Hello Reader,
On Friday, the results of the DC-CFA2 were announced. As the four winning operators celebrate, at least 16 others were left disappointed - and no doubt wondering if winning was even a realistic possibility.
Today, I want to talk about Singapore's DC-CFA2: what was so surprising about the results, why was it so hard, its inherent "unfairness" - and a suggestion for a future CFA.
Surprising winners
On Friday, 130 days after the deadline for submissions, IMDA and EDB announced the four CFA2 winners: Digital Realty, Equinix, Keppel Data Centres, and ST Telemedia Global Data Centres (STT GDC).
The CFA is part of Singapore's controlled-growth regime for approving new data centres in the established data centre hub. This second run, or "CFA2", challenged operators to design cutting-edge data centres that push the envelope on efficiency, leverage low-carbon fuel, and deliver economic benefits to Singapore.
In my informal chats with operators and industry players, some expressed surprise at the results. How did Equinix, which also won the first CFA, clinch a consecutive win? What did Equinix bring to the table that was so compelling that Singapore awarded one of four coveted slots to them - at the risk of a perception of partiality?
Also, why did Keppel Data Centres, which had already built at least two large data centres at Genting Lane during the data centre moratorium, and is known to have started earlier this year on a floating data centre, even participate in the CFA2?
A bar only giants can clear
To be clear, I am not hinting at any impropriety. On the contrary, what I've seen is scrupulous efforts to ensure a level playing field with well-attended briefings, multiple industry consultations, and detailed responses to questions raised by interested parties. Moreover, an inter-agency team gave expert feedback, while key personnel even recused themselves from industry events.
But the requirements were always biased towards operators with the deepest pockets, large global networks, a sizeable workforce, and extensive technical capabilities. This shouldn't be surprising if you consider the objective to push the envelope on sustainability, energy efficiency, benefits to the data centre ecosystem, and economic benefits. I've long said that small players have virtually no chance today.
According to IMDA and EDB, over 20 operators submitted bids this time. Though their identities were not disclosed, we could guess at many of them from the industry grapevine, as well as their efforts to publicise their data centre-related initiatives.
I think it is fair to say that many of them are very large players able to bring substantial financial and technical resources to bear. I also personally know of at least one global data centre operator that eventually opted not to make a submission due to the high costs of meeting the many criteria.
And perhaps the fraught and quickly evolving geopolitical environment we find ourselves in also played a not-insignificant part in the choice of winners. The schism between east and west, the increasingly vital role of digital infrastructure, and the profusion of laws with extraterritorial reach, are all weighing in favour of a certain profile of winners.
Room for the next STT GDC
Yet if resource depth is the crucial criterion, why ever run a CFA at all? Why not just have IMDA and EDB negotiate directly with the top global operators? This didn't happen because we intuitively know the logical endpoint here is a permanent closed loop, the kind that edges out innovation but rewards compliance-grade capabilities - not, I suspect, where breakthroughs come from.
This is my worry with criteria that screen heavily on incumbent capabilities and scale. They reward mastery of the current format at the very moment that format itself is changing. Think of the latest AI data centres with their extreme rack densities, liquid cooling to the chip, and unconventional power systems - arguably a different kind of facility. And data centres themselves have evolved more in the last few years than in the decade before.
To be fair, EDB does have its own track for innovative AI players that might need data centre capacity. STT GDC earlier this year launched a testbed for direct current (DC) power in data centres. But my point stays: Some of that knowledge or innovative mindsets could well sit in unexpected places. This means a regime with zero pathway for smaller players doesn't just exclude weak bids; it could exclude the source of the next design paradigm.
My suggestion is a modest one: when the next CFA is designed, ask whether the criteria leave any realistic path for a smaller, more innovative player. Right now, I don't think they do.
What's next
For now, it will be a time of building for the winners. Likely drawing on lessons from the pilot CFA, the stipulated timeline to get the CFA2 data centres up and running is tight, so a lot of work will have to be done quickly after the celebratory drinks this weekend.
What can we expect from these facilities? Roy Samuels, the APAC director for Technical and Reliability at CBRE Data Centre Solutions, echoed my words in his comment to my LinkedIn post this morning: "Looking forward to the promised 'cutting edge' technologies and renewables."
For those who didn't make it, perhaps the learnings carry over to their next bid. And hopefully the regulators will tune the criteria slightly.
As usual, you can reach me by hitting "reply" to this email.
Regards,
Paul Mah
