Shell Hong Kong · EV charging
Built entirely from public information — your own locator and Recharge pages, the EPD charger register, 中電 CLP and 港燈 HK Electric published tariffs, and a listed competitor’s audited accounts.
The finding
Not a saving I am claiming. A threshold you can test against one bill.
The whole model
Demand cost per unit
= rate ÷ (730 h × load factor)
Load factor measured against billed maximum kVA, not installed equipment. Put in your own rate and your own load factor and it answers itself.
中電 CLP Bulk, against a HK$1.36 spread
| How busy the site is | Demand charge costs | Left over |
|---|---|---|
| Nearly empty — 2% | HK$5.13 | −HK$3.77 |
| 5.5% — my estimate | HK$1.87 | −HK$0.50 |
| The turning point — 7.5% | HK$1.36 | HK$0.00 |
| Busy — 15% | HK$0.68 | +HK$0.68 |
| Very busy — 40% | HK$0.26 | +HK$1.11 |
The one audited window — Cornerstone Technologies, HKEX GEM:8391, FY2025
Audited, year ended 31 Dec 2025
| HK$M | FY2025 | FY2024 |
|---|---|---|
| EV charging income | 44.2 | 23.9 |
| Selling hardware | 64.6 | 64.8 |
| Installing it | 7.4 | 60.3 |
| Total revenue | 125.2 | 153.1 |
| Gross margin | 24.8% | 17.8% |
Charging income +85.3% to HK$44.2M — now 35.3% of revenue, up from 15.6%.
Revenue fell only because they chose to stop installing (−87.8%). Margin rose 7 points.
And still: adjusted LBITDA −HK$34.4M, net loss HK$115.2M, accumulated losses HK$617.8M.
Going-concern material uncertainty every period since 2020.
Growing the network fast, and improving the margin, did not make it self-funding. Scale was not the answer to unit economics. Adjusted LBITDA is a management-defined measure for their whole EV business — hardware, installation, charging and overhead — not the public network alone. Their property, tariff and funding structure are not Shell’s, and no per-point figure can be derived: the GO/HOME revenue split is not disclosed.
Who owns the levers
| Idea | Who has to agree |
|---|---|
| Loyalty points, redemptions | The platform partner builds it |
| Session-level measurement | Partner likely holds the data |
| Idle fees | Already live, already theirs |
| The electricity tariff | Not the app. Plausibly Shell alone. |
Every other idea needs somebody else to say yes first.
Three thresholds. They answer different questions.
6–12%
Is a site already on Bulk underwater?
Where the demand charge eats the whole margin. Centred on 7.5%, but it moves with price and site share — a band, not a constant.
29%
Which tariff should this site be on?
Where Bulk’s cheaper energy finally outweighs its demand charge. Different question, different number.
10.4%
Should you build another bay?
Carries HK$317,520 of capital over eight years. Why I would not lead with building more chargers.
I am not claiming any of these is your number. Two caveats I would rather say than have you find: the demand charge is billed on recorded monthly peak, not installed power — a quiet site is billed on the 100 kVA minimum instead. And the EV tariff applies to a supply used solely for EV charging, so on a shared forecourt supply, switching may need separate metering. Not automatically free.
What I am asking
Tariff code, twelve months of charges, interval demand. Whoever already owns that data can answer it faster than I could from outside.
If those sites are on the EV tariff with no demand charge,
this idea is dead and I will tell you so tonight.