunami. It killed ~18,000 people and caused $235 billion in damage.

For supply chains, the key impact was on suppliers:

  • A factory producing 60% of the world's airbags was damaged
  • A factory producing 25% of the world's touch screens was damaged
  • Numerous semiconductor and automotive suppliers were disrupted

Auto manufacturers worldwide had to shut down plants because they couldn't get airbags or other components. Disruptions lasted 6-12 months. The financial cost was in the tens of billions.

This was a natural disaster, so it was unavoidable. But the supply-chain fragility amplified the damage. A single disrupted supplier could shut down global auto production because companies had no backup suppliers.

The 2022 Suez Canal Blockage​

On March 23, 2022, the Ever Given (a massive container ship) got stuck crosswise in the Suez Canal, blocking all traffic.

For 6 days, shipping traffic was backed up. Hundreds of ships waited in queues. Supply chains everywhere experienced delays:

  • Automotive supplies were delayed
  • Consumer goods were delayed
  • Electronics were delayed

The blockage was only 6 days. But because ships were delayed by a week, they missed scheduled arrival times. Retailers had to expedite other shipments, paying premium shipping costs. Manufacturers had to adjust production schedules.

The 6-day blockage created ripple effects that lasted weeks. This illustrates the cascading nature of supply-chain disruptions.

The 2020-2022 Chip Shortage​

The semiconductor shortage combined multiple failure modes:

  1. Concentration risk: TSMC produces 54% of the world's advanced chips
  2. Lead-time vulnerability: Chip orders have 12-20 week lead times
  3. Geographic risk: Taiwan is geopolitically isolated and faces natural disaster risk
  4. Information opacity: Chip shortages weren't visible until they were severe (by which time they'd been propagating for weeks)
  5. Demand surge: Pandemic demand spiked unexpectedly, and everyone over-ordered, exhausting capacity

The result was a 24-month shortage that disrupted automotive, consumer electronics, and industrial production worldwide.

How Companies Can Reduce Fragility (Without Eliminating It)​

Companies can reduce fragility, but not eliminate it. The cost of total resilience is prohibitive.

1. Dual Sourcing for Critical Components​

For components that are critical and have long lead times, maintain two suppliers. This adds cost but provides insurance.

2. Geographic Diversification of Suppliers​

Don't rely on suppliers in a single region. Spread suppliers across geographies so a natural disaster in one region doesn't stop everything.

3. Shorter Lead Times​

Work with suppliers to reduce lead times. Shorter lead times mean less forecasting error and better ability to react to disruptions.

4. Supply Chain Visibility​

Invest in systems to see what's happening upstream. Real-time monitoring of supplier status lets you spot disruptions faster.

5. Safety Stock for Critical Components​

Maintain buffer inventory for components with long lead times or single suppliers. The carrying cost is worth the insurance.

6. Supplier Financial Health Monitoring​

Weak suppliers are more likely to fail. Monitor supplier financial health and diversify away from suppliers with weak balance sheets.

FAQ​

Isn't it obvious that companies should diversify suppliers?​

It seems obvious, but the incentive structure is wrong. A single supplier allows companies to negotiate lower prices due to volume commitments. Dual suppliers command higher prices. For a company with 3% profit margins, the 5-10% cost difference from diversifying eats up the entire margin.

Companies diversify only when forced to by crisis or regulation.

Why don't governments mandate supply-chain resilience?​

Some do (defense contractors, critical infrastructure). But for most industries, the cost would be passed to consumers. A 10% increase in input costs translates to higher prices. Governments are reluctant to mandate this unless national security is directly threatened.

Can artificial intelligence fix supply-chain fragility?​

AI can improve demand forecasting and inventory optimization, reducing some fragility. But it can't eliminate geographic concentration or single points of failure. An earthquake or geopolitical crisis will still disrupt supplies, no matter how good your AI is.

Will reshoring fix supply-chain fragility?​

Partially. Reshoring (moving production closer to markets) reduces transportation lead times and geographic risk. But it increases costs. Reshoring is happening in some industries (semiconductors, pharmaceuticals), but it's slow and expensive.

Summary​

Supply chain fragility is fundamentally a result of optimization for efficiency. Companies reduce costs by concentrating production, relying on single suppliers, and maintaining minimal inventory. These choices are rational individually but create collective vulnerability.

The fragility manifests through geographic concentration (Taiwan chips, Japanese auto parts), supplier concentration (TSMC's dominance), long lead times (forcing speculative ordering), and information opacity (disruptions take weeks to cascade downstream).

Reducing fragility requires higher input costs, lower profit margins, and more capital tied up in inventory. Most companies accept the fragility rather than pay for resilience. Crises (pandemics, earthquakes, wars) force temporary shifts toward resilience, but the underlying incentives eventually drive companies back toward efficiency-optimized, fragile structures.

Next​

→ The 2020-22 supply chain crisis