Where should the next charger go?

Buy the tiles you want to develop in. Inside each one the engine has already scored every candidate location and picked the best; you get the answer and the working, not a pile of layers to weight yourself.

How the engine decides

Six sources, weighted for this market, with one hard disqualifier.

Product note — no weight slidersLetting users tune weights feels empowering and produces garbage: it turns a calibrated model into a projection surface for whatever the user already believed. Weights are a market parameter the engine owns and publishes. The user's leverage is which tiles to buy, which is a decision they actually have information about.

Bengaluru · 49 tiles

z13 tiles ≈ 4.8 km · click a tile to inspect or buy

Geography drawn from real coordinates · synthetic analysis overlay

drag to pan · scroll to zoom · click a tile

Tile inspector

Click any tile on the map.

Blind spots

Unowned tiles whose free signal suggests they beat what you own.

Product note — sell the gap, not the fogThe free signal is real open data with real noise on it, so it is honestly wrong sometimes. Publishing it costs a little revenue and buys the thing a data business actually runs on: the buyer believing the number when it says no.

Recommendation

Why here

Every signal the engine used on this location, in the weight it carried.

Product note — evidence, not controlsThese are read-only on purpose. The job to be done is "defend this site in an investment committee", which needs a citable chain from source to score, not a sandbox. The only interactive number in the whole product is the ROI model, because that is where the user's own private information legitimately enters.

Hard constraints

Checked before scoring. A failure removes the site rather than discounting it.

Runners-up in your tiles

What the engine ranked below this, and why

Buy tiles

Licensing is per tile, per year. A tile is a real z13 map tile — the same addressing your GIS team already uses — and it carries every source the engine consumes, refreshed on its own cadence.

Bundles

Contiguous coverage is worth more than scattered coverage — the engine can compare across tile boundaries.

Product note — why tiles and not a seat licenceA seat licence prices access. A tile prices exposure: it scales with how much capital the investor is actually deploying, it renews when they expand into a new corridor, and it makes churn visible one tile at a time instead of all at once at renewal.

What a tile contains

Identical bundle in every market. What differs is the weight each source carries.

All tiles

Sorted by free-signal strength. Owned tiles show the engine's actual best score.

ROI model — —

A ten-year cash-flow model seeded from the site's own signals. This is the one place your private assumptions belong, so every input here is yours to move.

Assumptions

Seeded from site signals and the price engine.

Blended price from engine—

Annual free cash flow

Post-tax, after lease and opex

Cumulative discounted cash flow

Crossing zero is payback

What moves NPV

One-at-a-time sensitivity

Scenario comparison

Same site, three demand worlds

Product note — break-even before NPVNPV is a number people argue about. The price and utilisation at which this site breaks even is a number they can go and test against a real offtake conversation next week. Leading with the falsifiable one is what makes the model get used instead of admired.

Price engine

Sets an hourly price that maximises contribution margin under your operating rules — wholesale tariff, elasticity, regulatory ceiling and utilisation targets. Flat pricing is the benchmark it has to beat.

Market inputs

Site: —

Operating rules

Constraints the optimiser must respect.

Product note — rules, not autopilotA pricing engine that just maximises margin is unshippable: the first evening it charges triple, the operator is on the news. The rules are where an operator's reputation, regulator and offtake contracts live, so they are first-class objects with visible cost. Each toggle shows what it costs per day to hold that position.

24-hour operating picture

Three panels, one shared time axis — price, throughput, margin.

Optimised vs flat

Same demand curve, same day

Hourly decision log

Why the engine moved the price

Portfolio

Every site you have run the ROI model on, blended into one number.

Sites modelled

Click a row to reopen its model

IRR by site

Against your cost of capital

Glossary

Every abbreviation, metric and input in the product. The same definitions appear on the ⓘ next to each field — this page is just all of them at once.

Case study

The reasoning behind the prototype — framing, the calls I made, and what I deliberately did not build.