Per-load metering
Metering hardware at the electrical panel, with one metering point per relevant load and continuous readings.
Smart energy rails
We meter energy load by load in retail stores and show you how to cut the cost
A store gets one electricity bill for everything it uses. Glow installs metering at the electrical panel and separates how much the HVAC, the refrigeration, the lighting and the EV chargers each use.
That data shows what drove the month's cost, how much it cost and which action brings it down. We start with the EV chargers, the load nobody has separate data on today.
What Glow does
Glow is an energy metering and management company for retail chains. The service has three layers on top of the same installation.
Analytics on site data
Models trained on how each store actually runs, checked against its bill. That is where peak attribution, anomaly detection and the dollar value of each action come from.
Execution on the equipment
Integration with the HVAC and refrigeration controllers already in place, and with the charging system. It starts with customer approval and moves toward autonomous operation.
Every chain starts with its EV chargers, the one load on site that sells energy as well as consuming it, and adds the other loads when it decides to.
Why a store cannot see where its energy goes
A chain store can spend hundreds of thousands of dollars a year on electricity and get a single bill for all of it. Three things keep that bill from explaining the spend.
One meter for the whole store
Every load draws behind the same meter, and the bill arrives as a total, not a breakdown. Without per-load metering there is no way to tell which one drove the cost.
You pay for consumption and for the peak
The medium-voltage commercial tariff charges two things at once: the energy consumed and the peak demand. Peak demand is set by the 15-minute interval with the highest draw, and that single interval fixes the charge for the whole month.
Penalties and prices that change every month
Power factor can add up to 3.5% of the bill. CFE also publishes 17 separate regional price lists and updates them monthly, so the same consumption costs a different amount depending on where and when it happened.
Source: CFE commercial time-of-use tariff for medium voltage (GDMTH): energy, capacity and distribution charges, power factor formula, and monthly price lists by division.
How it works
The service runs in four steps. Each one depends on the one before it.
Meter
Each relevant load gets its own metering point. Metering runs continuously from day one, because consumption that went unmeasured in a given month cannot be reconstructed later.
Learn
Each store's history trains models of how that site runs. Against that reference, the system flags what falls outside the expected range and attributes the month's peak to the load that created it.
Decide
The system returns a specific action and what it is worth, priced with the tariff in effect for that region and that month. For example, if a fleet charges between 7 and 9 PM and the peak forms there, the action is to move that charging to a lower-demand window.
Execute
At first, every action is applied with approval from someone on the customer's side. As verified history builds up, the system starts executing on its own the actions that have already proven their result. Autonomous execution is a stage of the same system, not a separate product.
The platform
These are the pieces the system is built from and how data moves between them.
On-site acquisition
Metering hardware goes into the electrical panel, on the existing circuits, without replacing the store's equipment. A small site is covered by one metering point and a large one runs to ten, depending on how many relevant circuits it has.
Continuous time-series ingestion
Each point produces a continuous consumption series, ingested and kept at its original interval. Analysis is aligned to the 15-minute intervals the tariff uses to set billable demand, so what is measured can be compared with what is billed.
Bill ingestion and reconciliation
The platform ingests electricity bills and breaks them into their components: energy by time-of-use period, capacity charge, distribution charge, power factor, and that month's regional price list. With both sources in the same structure, measured consumption is reconciled against billed consumption, period by period. That reconciliation is what makes it possible to explain the bill load by load.
Models trained per site
There is no single model for the whole chain. Each store has its own hours, its own seasonality and its own mix of equipment, so models are trained on that site's data and retrained as new readings and new bills arrive. Against that baseline, the system identifies consumption outside the usual pattern and attributes peak demand to the loads running during the interval that set it.
Integration with on-site controllers
The platform connects to the HVAC and refrigeration controllers already in place, reads their state and applies the change to them once an action is approved. The founding team built this kind of integration before Glow existed, on the controllers running in 70% of the stores of one of the largest retail chains in Mexico.
EV charging and OCPI interoperability
Chargers come in as one more load: their consumption is metered along with everything else, and their sessions are recorded as operating data. Interoperability runs on OCPI, the international protocol for exchange between charge point operators and e-mobility service providers. That lets the platform work with the hardware already installed and enable roaming, so drivers coming from partner apps can find and use the store's chargers.
Roland Ferwerda, CEO of GreenBee, Glow's product partner, chairs the international body behind OCPI.
Delivering the results
Results go out through a web dashboard, email and WhatsApp, with the same structure on every channel: what drove the cost, how much it cost, what action is recommended and what it is worth. The channel depends on who has to act.
Closed verification loop
Every action taken is checked against the next period's bill. If the savings showed up, they are recorded as a verified result; if not, the difference goes back into the model as a correction. What one store learns is available to the next one.
Who it's for
Glow works with operators running several stores on the medium-voltage commercial tariff, where the same problem repeats site by site: one bill per store, with no detail by load.
Supermarkets and self-service
Heavy refrigeration and long hours, where the peak forms when several loads run at once.
Department stores
Large floor plates where HVAC and lighting dominate, with their own parking, which is where EV charging is going in.
Convenience stores
Many small, similar sites, where what one store learns carries over to the rest of the network.
Multi-site operators with parking
Sites where EV charging already runs or is about to be installed, usually under a third party.
Glow is a good fit when
- The site is on the medium-voltage commercial tariff (GDMTH) and pays peak demand charges.
- The operation has more than one site and wants to compare stores on the same metering basis.
- Chargers are run by a third party, with no clarity on what they consume or what they cost.
- HVAC and refrigeration already have controllers and are run without per-load consumption data.
Glow is not built for
Homes, or a single low-voltage location with no demand charges. In those cases, per-load metering does not pay for itself out of the savings it creates.
Operations start in Mexico. The same tariff structure and the same problem exist across the rest of Latin America, and that is the scope of the project.
How we work with each chain
Implementation moves in stages. Each one leaves the infrastructure in place for the next and ends with something the customer can use.
Weeks 1 to 4
Diagnostic
We review the site's billing history, rebuild its consumption profile and run a metering test on a first load.
From the customer: the last few months of bills and a visit to the electrical panel.
Result: where the cost is concentrated and which loads to meter first.
Months 1 to 6
Pilot
Metering goes into one store and visibility into the EV charging system comes online, with the dashboard and the first recommendations.
From the customer: one store, a maintenance counterpart and an installation window.
Result: charging consumption, cost and revenue separated from the rest of the site.
Months 6 to 18
Expansion
The rest of the store's loads come online, including HVAC, refrigeration and lighting, and the system starts executing actions with prior approval.
From the customer: which loads to add and who authorizes each action.
Result: the store is fully instrumented and savings are measured against the pilot baseline.
Month 18 onward
Autonomy
The system executes autonomously within the limits the chain defines, and the model rolls out to the remaining stores.
From the customer: the rules and limits the system has to respect.
Result: energy operations managed continuously across the network.
Once metering is in, adding HVAC, refrigeration or lighting is a configuration decision on the customer's side, not a new implementation.
Where the product stands
The platform is in active development. The version being built delivers visibility into the EV charging system: charger consumption, the cost that goes with it, and session data. Next comes execution with human approval on HVAC and refrigeration, and after that autonomous operation within the limits each customer sets.
The per-load metering, bill processing and analytics layers are not new. They are the technology the founding team ran at scale in retail before Glow existed. What Glow is building on top of them is the integration with EV charging.
We describe the real state of the product so it is clear what it delivers today and what it delivers later.
What this team built before Glow
The figures in this section come from work the founding team did before Glow was formed. They are not results from Glow operations.
500+
sites monitored
84,000
utility bills processed
~6%
savings identified
That work was done inside two of the three largest retail chains in Mexico, and it included proven integration with the HVAC and refrigeration controllers running in 70% of the stores of one of them.
The customers on those projects are covered by active confidentiality agreements, so their names are not published. The technical scope of the work can be described in a direct conversation.
Founding team
Four founders, with prior experience in data science applied to energy, commercial strategy, infrastructure finance and e-mobility policy.
Neftaly Ramírez
CEO
13 years in strategy and go-to-market for high-growth companies across Latin America. Venture-backed founder.
Neftaly Ramírez on LinkedIn
Paolo De Ferari
COO
13 years in corporate finance for infrastructure and energy projects, plus capital raising across Latin America.
Paolo De Ferari on LinkedIn
Andrés Villaseñor
CTO
10 years in data science and 8 in energy, focused on metering and electricity consumption analytics in retail.
Andrés Villaseñor on LinkedIn
Juan Camilo Ramírez
CPO
Sustainable transport advisor to the UN, the UK government and the European Commission. Product development for e-mobility companies.
Juan Camilo Ramírez on LinkedInGreenBee, strategic partner
Dutch company with more than 20 years in energy, telecom and charging system design across Europe and the Americas. Its CEO, Roland Ferwerda, chairs the international body behind the OCPI protocol. GreenBee works with Glow as a product partner on the design and integration of the charging system.
Partners and ecosystem
The current relationships around the product, with the type of agreement behind each one. A letter of intent sets out how the two sides would work together. It is not a service contract or a customer relationship.
Kigo
Letter of intentMobility app with 5.7 million downloads.
Parko
Letter of intentOperator with more than 350 parking sites across Mexico.
Solarbit
Letter of intentInstaller and Siemens EPC partner.
Catalyst
PortfolioPortfolio and acceleration relationship, with Tec de Monterrey and the Green Climate Fund among its backers.
GreenBee
Product and strategic partnerCharging system design and OCPI interoperability.
The letters of intent signed with mobility partners open up reach to more than 3 million drivers once a store's chargers are published for roaming.
Frequently asked questions
What has to be installed in the store?
Metering hardware at the electrical panel, with one metering point for each load you want to separate. It mounts on the existing circuits, the work is coordinated with the maintenance team, and it does not require changing the store's equipment or its energy supplier.
Does it work with chargers that are already installed?
Yes, when the operator or the manufacturer supports OCPI, the international protocol for exchange between charge point operators and mobility apps. OCPI is what lets the platform read sessions, publish availability and enable roaming without changing the hardware. If a charger does not expose OCPI, its consumption is still metered at the panel, though without session data.
Do the HVAC or refrigeration controllers have to be replaced?
No. The platform connects to the controllers the store already has, through their integration interfaces.
How are savings measured?
Against the bill. During the diagnostic and the pilot, a baseline is set from measured consumption and billing history. Each action taken is then compared with the next period's bill, covering energy, peak demand and power factor. If the effect does not show up on the bill, it does not count as savings.
What happens to operating data?
Consumption, billing and operating data for each site belong to the customer. Glow uses that data to deliver the contracted service: training the models for those stores, reconciling against billing and generating recommendations. Glow does not publish the identity of its customers. How data from visitors to this site is handled is described in the privacy notice; each chain's operating data is governed by its own service agreement.
Which countries does Glow operate in?
Operations start in Mexico, where CFE's medium-voltage commercial tariff applies. The scope of the project is regional across Latin America, because the tariff structure and the visibility problem repeat in the other markets. Glow Mobility Technologies LLC has been incorporated in Delaware, United States, since 2024.
How is the service priced?
A fixed fee per metering point, plus a share of the revenue from EV charging, the one load on site that sells energy as well as consuming it. Once autonomous execution is available on the other loads, a share of the savings verified against baseline is added. Pricing is set to stay under 3% of the site's energy spend and under a quarter of the savings created. Terms are agreed chain by chain, after the diagnostic.
How long before there is a result?
The diagnostic takes one to four weeks and returns a measured reading of consumption, not an estimate. The first verified savings come with the bill for the period after the first action is applied.
Contact
To talk about the product, a test at one site or the technical detail of the platform, write directly to either address.
Registered address
16192 Coastal Highway, Lewes, Delaware 19958, United States
Operations in Mexico
Acordada 32-203, Col. San José Insurgentes, C.P. 03900, Benito Juárez, Mexico City
Email goes straight to the team. There is no form in between.
Glow Mobility Technologies LLC, incorporated in Delaware, United States, in 2024, operating in Mexico with regional scope across Latin America.