For about six months my team has had the same circular argument. We have an inference workload that would be materially better run on-shore, close to the data it reads, and every time we priced it the answer came back the same way: there is no sensible capacity available in Singapore at a price that makes the business case work, so we keep it offshore and accept the latency. On Friday that argument became answerable, and the useful part was not the headline number.
What Opened, and the Figures Worth Keeping
Equinix opened SG6, its sixth International Business Exchange facility in Singapore, with an initial investment of roughly US$290 million. eWeek reported the launch on 6 November 2026, and W.Media covered it the same day. Phase one delivers 8 MW, reported at around 1,550 cabinets, with up to 20 MW of IT capacity targeted at full build-out.
The engineering details matter more than the money. The site is built for high-density deployments with liquid cooling capability and advanced cooling for power-intensive GPU infrastructure, and it is designed to reach a 1.20 PUE, which is a design target rather than a measured result. It is targeting Green Mark Platinum, the top tier of Singapore’s scheme, with rooftop and facade solar, NEWater, rainwater harvesting and lower-embodied-carbon construction materials. Equinix also intends to open a Co-Innovation Facility at the site with the National University of Singapore, studying liquid cooling, cognitive digital twins and battery energy storage.
Cyrus Adaggra, Equinix President of Asia-Pacific, framed the regional case around “AI adoption, cloud expansion and data-intensive workloads”, while Adaire Fox-Martin, Equinix Chief Executive Officer and President, noted that “Equinix has been investing in Singapore for nearly 25 years, and our new facility will further expand our presence”. Worth noting what was not disclosed: pricing, how much capacity is already committed to customers, and any AI performance benchmark. Those three absences are why this is a planning signal rather than a procurement decision.
The Thing Non-Singapore Readers Miss: Capacity Here Is Granted
In most markets you treat data centre capacity as something you buy. Here it is something that gets allocated. In August a further 200 MW was provisionally allocated across four operators, Equinix, Digital Realty, Keppel Data Centres and ST Telemedia Global Data Centres, at 50 MW each under the second call for application, and SG6 itself came from an earlier allocation. Equinix’s separate 50 MW award is for a seventh facility on Jurong Island. There is also an efficiency condition attached at the equipment level, with customer IT equipment expected to meet a recognised energy-efficiency certification or demonstrate equivalent efficiency under SS 715:2025.
Read that as an engineering constraint rather than a policy summary and the hiring implication falls out immediately. If supply is granted in fixed increments and carries an efficiency condition, then the valuable engineer is the one who gets more useful work out of an allocated megawatt. The engineer whose instinct is to solve a saturation problem by provisioning more is not merely expensive in this market, they are proposing something that is frequently unavailable at any price.
Our expert take
The number I would want from any infrastructure candidate in this market is not a platform or a certification. It is a utilisation figure they have personally measured and defended: what fraction of the capacity you were allocated was doing work someone valued. Almost nobody can produce it, which is exactly why asking makes the question useful. The ones who can produce it have usually worked somewhere that capacity was genuinely scarce, and that experience is the thing being priced here.
The 4 Hiring Signals I Read That Day
- Liquid cooling stopped being exotic. A flagship local facility opening with liquid cooling capability and a dedicated research programme on it means the skill moves from specialist curiosity to something your next infrastructure hire should have at least touched. I am not suggesting you require it. I am suggesting you ask, because in twelve months the answer will be a differentiator and today it is still cheap to find someone who has it.
- On-shore became arguable again, which creates roles rather than filling them. Our deferred workload is now worth re-pricing. Note the asymmetry though: capacity existing does not mean capacity is affordable, and pricing was not disclosed. What this unblocks is the analysis, and the person who can do that analysis credibly, comparing total cost of a local high-density deployment against the offshore status quo, is a hire I would make before committing to any migration.
- Efficiency compliance became a job requirement, not a procurement footnote. With an efficiency expectation attached to tenant equipment, someone on your side has to own it. In a small team that is a responsibility added to an existing role rather than a new seat, and it should be written into the job description explicitly, because engineers who have lived with this kind of condition work differently from those who have not.
- The research facility is a talent pipeline signal, on a long lag. A co-innovation programme with a local university on cooling, digital twins and energy storage will produce graduates and researchers with exactly the skills this market will be short of in two to three years. That is worth a relationship now and nothing in your headcount plan this quarter.
Re-pricing an on-shore workload
We help Singapore employers scope infrastructure roles against real power, cooling and cost constraints, then shortlist pre-vetted engineers against that scope.
Let us discuss itWhat Moves When, Because It Is Not All at Once
The single most common error I see after an infrastructure announcement is treating it as one event with one response. A facility opening reaches a hiring plan through several channels on very different clocks, and only the first one is actionable this month.
| What moves | Lag | What to do now |
|---|---|---|
| Your own job descriptions and screening questions | Immediate, you control it | Rewrite around density, cooling and utilisation this week |
| Re-pricing of deferred on-shore workloads | 2 to 6 weeks, once pricing is quotable | Commission the analysis, do not commit to the migration |
| Demand for high-density and cooling-literate engineers | 1 to 2 quarters, as tenants move in | Hire ahead of it if the skill is already on your roadmap |
| Competition for the same engineers from operators | 1 to 2 quarters | Check your senior infrastructure bands before, not after |
| New graduate supply from the research programme | 2 to 3 years | Build the university relationship, change nothing else |
Our expert take
The row I would watch hardest is the fourth. When a major operator brings a facility online and another is in build on Jurong Island, those sites need operations, power and cooling engineers from the same small local pool that your platform team recruits from. The bands move quietly, and the first sign is usually not a competing offer. It is a counter-offer you did not expect, on a renewal you assumed was routine. Checking your senior infrastructure bands this month costs you an afternoon.
The Mistake I Made the Last Time Capacity Opened Up
A previous cycle, a different operator, similar announcement. I took it as permission to stop optimising. We had been running a genuinely wasteful training pipeline, tolerated because we assumed we would simply move it somewhere with more room. Capacity arrived, we moved it, and the costs went up by more than the move saved, because the inefficiency travelled with it and the new environment metered things the old one had bundled. The engineer who eventually fixed it did so by profiling the pipeline and cutting the work in half, which is what we should have done before relocating anything.
That is why my response on Friday was to re-open an analysis rather than approve a migration. New capacity does not make a wasteful workload economic. It makes it slightly more affordable to keep being wasteful, which is a different and much worse thing to buy. If you only take one line from this piece, take that one.
Our expert take
My expectation for the first half of 2027 in Singapore: the job title that grows fastest is not a new one. It is the existing platform or infrastructure engineer role with efficiency accountability bolted on, and with a utilisation or cost-per-unit-of-work number in the objectives. Employers who add that accountability now, at current bands, will look prescient. Those who wait until an operator is bidding for the same people will pay for the same capability under a more fashionable title.
What to Watch Over the Next 30 Days
- Whether indicative pricing becomes quotable. Nothing was disclosed on launch. Until a number exists, any on-shore migration business case is a sketch, and treating it as more than that is how teams commit to moves they reverse.
- How much of the 8 MW first phase is already committed. Committed capacity tells you whether this genuinely relieves scarcity or simply houses demand that was already queued.
- Operations and cooling job adverts from the operators. These are the clearest early read on whether local senior infrastructure bands are about to move, and they appear before any competing offer reaches your team.
- Your own renewal conversations. Local data beats market narrative. If your infrastructure engineers renew on expected terms through November, nothing has reached your team yet.
If you want the longer version of the staffing argument underneath this, our method for building a data centre engineering team in Singapore covers the operations side, and hiring AI infrastructure engineers in Singapore covers the tenant side, which is where most readers of this piece actually sit.
Frequently Asked Questions
What did Equinix actually open in Singapore on 6 November 2026?
Does more data centre capacity in Singapore mean more engineering jobs?
Why does the 20 MW full build figure matter to a hiring decision?
What should we change in an infrastructure engineer job description this month?
Discussing it beats guessing at it
We help Singapore employers write infrastructure roles around the constraints they actually operate under, and shortlist pre-vetted engineers who have worked inside them.
Let us discuss it