Tag: Tasmania

  • Would data centres use “15% of Tasmania’s grid”? What the claim does — and doesn’t — tell us

    A widely repeated figure holds that the two Firmus facilities (St Leonards and Bell Bay) would together draw close to 15% of the power Tasmania’s grid can supply. The number is a reasonable shorthand, but it compresses several distinct concepts. This explainer separates them. DII takes no position on any individual project.

    Why this matters

    The claim that data centres would use 15% of Tasmania’s grid is neither wrong nor quite right; it depends on which measure of the grid is used and what new generation gets built. Those conditions are the real things to scrutinise. This explainer separates the concepts so the debate can focus on them.

    What the facilities would draw. St Leonards is approved for ~90MW in Stage 1 (44MW in Stage 1a), with a further ~300MW second stage planned subject to approvals. The proposed Bell Bay facility (GEE01) is ~288MW. Full build-out across both sites is therefore several hundred megawatts of near-continuous load — data centres run at high load factors, unlike most industrial users.

    What Tasmania’s system can supply. Tasmania’s generation is dominated by hydro — more than 2,600MW across 30 stations, with long-term average inflows equivalent to ~9,000GWh/year — plus ~568MW of wind and the 500MW Basslink interconnector. The energy constraint is real, not theoretical: very low rainfall in 2024-25 produced the lowest on-island hydro generation in 20 years. Two different denominators matter and are often conflated: installed capacity in MW — what can be generated at peak; and annual energy in GWh — what can be generated over a year, which for hydro is limited by rainfall, not turbines.

    Why the distinction matters. A 380MW+ combined load is modest against installed peak capacity but significant against annual hydro energy, particularly in dry years. Tasmania’s hydro storages have recently been managed for firming rather than baseload, which tightens the energy budget further. Whether new load is “manageable” therefore depends on what new generation is built, Marinus Link timing and the commercial terms of supply — questions currently under negotiation between Firmus, Hydro Tasmania and TasNetworks and not yet public.

    The factual bottom line. The 15% figure is neither scaremongering nor spin: it is roughly right as a share of energy at full build-out under today’s system and too high if new generation and interconnection arrive as planned. The honest answer is conditional and the conditions — new generation commitments, tariff terms, curtailment provisions — are the things worth scrutinising in the approval process.

    The Digital Infrastructure Institute is an independent research and education organisation. It takes no position for or against individual projects and accepts no proponent funding.

    Sources: DCD — Firmus 288MW Bell Bay; DCD — St Leonards approval; Hydro Tasmania — power stations; RenewEconomy — hydro storages shift to firming; Tasmanian Economic Regulator — Annual Energy Security Review 2024-25; IEA Hydropower — Hydro Tasmania fleet

  • How data centre jobs are counted (and why the numbers vary so much)

    Two very different job numbers are circulating about Tasmania’s proposed data centres and both can be true at the same time. This explainer sets out what each measures, so residents, councillors and journalists can compare like with like. DII takes no position on any individual project.

    Why this matters

    Job numbers are the most quoted and least comparable figures in Tasmania’s data centre debate. Different claims measure different things: construction versus ongoing roles, headcounts versus modelled estimates. Comparing like with like changes the quality of the whole conversation.

    Construction jobs vs ongoing jobs. Large data centres are construction-heavy and operations-light. A build phase can employ hundreds of trades and engineering workers for one to three years. Once operating, a facility runs with a small permanent team of technicians, electricians, security and facilities staff. Public debate often quotes one figure without saying which phase it refers to.

    The “FTE per megawatt” benchmark. Ongoing data centre employment is commonly estimated relative to IT load. Firmus has cited roughly 0.5 full-time roles per megawatt — about 52 ongoing roles at St Leonards and about 144 at Bell Bay if built to ~288MW. Published international benchmarks for large automated campuses sit at roughly 0.15–0.35 FTE per megawatt, with the most automated hyperscale sites running 20–30 permanent staff per 100MW (Hamm Institute, 2025) and industry figures of 25–40 operators per 100MW reported elsewhere. Firmus’s cited ratio is at or slightly above the top of that range. Smaller enterprise or edge facilities typically employ more people per megawatt; very large automated campuses employ fewer.

    Indirect and induced jobs. Proponent figures sometimes include supply-chain and spending effects (catering, maintenance contracts, local services). These are real but estimated by economic modelling, not headcounts and depend heavily on assumptions. Any figure described as “supporting X jobs” should be read as a modelled estimate.

    What a fair comparison looks like. When assessing a proposal, the useful questions are: how many ongoing FTE roles on site, over what ramp-up period; how many construction FTE-years; what share of each can realistically be filled locally given current skills; and what training commitments, if any, are attached. These are factual questions councils can put to any proponent.

    The trade-off being debated. Critics note that per unit of electricity consumed, data centres employ fewer ongoing workers than most industrial users. Proponents respond that value should also be measured in investment, rates revenue and digital capability. Both framings are legitimate — they are answers to different questions.

    The Digital Infrastructure Institute is an independent research and education organisation. It takes no position for or against individual projects and accepts no proponent funding.

    Sources: Pulse Tasmania — Firmus defends jobs and energy use; Pulse Tasmania — Launceston approval; Hamm Institute — Data Center Employment Forecast Analysis (Nov 2025); Latitude Media — data centre staffing benchmarks