Mexico's new solar rules push mid-market growth behind the meter
Latin America · Amaru · Mexico · 2026-07-30 · Likelihood: Likely
Forecast: The real opportunity is on-site solar in the 0.5 to 5 MW commercial band.
Risk is paperwork, not law. Rising Comisión Federal de Electricidad tariffs drive demand. Grid connection and surplus payment terms decide whether growth holds.
What this changes for you
- Capital. Your equity return depends on whether CFE tariffs keep rising and whether the peso holds, two variables that can compress paybacks without any change to the reform text.
- Operations. Your interconnection queue and per-plant reliability rulings from CENACE now determine whether surplus injection is economic or whether storage becomes a mandatory cost line on every project.
- Positioning. Holding equity in a USMCA-covered generation asset gives you recourse that a vendor or services model does not, and concentrating in high-tariff industrial states maximises the tariff arbitrage that drives demand.
Drivers
Permit-light corridor now exists — A new band opened that requires only CFE interconnection. — The distributed generation ceiling rose from 0.5 MW to 0.7 MW with no generation permit required. Industrial on-site self-consumption extends to 20 MW. Distributed capacity reached 5,437 MW at end-2025, a 22.2% increase from 4,447 MW at end-2024.
Tariff arbitrage, not policy generosity, drives demand — Rising CFE tariffs create the demand signal. — The DAC tariff in the central region rose from 6.65 pesos per kilowatt-hour in January to 6.75 pesos by June. Penetration remains roughly 14% of a potential 4.4 million users, leaving substantial runway.
Interconnection and compensation are the binding risk — CENACE reliability criteria, not the megawatt ceiling, constrain project economics. — December 2025 self-consumption rules allow surplus injection only subject to protection devices, connection studies and per-plant reliability criteria set by CENACE. Storage shifts from option to mandatory cost line if conditions are applied restrictively.
What we expect
Permit-light expansion holds (Likely) — The 0.7 MW exemption and the 0.7 to 20 MW self-consumption corridor operate as written; the Comisión Nacional de Energía registers and permits at workable cadence; Comisión Federal de Electricidad tariffs keep rising. Growth rotates from residential retrofit toward commercial and industrial systems. Asolmex projects distributed capacity between 8.3 and 11.3 gigawatts by 2030 under the new guidelines. Penetration remains roughly 14% of a potential 4.4 million users, so the runway is the story.
Compensation squeeze, self-consumption only (Possible) — The Comisión Nacional de Energía and the Centro Nacional de Control de Energía apply reliability and own-backup conditions restrictively. Own backup means storage capacity or contracted coverage with the state public utility for ramping, intermittency and variability. Surplus injection turns uneconomic, systems are sized strictly to daytime load, and storage becomes a mandatory cost line. Growth decelerates toward the pattern Reforma flagged. Engineering, procurement and construction firms carrying speculative pipelines take the loss; storage-integrated integrators gain share.
Active throttle (Unlikely) — Deliberate tightening: the medium-voltage compensation methodology is revised toward a self-consumption-only model, the design industry fought in the 2022 proposal that would have repealed net metering, plus anti-fragmentation rules against splitting projects to stay under 0.7 MW. The legacy-permit tripwire already bites: prior isolated-supply permits continue on the same terms only absent change in installed capacity or associated load demand, so any brownfield expansion reopens the permit file.
What to watch
- Installed additions — against the 529.53 MW first-half benchmark
- Application count — near 112,000 prior-year pace or falling
- CNE compensation rule — anteproyecto entering regulatory pipeline
- State leader share — Jalisco, Nuevo León, Chihuahua versus laggards
- Cogeneration registrations — issued under April 2026 CNE rules
Framing
Two Sheinbaum-era decisions redrew Mexico's solar map. The distributed generation exemption rose from 0.5 MW to 0.7 MW, and industrial on-site self-consumption was set at up to 20 MW for own use. Utility-scale stays rationed by a 46% private ceiling and a 9,550 MW renewables cap to 2030. Growth moves behind the meter.
Key judgments
- The corridor — A permit-light 0.7 to 20 MW band now exists that did not before. The distributed generation ceiling rose from 0.5 MW to 0.7 MW with no generation permit required, leaving only Comisión Federal de Electricidad interconnection to arrange, while industrial on-site self-consumption was set at up to 20 MW.
- Demand driver — Tariffs, not policy generosity, create the demand. The high-consumption residential (DAC) tariff in the central region rose from 6.65 pesos per kilowatt-hour in January to 6.75 pesos by June. Installed distributed capacity reached 5,437 MW at end-2025, a 22.2% annual increase from 4,447 MW at end-2024.
- The cap helps — The 46% private ceiling in generation, the 9,550 MW renewables cap to 2030 and the 54% Comisión Federal de Electricidad dispatch floor push private capital off the wholesale market and onto the customer's roof. Mid-market solar is structurally advantaged, not merely tolerated.
- The real risk — Compensation and interconnection, not the megawatt ceiling, are the binding uncertainty. December 2025 self-consumption rules let interconnected plants inject surplus without compensation, or sell it only subject to protection devices, connection studies and reliability criteria determined per plant by the Centro Nacional de Control de Energía.
- Positioning — Concentrate in high-tariff industrial states and the 0.5 to 5 MW commercial band, and own equity rather than sell equipment. Power generation is a United States-Mexico-Canada Agreement covered sector, so an equity-owning generator holds recourse a vendor or services model does not.
What could change our mind
- First-half 2026 distributed generation installed additions reported by the Comisión Nacional de Energía versus the 529.53 MW January-June benchmark reported from Comisión Reguladora de Energía statistics — Additions above 529.53 MW confirm Scenario 1 and validate the permit-light thesis. Additions below it, with application counts still holding, point to queue backlog and Scenario 2. Additions and applications falling together point to demand fear and Scenario 3. Non-publication of the series is itself evidence against Scenario 1.
- Application count in the same semiannual series, read against the 112,000 applications filed in the prior year and the 630,207 cumulative total, and against the 2024 comparator of 107,000 versus 65,100 — Applications sustaining near the prior-year run rate while installed megawatts lag isolates administration, not demand, as the constraint and moves toward Scenario 2. Applications falling materially below the prior-year pace moves toward Scenario 3. Both rising together is the cleanest confirmation of Scenario 1.
- A Comisión Nacional de Energía preliminary rule (*anteproyecto*) on distributed generation compensation entering the federal regulatory-improvement pipeline, or publication of a revised medium-voltage compensation methodology in the Diario Oficial de la Federación — Filing of such a preliminary rule is the cleanest early warning of Scenario 3, observable before publication. Absence of any filing through the window leaves the compensation regime for sub-0.7 MW net billing untouched and holds Scenario 1 as the base case.
- State-level distribution in the next Comisión Nacional de Energía table: whether the three leaders, Jalisco at 747.7 megawatts installed with 99,949 cumulative applications, Nuevo León at 543 megawatts and Chihuahua at 392 megawatts, take a smaller share of new additions than in the prior period — A declining leader share means growth is broadening beyond the cheapest geographies and confirms Scenario 1. A rising leader share, with laggards such as Tlaxcala still near 11.4 megawatts, means the market is saturating its best states and caps the mid-market runway toward Scenario 2. Treat state counts as approximate.
- Cogeneration uptake by mid-sized industrial users with thermal load, evidenced by registrations issued under the April 2026 Comisión Nacional de Energía cogeneration rules confirming distributed generation may be carried out through cogeneration systems — Registrations issued at visible cadence show the regulator administering the new figures permissively and reinforce Scenario 1, opening an underpriced adjacent segment. No registrations issued through the window indicates administrative friction and supports Scenario 2, where nominal openings are neutralized in processing rather than in law.
Who matters
- Claudia Sheinbaum — Sets the energy architecture that caps private wholesale participation while widening permit-exempt distributed and self-consumption thresholds — Committed to a 54% Comisión Federal de Electricidad dispatch floor and a 9,550 MW renewables cap to 2030, which limit how far she can liberalize large-scale generation
- Comisión Nacional de Energía — Successor regulator to the extinct Comisión Reguladora de Energía; issues self-consumption registrations, permits and the interconnection statistics series — Unverified processing capacity and no confirmation the prior statistical publication cadence has been continued
- Centro Nacional de Control de Energía — Determines per-plant reliability criteria that condition surplus sales and connection studies under the December 2025 self-consumption rules — Surplus sales may not affect system operation, giving reliability judgement primacy over project economics
- Comisión Federal de Electricidad — Sets the tariffs that create the arbitrage and controls the interconnection the exemption still requires — Tariff path is politically exposed to subsidy decisions; grid saturation limits medium-voltage connections
- Mid-sized installers and industrial offtakers — The capital deploying into the 0.5 to 5 MW commercial band — Dollar-denominated module and inverter imports against peso-quoted contracts; balance-sheet exposure if pipelines stall
What changed
Used to be: The distributed generation exemption sat at 0.5 MW with no permit-light corridor above it for industrial self-consumption.
Now: The exemption is 0.7 MW with no generation permit required, and industrial on-site self-consumption is set at up to 20 MW for own use.
Two Sheinbaum-era decisions created a permit-light band from 0.7 MW to 20 MW that did not exist before. Below 0.7 MW, only CFE interconnection is required. Above that, industrial users can self-consume up to 20 MW on site. Utility-scale access remains rationed by a 46% private ceiling and a 9,550 MW renewables cap to 2030.
The tariff side amplifies the structural shift. The high-consumption residential DAC tariff in the central region rose from 6.65 pesos per kilowatt-hour in January to 6.75 pesos by June, it is tariff arbitrage, not policy generosity, that fills pipelines. Distributed capacity reached 5,437 MW at end-2025, a 22.2% annual increase from 4,447 MW at end-2024, against roughly 14% penetration of a potential 4.4 million users.
The binding uncertainty has moved. The megawatt ceiling is no longer the constraint, surplus compensation and interconnection administration are. December 2025 self-consumption rules allow surplus injection only subject to protection devices, connection studies and reliability criteria determined per plant by CENACE, converting storage from an option into a potential mandatory cost line.
What would prove us wrong
- CFE tariff increases stall through an expanded subsidy decision — Tariff drift is the primary demand driver, not regulatory generosity, so a subsidy-driven freeze lengthens paybacks and compresses the mid-market opportunity regardless of how permissively the 0.7 MW and 20 MW thresholds are administered.
- CENACE applies reliability criteria to make surplus injection uneconomic — If per-plant connection studies and own-backup requirements are applied restrictively, systems get sized strictly to daytime load, storage becomes mandatory, and growth decelerates toward the compensation-squeeze scenario.
- CNE first-half 2026 additions fall below 529.53 MW and applications also drop — Both metrics falling together signals demand fear rather than administrative backlog, pointing toward deliberate throttling and breaking the permit-light thesis.
What it means for you
- Own equity, not equipment — Power generation is a USMCA-covered sector, so an equity-owning generator holds recourse that a vendor or services model does not, structure matters as much as project selection.
- Size storage into your base case now — CENACE's per-plant reliability criteria can make surplus injection uneconomic at any time, so treating storage as a mandatory cost line from the outset protects margin rather than surprises it.
- Watch the CNE first-half 2026 numbers before committing speculative pipeline — Additions above 529.53 MW confirm the permit-light thesis; additions below it with application counts still holding isolates administration, not demand, as the constraint and changes how you price pipeline risk.
Methodology
Confidence is moderate. Regulatory architecture is verified from primary instruments and is unambiguous on capacity thresholds and permit exemptions; administration is not. There is no retrieved data on Comisión Nacional de Energía interconnection processing times, no first-half 2026 statistics, and no confirmation the Comisión Reguladora de Energía statistical series continues under its successor. Two hypotheses are retained deliberately: genuine mid-market liberalization, or a nominal opening neutralized by compensation and reliability conditions. Current market metrics rest on a single unregistered outlet and should be re-verified against regulator publication before a capital decision. Measurement bias runs toward overstatement: applications are a leading proxy, not a substitute, and cumulative counts include withdrawn and stalled files. Announced megawatts in this segment have systematically overstated near-term delivery, so treat the upper end of the 8.3 to 11.3 gigawatt range as aspiration.
Sources