Home Scenario Lab Materials Watch Bottlenecks Briefs Policy Monitor Risk Map AI Tools Pricing Get Access
MEXEL METHODOLOGY

The Constraint Chain.
How we think.

Most intelligence stops at the headline. Mexel maps the full chain from headline to hidden constraint to market signal — so you can act before the consensus catches up.

01
Headline
The visible event
→
02
Hidden Bottleneck
The underappreciated physical constraint
→
03
Transmission Chain
Step-by-step propagation
→
04
Sector Implication
Who is exposed and how
→
05
Decision Signal
The trigger to watch
Example Constraint Chain — AI Infrastructure & Grid
Headline
AI data-centre build-out is accelerating
Obvious explanation
Power demand is rising — utilities need more capacity
↓ Hidden constraint
Large power transformers / grain-oriented electrical steel / high-voltage switchgear — lead times have extended to 24–48 months
Process reason
LPT manufacturing is a global oligopoly; GOES mills book 12–18 months forward; qualified EPC crews are fully booked. None of these can be resolved in under 2–3 years.
Transmission chain
AI load growth → Grid interconnection requests surge → Transformer order backlog >36 months → Utility interconnection queue extends → Data-centre commissioning delayed 1–3 years
Trigger to watch
Transformer lead times rising above 36 months in OEM earnings commentary
What would prove this wrong
New Western LPT factories announced and funded; AI demand profile shifts to nuclear or storage
BOTTLENECK KNOWLEDGE LAYER

22 canonical
constraint cards.

Each card maps a hidden bottleneck — its process role, failure mode, supply concentration, lead-time risk, and decision triggers. Updated as the landscape shifts.

Filter by category
FEATURED CONSTRAINT CHAINS

Three chains in detail.

Semiconductor
The Fluorine Chain
China controls 55–60% of acid-grade fluorspar. Every fluorinated semiconductor input — HF, WF6, NF3, LiPF6, UF6 — traces back to acidspar. A single export-quota tightening hits fabs, batteries, and nuclear simultaneously.
China fluorspar controls→ HF supply pressure→ WF6 / NF3 / LiPF6 constraints→ Fab throughput + battery electrolyte margins
Model this scenario →
Grid Equipment
AI Load → Grid Bottleneck
AI data-centre load growth is driving transformer orders at rates not seen since 2002. LPT lead times have quietly extended to 24–48 months. The commissioning delay is a physical constraint that cannot be resolved with capital alone.
AI load growth→ Interconnection queue→ Transformer / GOES backlog→ Data-centre commissioning delayed 1–3 years
Model this scenario →
Industrial Mfg
Gas Turbine Duopoly
GE Vernova and Siemens Energy control large-frame gas turbine supply. Order books extend past 2027–2028. The constraint is not design or permitting — it is single-crystal blade casting capacity at Howmet and Precision Castparts.
Firm capacity demand→ OEM order saturation→ Blade casting bottleneck→ Power capacity unavailable until 2028–2030
Model this scenario →
Take it further
Bottleneck intelligence,
in your workflow.

Model scenarios, generate rapid response memos, and map constraints across sectors — with Mexel's AI tools pre-loaded with these cards.