Somewhere under the Snowy Mountains right now, a tunnel boring machine named Florence is grinding through rock that has, at various points over the past couple of years, refused to behave the way the geotechnical reports said it would. That single fact — a machine the size of a small apartment block stuck in fractured, water-bearing ground — has done more to reshape the public conversation about Snowy 2.0 than any press release out of Canberra.
I’ve been tracking this project on and off since the 2018 final investment decision, back when Snowy Hydro was still talking about a 2021 completion and a price tag around $5.1 billion. Neither of those numbers survived contact with the mountain. Let’s be careful with what we say next, because this is a project where precise figures get quoted with false confidence all the time. What we can say, based on Snowy Hydro’s own public disclosures and testimony to Senate estimates, is that total project cost is now running past $12 billion, and the commissioning timeline has been pushed out repeatedly since the original schedule.
Where the tunnelling actually stands #
Snowy 2.0 isn’t one tunnel, it’s a network of them: a headrace tunnel, tailrace tunnel, access tunnels and an underground power station cavern excavated deep in Kosciuszko National Park, linking the existing Tantangara and Talbingo reservoirs. Three tunnel boring machines have been doing the heavy lifting — Florence, Lady Eileen Hudson and Kirsten — each named after women with a connection to the Snowy Scheme’s history, which is a nice touch even if it hasn’t made the ground any softer.
Florence is the one that’s caused the real headaches. It hit a fault zone with squeezing, unstable rock that isn’t behaving like the pre-construction geotechnical modelling assumed, and progress on that stretch slowed to a crawl for extended periods. Snowy Hydro’s construction updates have described remediation work, ground support changes and revised excavation methods through the affected sections. The company’s main contractor is Future Generation Joint Venture, led by the Italian firm Webuild, which took a larger role in the venture after the original Australian partner, Clough, went into voluntary administration in late 2022.
None of this is unusual for hard-rock tunnelling in fault-affected terrain — it happens on metro rail projects too — but the scale of Snowy 2.0’s tunnels, and the fact the whole project’s value proposition depends on getting water between two reservoirs 27 kilometres apart, means every month lost underground shows up directly in the project’s cost line.
The cost line, in real terms #
It’s worth walking through how we got from $2 billion to past $12 billion, because the jump looks abrupt if you only see the headline number.
- 2017 feasibility estimate, under the original Turnbull government announcement: in the order of $2 billion.
- 2018 final investment decision, Snowy Hydro’s own figure: around $5.1 billion for the scheme itself.
- By the early 2020s, with transmission links and remediation folded in, public estimates were sitting closer to $6 billion.
- Through 2022 and 2023, cost blowouts tied to tunnelling delays, contractor changes and site conditions pushed disclosed figures past $12 billion, a number Snowy Hydro’s leadership has confirmed publicly and which has been the subject of scrutiny at Senate estimates.
Snowy Hydro is a Commonwealth-owned corporation, so this isn’t shareholder money in the ASX sense — it’s public money, running through a government-owned generator, and that’s part of why the cost trajectory gets political attention that a private developer’s blowout might not. I’d argue the original 2017 costing was never especially credible; a project of this geological complexity needed longer in front-end investigation before anyone signed off on a number, and the industry has been paying for that compressed timeline ever since in the form of headline surprises.
Power versus energy — the number that actually matters #
Here’s where I want to slow down, because this is the distinction that gets flattened in most coverage. Snowy 2.0 is rated at 2,000 MW of generating and pumping capacity. That’s the power figure — how fast it can push electricity into or draw it from the grid at any instant.
The energy figure is different, and it’s the one that actually explains why this project matters. Snowy 2.0 is designed for around 175 hours of storage at full output, which works out to roughly 350,000 MWh of stored energy. Compare that with a big grid-scale battery: even a large one, in the 300–500 MW class, typically stores two to four hours of energy at rated power, so somewhere in the range of 1,000–2,000 MWh. Snowy 2.0’s storage duration is measured in days, not hours. That’s not a knock on batteries — I’ve written before that pumped hydro and batteries firm the grid in genuinely different ways, and the two are complementary rather than competing technologies for most purposes. But if you’re trying to ride through a multi-day wind lull across the southern states, duration is the whole game, and that’s the job Snowy 2.0 was designed for.
The timeline, again #
Original completion: 2021. Then 2025. Then guidance moved to 2027, then 2028 for full commissioning of all units, with first power flowing earlier as individual units come online in stages. That’s the pattern with this project — the date keeps moving in roughly two-year increments, and I’d treat any specific completion year quoted publicly right now as the current best estimate rather than a fixed commitment. The Australian National Audit Office examined the probity of the original investment decision process and was critical of aspects of the due diligence undertaken before the 2018 commitment, which is relevant background for anyone wondering how a project got this far off its original numbers.
The knock-on effects matter too. HumeLink, the transmission project meant to carry Snowy 2.0’s output into the grid, has its own schedule pressures, and NSW’s broader transition — coal exits meeting a Renewable Energy Zone rollout that’s already running late — depends on both pieces landing somewhere close to together. A pumped hydro plant with nowhere to send its power is an expensive cavern with good acoustics.
Why it still matters for the grid, even delayed #
AEMO’s Integrated System Plan has long treated large-scale, long-duration storage as one of the pillars the NEM needs as coal retires, alongside transmission and firmed renewables. Snowy 2.0 was meant to be the flagship of that pillar in the southern states, much as Kidston is doing the same job at smaller scale in north Queensland. The project’s delays haven’t changed AEMO’s underlying view that long-duration storage is needed; they’ve changed the assumptions about when it arrives, which is a live risk given coal capacity is leaving the grid faster than firm replacement is being built in some states.
My honest read: the tunnelling problems are a genuine engineering setback, not a sign the concept is wrong. Long-duration pumped hydro remains one of the few proven ways to store days’ worth of energy rather than hours. But the cost and schedule slippage has been large enough, and repeated enough, that I think it’s fair to ask whether the project’s economics still stack up the way they did in 2017, especially against how fast battery costs and deployment have moved in the meantime — a question I don’t think Snowy Hydro or the Commonwealth have answered publicly with anything more than reaffirmed commitment.
A small aside on patience #
I spend a fair bit of time on long rides through undulating country, and there’s a particular frustration in climbing a hill only to find the gradient’s steeper than the map suggested. Tunnelling through a fault zone is the engineering version of that, except you can’t just change gear — you stop, you re-survey, you change your ground support method, and you eat the delay. Nobody who’s actually dug a tunnel through the Snowy Mountains’ geology seems surprised by what’s happened with Florence. The surprise, if there is one, is that the original schedule assumed it wouldn’t.
What to watch next #
The things I’d keep an eye on: Snowy Hydro’s next construction update on tunnelling progress through the affected zones, any further cost disclosure at the next Senate estimates round, and whether HumeLink’s own schedule starts to converge with or diverge further from Snowy 2.0’s. AEMO’s Integrated System Plan updates are the place to check how much weight the market operator is still putting on this project’s timing in its reliability planning, and Snowy Hydro’s own project pages remain the primary source for construction status, rather than any secondhand account, mine included.
Whether Florence clears its current stretch in months or drags on longer, the project’s basic proposition hasn’t changed: 2,000 MW of power, roughly 350,000 MWh of storage, built for a grid that increasingly runs on wind and solar and needs something to cover the days when neither shows up. The mountain, so far, has had its own opinion about the schedule for delivering that.
— Anjali Rao, Grid & Storage Correspondent
Photo by Gavin Allanwood on Unsplash