Mexico Quietly Turns Self-Generation Into a Serious Solution

Mexico just published the new “Autoconsumo” DACGs — and quietly turned self-generation into a serious solution. If you were planning to oversize and “sell the energy leftovers,” the document has a polite but firm answer: no.

Last week, I heard a detail that made this real very fast: the first cogeneration permit under this new Autoconsumo (auto-consumption) framework is already operating — the first one in Mexico actually running under these rules. No press release. No ribbon-cutting. Just a permit, a meter, and electricity doing its job. That’s the key point: this is not a theoretical regulatory exercise.

The winners won’t be the people with the most panels. They’ll be the ones — customers and suppliers — who treat auto-consumption like an operating system, not a one-time installation.

The new DACGs were published in the DOF as “Disposiciones Administrativas de Carácter General para regular la figura de Autoconsumo de Energía Eléctrica (General Administrative Provisions to regulate the figure of Auto-Consumption of Electrical Energy).” It’s not light reading, unless contractual definitions and acronyms are your idea of fun.

Here’s a simple way to approach documents like this without losing a full day: don’t start with definitions. Start with the sections that can kill your project. In this case, those are termination triggers, excess-sale mechanisms, and private-wire rules. Everything else is commentary. And yes, your favorite AI model can help.

Idea 1: Your Permit Can Die From Underuse

The most important sentence in the DACGs is not about solar, batteries, or interconnection. It’s about when the regulator can decide your permit has lost its purpose.

The DACGs state that the permit’s “object” is considered gone if, after the plant is operating, the ratio of aggregated peak demand (from autoconsumo users) to installed capacity stays below a threshold for a continuous 12-month period: below 50% for conventional sources, below 30% for renewables — unless a technical justification is submitted and accepted.

Let’s do the math. If you build 10MW of renewable generation, you need roughly ≥ 3MW of aggregated peak demand to stay above the 30% threshold. If your demand is 2MW, the ratio is 20%. You’re in the danger zone for a year. The sun doesn’t care. The regulator does.

I used to think the biggest risk in industrial self-generation was interconnection timing. It still is, but this ratio is just as critical — a business risk disguised as a technical requirement. Factory expansion plans change. Shifts get cut. Efficiency programs land. Suddenly, your design becomes a permit problem.

This pushes a new sizing logic. In 2026, the conservative move is not “maximum roof utilization.” It’s “minimum structural excess.” That usually means sizing closer to the load curve, designing controllability, or pairing with storage justified by operations.

There’s also a second-order effect few people mention: energy efficiency can increase regulatory risk if demand falls faster than capacity. This isn’t an argument against efficiency. It’s an argument for planning your load trajectory as carefully as your generation.

The DACGs do leave a door open: a technical justification can be submitted shortly before the 12-month window closes, and the regulator may grant up to an additional 12 months to restore compliance.

Idea 2: Selling Excess Is No Longer a Side Business

The DACGs allow interconnection with excess injection, but they narrow the commercial path significantly.

If you sell excess, you can sell it only to CFE, through contracts defined under this framework. Any idea of a mini-merchant strategy is effectively removed, unless subasta pricing somehow works in your model.

If the plant is intermittent and injects energy, backup is mandatory, either through storage or by paying for it via the state utility. This is a system stability requirement, not a philosophical one, and it shifts cost from “optional optimization” to “required capability.”

Operationally, the process is strict. The model contract requires notifying the buyer at least 72 hours before submitting offers to the short-term market, after which the accepted amount must be offered by your market representative.

The price formula is designed to remove upside.

For clean energy plants, compensation is the lower of:

• 0.9 × the levelized cost of energy from the last long-term auction, and

• 0.9 × the hourly local marginal price (PML).

If PML is negative, it is treated as zero.

This is intentional. You don’t capture price spikes, and you don’t get compensated when prices go negative. Excess sales become “nice to have” revenue, not project-defining revenue.

Building a business case around excess sales under this scheme is like buying a food truck because a mall allows you to sell there — but only on approved days and at posted prices. You can still make money. It just can’t be the core strategy.

CFE must publish the relevant levelized costs by the end of January each year, meaning expected compensation is partially anchored to a public reference.

Idea 3: The Private Wire Is Now a Regulated Asset

Autoconsumo is defined around a private network (Red Particular) and on-site needs, with a ≥ 0.7MW threshold. This matters because the DACGs regulate not just generation, but site electrical architecture.

Private networks must be independent, tied only to the permitted installations, and cannot interconnect with other private networks. This effectively ends certain “campus federation” designs where private wires are bridged over time.

Each private network is also limited to a single active interconnection point with the transmission or distribution system. If your campus growth plan relies on multiple interconnection points, the architecture must be rethought early.

Two Types of Interconnection

The DACGs define two operational modes.

Without selling excess, CENACE studies assume zero net capacity at the point of interconnection, and reverse-power or low-consumption protection is required. In simple terms: connect, but don’t push power back.

With excess sales, protections still apply, but registration and market representation are required for both shortfalls and injections. This added operational burden is why many industrial projects should default to “interconnected, no sale.”

On compliance: the autoconsumo registry must be updated for modifications within 15 business days, or by March 31 at the latest. Resolution can take up to 90 days.

A “modification” includes changes in demand, installations, and explicitly, the addition of EV charging infrastructure. Decarbonization roadmaps now come with regulatory housekeeping.

Quick reality check: Nobody budgets for long-term compliance operations. We budget once for panels, inverters, and legal work.

Why This Matters for 2026

The DACGs make auto-consumption a real, enforceable path for industrial self-supply — but with discipline.

Industries will stop treating self-generation as a procurement project and start treating it like a production line: sized to demand, controllable, auditable, and resilient to business change.

The quiet message is clear: Mexico is making room for self-supply, not for disguised merchant generation. The economics reward consumption, not speculation.

A fair objection is that the DACGs add friction and cap upside, potentially slowing investment when industry needs reliability. Some will read the termination triggers and exclusive sale rules and conclude that autoconsumo is being constrained, not enabled.

That view is understandable.

The counterpoint is Autoconsumo aislado: exclusive on-site supply, which under the DACGs does not require studies, registry, or market representation. It’s not for every site, but for some industries it’s the cleanest path to reliability when interconnection queues are heavy.

If there’s a single-sentence strategy for 2026–2027, it’s this: build autoconsumption to cover core load, control the rest, and treat excess sales as optional — not foundational.