Live · Oct 2026
Winner · Hong Kong AI Summit hackathon 2026
Drop a data centre on Hong Kong's power grid and watch a whole day play out. The map shows what overloads and when, what that costs, and the cheapest fix. Built in 36 hours for CLP's track at the Hong Kong AI Summit hackathon, and it won.
Open source · View on GitLabGridReady · the story
Scroll
Flat AI-training load · Hong Kong summer evening peak · day two with the fix
A data centre draws almost the same power at 3am as at 7pm, so the line feeding it is over its rating all day. Four new transformers bring the worst hour from 258% down to about 90%.
Live · drag a data centre onto the map, then Review and Fix
Start with the one-click example, a 300 MW AI campus at Sandy Ridge. Zoom into a district to see local 11kV substations and spare capacity.
Clips from the demo video
Recorded from the local build the night before the pitch
PLACE
a data centre at a substation, with its MW and load shape (flat, AI training, cloud)
FLOW
DC power flow over the CLP network for 48 half-hour steps, in a Web Worker
BREAK
every line and transformer over its rating, and when
COST
transformer wear for CLP, outage risk for customers, in HK$ per day
FIX
upgrades, transfers, phasing or a different site, ranked by cost and lead time
Rules first, so the demo cannot invent a number. Claude only rewrites the brief, and every figure is checked.
Indicative, from public data and estimated ratings
GridReady · the write-up
The Hong Kong AI Summit hackathon, theme BUILD ACROSS, run by the Erasmus AI Society and Maastricht AI Network and hosted by HKUST Business School. About 600 people registered. We took the Urbanization track, set by CLP, Hong Kong’s main power company: how might AI help CLP become a trusted energy partner for future households and businesses?
Judging came in two rounds. First a 90-second video submission, then a 3-minute pitch battle in front of the judges. GridReady won, and with it the HK$7,500 prize.
Data centres are Hong Kong’s fastest-growing load, and grid connection is the bottleneck. Sandy Ridge alone is a HK$23.8bn campus. A developer can buy land and wait months to learn whether CLP can supply it. GridReady answers in seconds, for both sides.
The 400kV backbone and substation sites are real, from OpenStreetMap: 90 nodes, 151 lines, 7,477 MW at peak. The 132kV layer and the ratings are estimates, and the app says so on every screen. A DC power flow runs for every half hour of a summer weekday in a Web Worker, fast enough to re-solve the day while you drag a slider. The pilot we pitched is CLP running it on its own data.
Verdicts, problems and fixes are rule-based. Claude only rewrites the brief into plain English, and if any figure in its answer isn’t in the computed facts, the rule-based text is shown instead. Costs come from IEEE C57.91 transformer ageing and the value of lost load, which turns a red line into a figure a business acts on.
I ran it as an orchestrator of AI agents: one wrote the engine against a single contract file, others built the data and UI on separate branches. GitLab CI deploys main behind this site’s nginx, at the address on the pitch’s QR code.
End of GridReady