Singapore stopped issuing new data center capacity in 2019 and the pipeline moved across the strait into Johor, so the two countries only make sense read together: the demand is Singaporean, the land and the power are Malaysian. Neither TNB nor SP Group publishes substation headroom or a connection queue, so each cluster gets a 0 to 100 Connectability Score from what open data does show — how near and how many high-voltage substations, how much capacity is already built nearby, and the distance to an exchange and a cable landing.
| # | Cluster | DC sites | Connectability | Substations ≤5 km | Nearest substation | To exchange | To landing |
|---|
Connectability Score. Each cluster gets a score from 0 to 100 from four public indicators: distance to the nearest high-voltage substation, how many sit within reach, how much data center capacity is already built nearby, and distance to an exchange or cable landing. The (i) beside the ranking gives the weights and scales. The same score, weights and colour bands are used on every power page, so clusters can be compared across markets.
No headroom is published. Neither TNB in Malaysia nor SP Group in Singapore publishes spare capacity per substation, and neither runs a public connection queue. So the score rests on proximity, not available capacity; a high score is a reason to ask the utility, not a connection.
Network density. The count of mapped substations within 5 km of a cluster. Neither TNB nor SP Group publishes headroom or a connection queue, and the World Bank has no Malaysian grid layers, so this is the same proxy the UAE page uses. A dense network means an existing supply that can be reinforced; a thin one means a new line. It is a proxy for connectability, not for available megawatts, and it inherits OpenStreetMap's coverage: Singapore is mapped far more completely than East Malaysia.
The corridor. Johor and Singapore are treated as one market because the constraint and the response sit on opposite sides of the causeway. A Johor cluster 30 km from Singapore is nearer to Singaporean fibre than to most Malaysian load centres, so exchange distance is measured across the border rather than within each country.
Connectivity. PeeringDB facilities and the exchanges they host, and TeleGeography cable landings. Singapore is one of the densest cable landing points in Asia and holds most of the region's peering, so a Johor site inherits much of that connectivity at a fraction of the land cost — which is the whole argument for the corridor.
Generation and climate. Power stations over 5 MW and transmission lines come from OpenStreetMap, so they are indicative rather than complete. ERA5 gives cooling degree days and free-cooling hours at each cluster; both countries are equatorial and humid, so free cooling is close to zero and the cooling load is the year-round base case, not a summer peak. Ember gives each country's generation mix and carbon intensity separately, because they differ: Malaysia still burns a lot of coal, Singapore is almost entirely gas.
What is missing. Substation capacity and the connection queue (TNB and SP Group requests), Singapore EMA and Malaysia Energy Commission tariffs and statistics, Johor state land and plot data, and water stress: WRI Aqueduct ships as a download that has to be prepared by hand, so neither country is in it yet. These are tracked as open requests.