of global semiconductor fab capacity sits in four countries that all score moderate-to-high on climate vulnerability.
Your climate risk map and your compliance risk map should be the same document.
Every bubble is a country. The further right, the more climate-vulnerable. The higher up, the more semiconductor capacity at risk. The bigger the bubble, the more chips it outputs.
Each cell shows a country's exposure per risk dimension — physical climate hazards on the left, regulatory and structural pressure on the right. The right-hand column counts how many dimensions sit at elevated exposure or higher.
Reading the table: Hazard cells are scored from each country's documented hazard profile weighted by ND-GAIN vulnerability. Regulatory cells reflect UFLPA entity-list exposure (CBP) and CSDDD scope (EU 2024/1760). Concentration reflects share of global fab capacity and output (SEMI). Exposed counts dimensions at severity ≥3 on the 0–5 scale. Green cells mean the country drives the rule instead of being exposed to it: ENF = enforcing jurisdiction (the U.S. enforces UFLPA at its border), REG = regulator/author (EU member states wrote and transpose the CSDDD).
The industry's fix is making the problem worse. Diversification destinations are more climate-vulnerable than the countries they're meant to replace.
The industry is moving supply chains from moderate-vulnerability countries to high-vulnerability countries. Vietnam's ND-GAIN vulnerability score is 23% worse than China's. India's is 27% worse. That is not derisking. That is trading one category of exposure for another — and calling it a strategy.
A Kraljic-style positioning grid adapted to convergence risk: business impact on one axis, convergence exposure on the other. The quadrant your sourcing countries land in suggests the response strategy.
Reading the grid: Taiwan sits deep in STRATEGIC — highest value at stake, highest disruption exposure, no substitute at its technology node. The diversification destinations (Vietnam, India, Philippines) cluster in BOTTLENECK: modest business impact today, but disproportionate convergence risk. Positions derive from the same data as the Convergence Risk Score.
Each country scored on three dimensions: climate vulnerability, regulatory pressure, and supply chain concentration. Click a card for details — and set your own weights below: your risk appetite, your ranking.
Enter where you source from and roughly how much. CONVERGE scores your footprint with the same method as the country profiles — including whatever CRS weights you set above. Everything runs in your browser: nothing is stored, nothing is sent anywhere.
Select a country to see the cascading impact on global semiconductor supply.
The regulations targeting supply chains are accelerating — and they target the same countries where climate risk is highest.
Major supply chain disruptions from 2011 through July 2026 — each with a source link — and the lessons they encode for semiconductor risk.
Every disruption in the record below followed the same arc. The visible damage ends early — the regulatory exposure peaks last, because emergency sourcing decisions get made faster than due diligence can follow.
This is CONVERGE's core claim in one sentence: the scramble is where climate risk turns into compliance risk — which is why the two risk maps have to be the same document.
Trend analysis from historical patterns, climate projections, and regulatory acceleration curves.
CBP reviewed 7,325 shipments in FY2025 (up ~51% YoY) — more than 18,000 shipments worth ~$3.81B since June 2022, with only ~6.5% of reviewed shipments released. In 2026 the FLETF added lithium, steel, copper, aluminum, and PVC as high-priority sectors: enforcement is expanding from solar into the entire electronics materials chain.
EU Corporate Sustainability Due Diligence Directive requires supply chain human rights and environmental due diligence. Companies with 1,000+ employees and €450M+ turnover must comply first. Semiconductor OEMs with EU revenue are in scope.
Taiwan produces 90%+ of advanced chips (<7nm). The stress tests keep coming: Dec 2025 M7.0 quake (fourth strong quake near the fab belt since Nov 2025), Super Typhoon Bavi skirting the island in Jul 2026, 2021's worst-in-56-years drought. Seismic design has held every time — but no historical precedent exists for a full disruption at this scale. The Thailand 2011 analog would understate impact by 10-50x.
Companies diversifying from Taiwan/China to Vietnam, India, and Philippines are trading concentration risk for climate vulnerability. Average ND-GAIN vulnerability of emerging destinations (0.466) exceeds established hubs (0.386). Risk is being redistributed, not reduced.
This risk stopped being predictive in February 2026: the Strait of Hormuz closure (~20% of global oil) stranded ~2,000 ships, pushed Brent to a $126 peak, and quadrupled insurance premiums — the largest energy supply disruption since the 1970s. A June reopening deal remains contested as of mid-July. The 2024 Red Sea crisis already showed the cascade mechanics: Cape of Good Hope rerouting added 10-14 days, container rates spiked 200-300%, and demand signals amplified 3-4x through the tiers (bullwhip). Energy-intensive fabs across Asia are absorbing the cost shock now.
How disruptive each hazard typically is in each month across the semiconductor manufacturing belt, scored 0–5. The dangerous months are where several rows darken at the same time.
Reading the heatmap: Every cell is a severity band from 0 to 5 — how disruptive that hazard historically is in that month (0 = no activity, 5 = peak season / hard deadline). Cells at 4 or 5 show their number; the current month is outlined. Jul–Oct is the structural convergence window: NW Pacific typhoon season peaks over Taiwan, Philippines, and Japan at the same time as South Asian monsoon flooding and Thailand's flood season — while regulatory deadlines (CSDDD July 2027) land in the same quarter. Severity bands are indicative, compiled from historical event frequency in the disruption record below.
The window between now and Q3 2027 represents a critical transition period: UFLPA enforcement intensifying, CSDDD compliance deadlines approaching, climate events increasing in frequency, and maritime chokepoints under persistent geopolitical pressure. Each risk domain reinforces the others — a climate event at a chokepoint triggers shipping delays that trigger bullwhip effects that trigger regulatory scrutiny. Companies without supply chain climate risk intelligence by mid-2027 face regulatory exposure on multiple fronts simultaneously.
Every number traces to a named primary source. No LLM-generated statistics.
ND-GAIN Country Index (2023), University of Notre Dame. 185 countries scored on 36 indicators across food, water, health, ecosystem, habitat, and infrastructure. Higher vulnerability score = less prepared.
SEMI World Fab Forecast (2025-2026). Installed fab capacity by country/region. Output percentages weighted by wafer starts and technology node.
UFLPA enforcement: CBP Statistics — 18,000+ shipments (~$3.81B) reviewed since June 2022; FY2025: 7,325 shipments (+51% YoY); 2026 FLETF priority sectors: lithium, steel, copper, aluminum, PVC. CSDDD: EU Directive 2024/1760 + Omnibus I amendments (Directive 2026/470).
Composite metric: CRS = (Climate × w₁) + (Regulatory × w₂) + (Concentration × w₃). Baseline weights 0.4/0.3/0.3 — but the Risk Profiles view lets you set your own (plus an entropy-based data-driven preset), and every ranking is robustness-tested across 120 weighting combinations. Climate uses ND-GAIN vulnerability normalized 0-100; Regulatory scores UFLPA entities, CSDDD scope, enforcement intensity; Concentration uses fab capacity + output share.
Thailand 2011 floods: World Bank ($46.5B losses), Swiss Re (2021 retrospective). 25% of global HDD production disrupted. Recovery took approximately two years.
Fetched client-side at page load, no backend: Open-Meteo 7-day forecast (max daily wind gusts, precipitation) for 15 representative fab/industry clusters, GDACS (EU/UN) Orange and Red disaster alerts, and USGS M5.5+ earthquakes from the past 7 days — each within 300 km of a site. Thresholds: gusts ≥90 km/h (storm) / ≥118 km/h (hurricane force); rain ≥100 / ≥180 mm/day; quakes M5.5+ flag / M6.5+ critical. Forecasts are model output, not observations — a signal to look closer, not a verdict. If the APIs are unreachable, the page falls back to the static view.
Built as part of Terra.do "AI for Climate: Practical Tools" (Cohort 2). Data verified through a four-stage pipeline: Research → Structure → Narrate → Produce. Built with Chart.js, vanilla JS, and deployed on Vercel.