A 3 MW rooftop and carport solar array on a Mexican manufacturing site is a well understood piece of engineering. The modules, inverters, and racking are commodities. Financing them against a CFE bill that runs to USD 1.7 to 1.9 million a year for a continuous 2 MW plant is straightforward math. Demand and distribution charges alone reach roughly 475 MXN per kW-month in the Aguascalientes GDMTH schedule as of July 2026 (CFE tariff schedule, July 2026). The array pays for itself. That part is easy.
The part that stops projects is the point where the array meets a substation built in the 1990s for one-way power flow. That substation was designed to take energy from the grid and step it down to the plant. It was never designed to sit still while a behind-the-meter generator pushes current back toward the transformer, trips protection relays set for a single direction, and raises the short-circuit duty at a bus that already sits near its interrupting rating. The interconnection, not the array, is the binding constraint.
Our thesis is simple. The interconnection study has to come before the module order, not after. An older substation with limited short-circuit headroom, protection coordinated for radial flow, a transformer sized only for load, and open Código de Red obligations can block or de-rate a project that penciled out cleanly on a spreadsheet. The engineering checks below decide whether the number in the finance model survives contact with the physical asset.
What changed, and what did not
Two things moved the ground under industrial self-generation, and it is worth separating them because they operate on different tracks.
The first is the interconnection rulebook, which predates the current reform and still governs the physical connection. Mexico runs two interconnection manuals. Plants under 0.5 MW connecting to the distribution network follow the Manual de Interconexión de Centrales de Generación con Capacidad menor a 0.5 MW, published in the DOF on December 15, 2016. That route is light. It carries no CRE generation permit, it uses a standard interconnection contract with the distributor, and it is the regime most rooftop and small carport arrays sit under. Plants at or above 0.5 MW, or those connecting at a level that requires network studies, fall under the Manual para la Interconexión de Centrales Eléctricas y Conexión de Centros de Carga, published in the DOF on February 9, 2018. That second route requires interconnection studies run through CENACE, and those studies are where short-circuit duty, load-flow, and protection coordination get tested against the real network.
The second change is the permit and market regime. The Ley del Sector Eléctrico was published in the DOF on March 18, 2025, replacing the 2014 Ley de la Industria Eléctrica, with its reglamento published October 3, 2025 and effective October 4, 2025. Under the LSE the self-supply generation-permit threshold rose from 0.5 MW to 0.7 MW, and interconnected self-supply projects between 0.7 and 20 MW now move through a simplified permit administered by the Comisión Nacional de Energía (LSE and reglamento, DOF 2025). Surplus energy is sellable only to CFE, and the old autoabasto legacy structure is closed to new entrants.
The practical read is this. The permit threshold tells you which piece of paper you need. The interconnection manuals tell you whether the copper will hold. A project can clear the 0.7 MW permit path and still fail the CENACE interconnection study on short-circuit or protection grounds. The two are not the same gate.
What an interconnection surprise costs
An interconnection problem never arrives labeled as one. It shows up as capex you did not budget, as a de-rated array that undercuts the savings case, or as a schedule slip that pushes the whole return to the right.
Start with the transformer. A legacy plant substation transformer was sized for load, with a service factor for growth, and nothing more. A behind-the-meter array that back-feeds during a low-load shift, a weekend, or a holiday can push reverse power through that transformer above its nameplate. If the CENACE study or the distributor flags reverse-flow limits, the fix is either a transformer upgrade or an export-limiting control scheme that caps the array's output. Export limiting is cheaper, and it also quietly caps the energy yield that justified the investment. That is the de-rate that finance rarely sees coming. An array modeled at a 22% capacity factor that gets curtailed on weekends does not deliver the kWh the model assumed.
Short-circuit headroom is the second cost. Every inverter adds fault current, modestly per unit but cumulatively across a large array. If the plant's main bus or switchgear already sits near its interrupting rating, the added contribution can push it over, and the remedy is new switchgear or a current-limiting reactor. Medium-voltage switchgear replacement on a live industrial site is a seven-figure line and a multi-week outage, not a change order.
Protection is the third. Relays on a radial substation are set to see fault current flowing one way, from the grid into the plant. Add a generator behind the meter and the fault-current picture becomes bidirectional. Coordination has to be restudied, and the array needs interconnection protection that disconnects it cleanly when the grid goes away, so it never energizes a dead line. Getting this wrong is not a cost question, it is a safety and liability question, and it is the single most common reason a distributor or CENACE withholds sign-off.
Put in ranges a CFO can carry into a budget meeting. A clean interconnection on a substation with headroom adds a modest protection and metering scope to the project. A constrained one can add transformer or switchgear capital that runs into the hundreds of thousands of dollars, plus an outage, and can shave a fifth or more off the deliverable energy if export limiting is the accepted fix. The study that tells you which case you are in costs a fraction of either outcome and takes weeks. Ordering modules before that study is buying an asset you cannot yet size.
The interconnection checklist
The table below is the set of physical checks that decide the project. Run these before committing capital, not after the panels arrive. The study and the protection review sit on the critical path ahead of procurement, which puts the schedule risk at the front of the job.
| Parameter | Why it matters | Who signs off | Typical blocker |
|---|---|---|---|
| Interconnection route (under 0.5 MW distribution vs 0.5 MW and above with studies) | Sets whether a full CENACE study applies | CFE Distribución or CENACE, per manual | Assuming the light route when array size forces the study route |
| Self-supply permit (0.7 MW threshold, 0.7 to 20 MW simplified) | Determines CNE permit path under the 2025 LSE | Comisión Nacional de Energía | Treating the permit as the only gate and skipping the physical study |
| Transformer capacity and reverse power | Legacy unit sized for load, not back-feed | CENACE / distributor study | Reverse flow above nameplate on low-load shifts |
| Short-circuit duty at the bus | Inverters add fault current to aging switchgear | CENACE interconnection study | Existing gear already near interrupting rating |
| Protection coordination | Radial relays assume one-way flow | Plant protection engineer + CENACE | Miscoordination once flow is bidirectional |
| Anti-islanding and disconnection | Array must not energize a dead grid | Plant engineer, IEEE 1547 as reference | Inadequate interconnection protection |
| Código de Red compliance | Power factor, harmonics, voltage, metering | CRE / CFE compliance review | Inverter harmonics or power-factor drift at the point of common coupling |
Why interconnection comes first
Treating this as a routine engineering step is the error. The grid the array connects to is tight and getting tighter. CENACE declared two states of emergency on February 18, 2026, when operating reserves fell below 3% against the 6% target and 21 states saw supply interrupted. A Tamaulipas plant had tripped as cloud cover cut solar output (Mexico News Daily, February 20, 2026, citing CENACE). CENACE's own summer 2026 forecast put demand near 54,000 MW with a worst-case operating reserve margin around 7%, described as apretado by CENACE director Ricardo Mota Palomino, reported April 16, 2026.
A grid running that close to its limit is a grid where the distributor and CENACE scrutinize every new injection point. A plant that back-feeds badly, with poor power factor or harmonic distortion, draws regulatory attention rather than avoids it. The Código de Red, the grid code first published by CRE in the DOF in April 2016, sets the power-factor, harmonic, voltage, protection, and metering obligations that a behind-the-meter generator has to meet at the point of common coupling. An array that pushes the site's power factor or total harmonic distortion outside the code's limits is not a compliance footnote. On a stressed grid it is the kind of thing that gets an interconnection request slowed or a penalty applied.
If nobody has checked your inverter specification against the Código de Red power factor and harmonic limits at your own point of common coupling under real plant load, that is the item most likely to stall the interconnection request. Talk to an advisor.
There is a demand-side reason to move as well. Transmission expansion has run far behind load. Demand grew 3.4% in 2022 and 3.5% in 2023 against transmission expansion of 0.09% and 0.10%, and CFE's 2023 grid investment came in near 21% of the level PRODESEN recommended (IMCO). Behind-the-meter solar is one of the few ways an industrial site can add supply without waiting on that transmission. The value is real. It is only capturable if the interconnection clears.
The risk: a newer substation changes this
This constraint is site-specific, and plenty of sites do not carry it. A plant with a recently upgraded substation, spare transformer capacity, switchgear with short-circuit margin, and a modern protection scheme may connect a multi-megawatt array with only a protection relay addition and metering. For those sites the study confirms headroom quickly and the module order follows without drama. Front-loading the study still costs weeks and a study fee that turns out, in hindsight, to have been insurance you did not need. That is a fair objection.
The reason the sequence still holds is asymmetry. The downside of studying first is a few weeks and a modest fee. The downside of ordering first and studying second is stranded capital, an unplanned switchgear outage on a live plant, or a curtailment scheme that guts the savings case after the modules are already on the roof. When the cost of being early is small and the cost of being wrong is large, you pay for the study.
The second risk is regulatory drift. The LSE reglamento took effect October 4, 2025, and CNE and CENACE are still settling the operational detail of the simplified permit path. Interconnection queue timing under the new institutions is not yet a published, reliable number, and on our read a project should build schedule contingency around it instead of assuming the pre-reform timelines hold. If your model depends on a fast interconnection approval, that assumption is the soft spot, not the hardware.
The order of operations before you buy
Commission the interconnection study before you issue the purchase order for modules and inverters. Size the array to what the study says the substation can accept, not to what the roof can hold. If the study shows reverse-flow or short-circuit constraints, price the two fixes side by side. A transformer or switchgear upgrade preserves full energy yield at higher capex and an outage. An export-limiting control scheme avoids the capital and the outage but caps the kWh, so model the curtailed yield honestly and rerun the payback before you accept it.
Match the paperwork to the route. If the array lands under 0.5 MW on the distribution network, the December 2016 manual and a distributor interconnection contract are the path, and no generation permit applies. If it is larger, plan for the February 2018 manual, the CENACE studies, and the CNE simplified permit that applies to self-supply between 0.7 and 20 MW under the 2025 LSE. Do not let the lighter permit threshold lull the team into skipping the physical study. They are different gates.
Specify interconnection protection and anti-islanding to a recognized engineering reference. IEEE 1547-2018 is the standard the industry uses for this. It requires the resource to cease to energize the grid within two seconds of an unintentional island forming, and it defines abnormal-condition ride-through in three categories, I, II, and III, of increasing coordination with bulk-system needs (IEEE 1547-2018). Mexican interconnection is governed by the CENACE manuals and the Código de Red, and IEEE 1547 is the engineering benchmark most inverter and protection vendors already design to, which makes it the practical reference point for the protection scope.
Close the Código de Red items in the same design pass. Confirm the inverters hold power factor and harmonic distortion inside the code's limits at the point of common coupling under the plant's real load profile, not at a lab condition. On a grid that hit sub-3% reserves in February 2026, the compliance margin at the connection point decides whether the interconnection request keeps moving or gets flagged.
The timing pressure is straightforward. The savings case is strong and the grid is tight enough that self-generation is worth doing. The interconnection is the item on the critical path with the longest tail and the least certainty. Start it first, and the module order becomes a decision you can actually size.
Solar and BESS Feasibility Study
We run the checks that decide the project, meaning transformer reverse-flow headroom, short-circuit duty at the bus, protection coordination and the open Código de Red items, and we confirm which interconnection route your array falls under. It takes your single-line diagram, substation nameplate data and twelve months of CFE bills.
Talk to a Mexico Energy Partners advisor or email info@mexicoenergypartners.com. Mexico Energy Partners sells no equipment and is compensated only by the client.
Sources
- CFE tariff schedule (GDMTH, Aguascalientes), July 2026.
- Manual de Interconexión de Centrales de Generación con Capacidad menor a 0.5 MW, SENER, DOF, December 15, 2016.
- Manual para la Interconexión de Centrales Eléctricas y Conexión de Centros de Carga, SENER, DOF, February 9, 2018.
- Ley del Sector Eléctrico, DOF, March 18, 2025, and its Reglamento, DOF, October 3, 2025 (effective October 4, 2025).
- Código de Red (Disposiciones administrativas de carácter general en materia de eficiencia, calidad, confiabilidad, continuidad, seguridad y sustentabilidad del Sistema Eléctrico Nacional), CRE, DOF, April 2016.
- IEEE Std 1547-2018, IEEE Standard for Interconnection and Interoperability of Distributed Energy Resources with Associated Electric Power Systems Interfaces.
- CENACE, states of emergency, February 18, 2026, reported by Mexico News Daily, February 20, 2026.
- CENACE summer 2026 demand and reserve-margin forecast, Ricardo Mota Palomino, reported April 16, 2026.
- IMCO, analysis of Mexican electricity demand growth versus transmission expansion, 2022 to 2023.