The Big Picture
T-MEC utilization jumped from 44.8% to 88.7%. Regulatory compliance changed. Supply chains did not. Trace the inputs inside any pump or valve and you find the exact point where regional integration breaks.
Automotive assembles $310B regionally. Drivetrain electronics come from East Asia. Aerospace: $93B in final assembly, avionics and precision alloys 40-60% non-regional. Medical devices: $45B, semiconductor components almost entirely imported.
Final goods stay in the bloc. Subsystems don't. Every sector, same shape.
Automotive's scale obscures the mechanism. Thousands of codes, hundreds of flows, every data point contested.
Pumps and valves: $4.3B. Small enough to trace every input. Large enough to matter. If the bloc cannot close the gap here, it will not close it at $310B.
2026: the T-MEC joint review. Tighten, relax, or restructure rules of origin. That decision needs supply chain facts, not assumptions.
Start with the chain itself.Factory, power plant, refinery, data center. Pumps move fluids. Valves control flow. Not niche. Infrastructure.
Motors for force. Converters for power regulation. Circuits for control logic. Seals for containment. Wiring for connection. Data processing for management.
Each one, a different layer of complexity.
Gaskets and wiring: 43-50% non-regional. Converters: 83%. Processor circuits: 96%.
Higher complexity, wider deficit. Mexico imports $18.76B in processor circuits. Ninety-six percent from outside North America.
This gradient is not unique to Mexico.Mexico exports 93% of its pumps and valves to the bloc. Sources inputs 80-96% outside it. Canada exports 76-95% regionally. Sources converters 67% and circuits 81% outside. The United States: 98.8% of its processor circuits from outside. 83.5% of its converters.
Not a Mexican problem. A North American one.
Gaskets: 37-50% non-regional across all three. Bulk, low-value, proximity works.
Processor circuits: 80-99%. Converters: 67-84%.
Rules of origin require regional content. The region does not supply it where it matters most.
Three failures. Three mechanisms.Capacity exists on both sides. Mexico exports DC motors north. The US makes AC motors. Yet Mexico imports $1.66B in motors, 80-87% non-regional. China supplies $667M. The US: $277M.
The bottleneck is not production. It is verification. Pentair, Parker Hannifin, and Illinois Tool Works each disclose 40,000+ global suppliers in their 10-K filings. Regional verification at that scale is not a process improvement. It is a structural barrier.
The region makes converters. But concentrated. No redundancy. Mexico: $4.06B, 83% non-regional. US: $17.11B, 84%. Canada: $2.52B, 67%.
One or two facilities per country. Disruption breaks supply. Concentration reinforces itself.
No regional substitute at scale. Mexico: $18.76B, 96%. US: $28.25B, 99%. Canada: $1.11B, 81%. Combined: $48.1B, 96% from outside.
There is no regional supplier to verify. Tightening rules penalizes exporters with zero substitution options.
The three failures have geography.Coahuila: $220M in pumps and valves. Thin upstream. Sonora: $195M. Limited input base. They integrate the product. They import what makes it work.
Jalisco: $8.6B in upstream inputs, 13 HS4 categories. $43M in pumps and valves. CDMX: $1.5B upstream, $143M final goods.
They make what the region needs. Not connected to the assemblers who could absorb it.
Chihuahua: $798M final goods, $11.3B upstream, 15 HS4 categories. Nuevo Leon: $639M final, $6.4B upstream, 16 categories.
Assembly and upstream in the same place. Traceability simpler. Verification cheaper.
Ontario: $880M in pumps and valves. Honda's CAD $15B EV supply chain commitment. Upstream capability building but T-MEC utilization at only 53% in 2025. Capacity exists. Linkages do not.
Proximity helps. It does not explain who captures the preference.The T-MEC preference means margin, volume stability, contractual priority. Qualifying means paperwork: input traceability, process audits, compliance files. That cost is fixed. Same paperwork for a $500K shipment and a $50M one. The Federal Reserve estimates compliance costs at 1.4-2.5% of shipment value, equivalent to a tariff. Some suppliers declare inputs "non-originating" rather than complete verification. Large firms spread it. Small firms absorb it whole.
OEM audits a supplier. Switching means starting over. Incumbents protected. Alternatives shut out. Cost of proving arrives before volume. Depends on the relationship. Repeats with every new buyer.
Non-portable verification blocks new contracts. No contracts means no production runs. No runs means no process learning. No learning means the capacity gap persists. That persistent gap then justifies continued extraregional sourcing. A reinforcing loop.
Failures are distinct. Instruments must be too.Motors and seals. SAT, CBP, and CBSA jointly recognize certifications. Shared testing labs. Portable compliance files reusable across buyers under a common trilateral standard. Rules of origin stay. The infrastructure for demonstrating compliance changes.
These instruments address the supply chain. The stakes extend beyond it.Assemble without producing subsystems and the learning stays where the component is made. Not where it is integrated. Employment without capability accumulation.
Three failures. Three instruments. One review cycle to deploy them.