Will Singapore's protracted data centre winter kill the ecosystem?

Is Singapore in the grip of a data centre winter? And is it killing the data centre ecosystem here?

We read about the AI and data centre boom in the news a lot these days.

But did you know that Singapore had a 4-year moratorium on new data centres that only ended last year?

Global surge of data centres

The rest of the world is having a data centre boom, with AI workloads threatening to unbalance national grids with pressure growing from Ireland to Germany and China.

Love it or hate it, I wrote previously about how the moratorium in Singapore was a last resort (Read: https://lnkd.in/gvMeyxgf).

The next generation of data centres requires much more power than before. Just 4 or 5 such 'hyperscale' facilities a year, and the total data centre footprint in Singapore could easily double in just 2 years.

The flip side of that would be today's minuscule 1.65% data centre vacancy, among the lowest in all tracked cities globally (Read: https://lnkd.in/guCrU7WS).

There are just a handful of new data centres, even counting the four DC-CFA winners (Read: https://lnkd.in/g7xXd_dR).

Upcoming data centres

Below are the notable upcoming data centres in Singapore.

𝘜𝘱𝘤𝘰𝘮𝘪𝘯𝘨/𝘚𝘵𝘪𝘭𝘭 𝘣𝘶𝘪𝘭𝘥𝘪𝘯𝘨 𝘶𝘱

✅ DC Tuas (67MW).
✅ Keppel at Genting Lane (Unknown).
✅ Keppel Floating Data Centre Park (Unknown).
✅ Meta's hyperscale data centre (150MW).

𝘕𝘰𝘵 𝘺𝘦𝘵 𝘴𝘵𝘢𝘳𝘵𝘦𝘥 (𝘋𝘊-𝘊𝘍𝘈)

⛔ AirTrunk-ByteDance (20MW)
⛔ Equinix (20MW)
⛔ GDS (20MW)
⛔ Microsoft (20MW)

Feeling the squeeze

I'm not here to talk about the data centre operators today, but the many vendors and specialists that serve the data centre industry.

They work in the background to build the critical infrastructure and systems powering modern data centres. But with limited facilities coming online, they can't rely on local opportunities for growth.

In fact, almost everyone I meet from the data centre space talks about past or planned trips to regional countries for various data centre projects.

Sure, the DC-CFA2 should be coming in the second half of this year. Yet how much hope can it offer when last year's DC-CFA winners have yet to start work?

My question: Would the ecosystem still be around in another couple of years?

👉 One contrarian view I heard is that the DC-CFA winners, with more stringent design requirements to meet, are not in a rush to build. True, not true? What are your thoughts about this?