AI Data Centre Jobs: The Career Test Comes After Construction

In research published on November 9, Reserve Bank of Australia staff project A$115 billion in data centre investment between mid-2026 and the end of 2028, in real terms and excluding supporting infrastructure. They link the expansion to cloud computing as well as AI. [1]

I see credible employment opportunities in that expansion. I also think the career test comes when a facility opens: whether the people who built it can find continuing work. Project continuity and skills useful across sectors deserve as much attention as resistance to automation.

Before construction, a proposal can disappear

On November 9, AAP reported that Stockland had shelved its proposed data centre in Kogarah, near Sydney Airport, on November 7. The report does not establish that workers lost jobs. [2]

That decision alongside the investment forecast is the writing on the wall I take seriously. A growing industry can contain projects that never proceed. For someone planning a career move, the relevant question is how far a particular employer's work has progressed.

I would look for an actual vacancy tied to a defined project stage, and consider whether the employer has other work if that project stalls. An announced development can justify watching a company. It cannot settle when someone will be needed or how long they will stay.

The building phase offers a real opening

SEEK's July 2026 report recorded a 109.5% annual rise in advertisements mentioning data centres, while total advertisements fell 6.0%. Its examples include engineers, project managers, data technicians and heating, ventilation and air conditioning (HVAC) technicians. [3]

These are SEEK's platform observations. They describe July advertisements; the report does not establish completed hires, current November availability or employment duration. [3]

I would take that pocket of demand seriously while checking whether it persists in today's local vacancies. It cannot support a blanket invitation for every displaced office worker to become an infrastructure specialist. A vacancy matters only if someone can realistically qualify for it and reach its location.

Headline spending also needs translation. The RBA staff assume about 70% of investment goes on imports, limiting its direct contribution to Australian economic activity. [1] That percentage cannot be turned into a share of local jobs.

Before treating billions as a promise of employment, I want to see the roles and their expected duration. A large equipment bill tells an applicant very little about either.

Opening day changes the work

RBA business contacts expect ongoing operations to be less labour intensive than the investment phase. The staff describe their operational assessment as preliminary. [1]

I read this as a reason to examine building a facility and staffing it afterward separately. A thriving data centre could coexist with a much smaller requirement for workers once construction finishes. That is my inference; the research does not provide measured future employment outcomes.

I would treat a future operations role as a separate employment prospect requiring evidence of its own. An employer should explain whether an offer covers a defined project, movement between sites or ongoing operations. Those arrangements imply different career paths even when every advertisement contains the same sector label.

Completion need not interrupt employment. Successive builds could keep a contractor's workforce busy, while maintenance could provide another route. I would judge the prospect by the actual sequence of work available.

Skills need somewhere else to go

The RBA analysis describes transferable skills among electricians, plumbers, HVAC technicians and site supervisors. It also identifies potential investment in supporting electricity infrastructure and equipment replacement, with uncertain replacement timing and substantial imports. [1]

I consider that flexibility the strongest career argument in the evidence. Someone who can move between data centre projects and other facilities has more possible destinations when one site finishes. Transferability still needs available employers; it does not manufacture the next vacancy.

For a worker already in a relevant trade, I would investigate whether a data centre role adds experience useful on other sites. For a career changer, I would compare the qualifications required by actual vacancies across several sectors before choosing training. A course built around an exciting industry name deserves scrutiny if its skills have few other buyers.

Location matters, too. A career sustained by successive projects may require travel or moves between sites. That is a possibility to establish with the employer, especially for someone whose family or other commitments limit mobility. National investment totals cannot resolve a local employment problem.

Make the promise extend beyond the opening

I think employers benefiting from this expansion owe applicants a clearer account of continuity. They should explain what happens to project staff at completion, what conditions govern transfers and which maintenance or operating roles actually exist. A possible future assignment should be described as a possibility.

That would let people choose project work knowingly. A finite assignment can be worthwhile when its duration fits their plans and it develops useful experience. The harm comes from encouraging someone to treat a temporary surge in demand as settled career security.

My prognosis is cautiously positive for workers with relevant skills and credible routes between projects. It is much less reassuring for anyone relying on one proposed facility or an assumed move into operations. These Australian signals cannot establish how the global employment balance will turn out.

I would pursue the opportunity with the next stage already in view. The facility may be designed to operate for years; a career needs its own plan for what follows construction.