Transcription
In December 2022, a sinkhole opened in the Kosiosco National Park in southeastern Australia. That was already strange. I mean, while sinkholes exist in Australia, they are relatively uncommon in the granite alpine geology of this national park, making the collapse really rather unusual.
But not only did the sinkhole appear, it did so quite unfortunately above a 2400 ton subterranean tunnel boring machine named Florence. Florence had such a weird name for a tunnel boring machine. Florence had until then been drilling for around 8 months and her aim was to create a tunnel running approximately 15 km or 9.3 mi beneath the park. But she'd only managed to advance 150 m before the ground above her caved in. This measly yield was just a taste of the woe suffered by Snowy 2.0, an enormous pumped hydro battery project intended to revolutionize Australia's energy.
The project was announced back in 2017 with a price tag of around $2 billion Australian dollars with the aim that power could be flowing by 2021. But by 2022, the project had not seen success. Its cost had blown past $5 billion and was moving steadily upwards. Meanwhile, Florence had gotten stuck in saturated mud while crews above her pumped cement into the earth to stop it swallowing up the entire hillside. The project has faced brutal geological challenges, various cost overruns, surprise, surprise, and has left a troublesome environmental footprint that few can still defend. Snowy 2.0 was meant to stabilize Australia's energy future, but for most of the past 5 years, the only thing unstable has been the project itself.
So, let's take a look, shall we? Today's video is brought to you by Squarespace. And look, I've talked before about how most people overcomplicate starting an online project. Whether it's a portfolio, a business, a newsletter, a course, or just a place to put your work. The hardest part is usually just getting started. And that's why I personally like Squarespace. Squarespace is an all-in-one platform that lets you build a professional website without needing to know how to code or spend weeks figuring things out. One feature that I really like is their blueprint AI website builder. You answer a few questions about what you're building, and it helps you generate a site with structure, design direction, and content suggestions, so you're not just staring at a blank page and wondering, "Where the hell do I begin?" The second thing that I love is their design tools and templates. Everything looks polished right out of the box, and it is easy to customize with their drag and drop editing, so you can make it feel like your own instead of looking just like every other website on the internet. And if you're trying to grow a business, Squarespace also includes built-in SEO and analytics tools so you can understand where your traffic is coming from and help more people discover what you're creating. Whether you're launching a side project, showcasing your work, selling products, offering services, or building an audience, Squarespace gives you everything you need in one place. Go to squarespace.com/mega projects for a free trial. And when you're ready to launch, use the code mega projects to save 10% off your first purchase of a website or a domain. And now back to today's episode.
Why Snowy 2.0? Now, the idea behind Snowy 2.0 is older than most of the politicians who championed it, and it is really rather simple. Take two reservoirs sitting at different heights. When electricity is cheap and abundant, say on an afternoon in sun-kissed Australia, solar panels produce more power than the electricity grid needs. That excess energy can be used to pump water from the lower reservoir up to the higher one, so that later when demand rises and renewable generation drops, the water is released back downhill through turbines, generating electricity on the way. Simple as that.
In Snowy 2.0's case, the lower reservoir is Taingo and the upper is Tantagara, two legs that lie in the state of New South Wales on the edge of the snowy mountain range or the snowies. Between them sits roughly 700 m of vertical drop, which is of course the asset that makes this whole concept work. To link them, engineers planned about 27 km of tunnels carved through the Snowy Mountains, feeding an underground power station housing six reversible turbine units. Together, those turbines were rated at 2200 megawatt of capacity with up to 350 gatt hours of stored energy on tap. Snowy Hydro, the company behind the Snowy 2.0 project, liked to put that storage figure in domestic terms. Enough, the company said, to power 3 million homes for a week. Treat that figure as a claim rather than a precise promise, though.
Now, to see why this mattered in 2017, we have to look at what was happening to the Australian grid. Wind and solar were coming online at pace, becoming cheaper every year, but highly dependent on weather and sun cycles. The coal plants that had carried the country's base load for decades were aging, becoming both expensive to maintain, and increasingly uneconomic against renewables during daylight hours. Politicians warned about the risk of blackouts in South Australia and rising gas prices. Above all, they argued about what could be done to fix it. What the grid needed was dispatchable storage. Basically, something large enough to soak up renewable surpluses and big enough to cover the evening peak when the sun went down. Batteries could do it at small scale. Pumped hydro could do it at a national scale and for decades, not just hours.
There was also a bit of a story to tell here. And the prime minister at the time and his government, Michael Turble, understood the value of telling that story. The original Snowy Mountain scheme, built between 1949 and 1974, had become one of the country's founding industrial legends. 16 dams, seven power stations, more than 100,000 workers drawn from across postwar Europe. It had given Australia a self-image as a nation that built difficult things in hard places. And so, Turnbull green lit Snowy 2.0, sold as the perfect 21st century mega project. Built on the legacy of Australia's most celebrated engineering achievement, it promised to transform the old snowy scheme into a clean energy powerhouse for the renewable era. Supporters believed it could solve a looming energy storage crisis, secure the grid against blackouts, and prove that nationbuilding infrastructure still had a place in Australia's future.
But in truth, the idea wasn't entirely turbos to begin with, and its introduction had already proved anything but smooth. The idea of expanding the Snowy scheme with pumped storage had actually been on the cards for decades. Engineers had discussed the proposal back in 1966 and again in 1980 and once more in 1990. Each time, the numbers were just filed away. At the time, the country had cheap coal. The grid was stable enough and nobody could see who would pay for several billion dollars of tunnels through a national park. So what would become Snowy 2.0 sat collecting dust for the better part of 50 years.
What pulled it off the shelf was the energy debate of the mid-2010s. As mentioned before, coal plants were now closing. Renewables were arriving faster than anyone had foreseen and the politics of blackouts became rather poisonous. So 2 years after being elected prime minister of Australia in 2015, Malcolm Turnbull announced plans for Snowy 2.0 alongside a price tag of around $2 billion. Turnbull added that construction would commence in 2019 and the project would be completed in 2021 subject to feasibility. Led by Snowy Hydro, an Australian company that manages the original Snowy Mountain scheme, the project was an electorally favorable promise for what had become by 2017 a serious domestic energy concern. Big enough to sound serious, small enough to sound deliverable. Turnbull became the political champion, while Paul Broad, Snowy Hydro's chief executive, was the public face. Bullish, confident, willing to dismiss critics as people who simply did not understand Pumped Hydro.
But there were warning signs. An ABC article from May 2017 claimed the plans for the project had been approved by the prime minister only 2 weeks after being put forward by Snowy Hydro. Somewhat prophetically, the article warned that Snowy 2.0's true cost could skyrocket, although amusingly it pitched skyrocket at eclipsing $3 billion. Turnbull's optimistic $2 billion estimate didn't live out the year. By December 2017, a feasibility study put the cost of between 3.8 and 4.5 billion. By 2019, when major works contract was signed with future generation joint venture, the figure had reached 5.1 billion. Construction had barely started and the budget had already more than doubled.
In 2018, the Australian Commonwealth, the federal government, bought out stakes in Snowy Hydro owned by the government of New South Wales and nearby Victoria for about $6 billion, making the company fully federally owned. From that point on, every cost overrun, every missed deadline, every cubic meter of tunnel landed on the Commonwealth balance sheet. No state government to share the pain.
On the ground, things moved. Access roads were cut into the steep country around Lobs Hole, the main construction site deep inside Kazco National Park. Worker camps went up. Environmental approvals were chased through state and federal processes, mostly successfully. Given that bureaucratic obstacles could pillar large federal projects, all seemed to be going rather well so far. Then all was very much not well, and primarily for two reasons that few could have seen coming. The Black Summer bushfires of 2019 and 2020 tore through Snowy 2.0's construction zone, charring the countryside and forcing evacuations of locals. Then COVID arrived months later, disrupting supply chains, shutting borders to overseas specialists, and slowing everything down that depended on any sort of travel.
By the end of 2021, Snowy 2.0 was still laboring forward. Tunnels were being bored, the personnel camps were reoccupied, and their tunnel boring machine called Florence was being assembled at Tatangara, the upper reservoir. But the project's budget had tripled before the hardest challenges had even been touched.
The $2 billion mistake. In March 2022, school children were invited to Tatangara to sign their names on the cutterhead of the exciting 2,400 ton tunnel boring machine, Florence. The machine was actually named after Florence Violet Mckenzie. Widely regarded as Australia's first female electrical engineer, Snowy Hydro presented the machine as a dual mode marvel capable of grinding through hard rock with a rotating disc cutter or switching to a slurry mode that could handle soft waterlogged ground. She'd also have no problem cutting through the 15 km of tunnel from Tatangara down toward the underground power station. And Florence's dual mode capability was a crucial feature. The geography beneath Kiosco Park was known to be unusually tricky, featuring patches of fractured waterbearing ground as well as well, just solid rock.
To run in soft ground mode, Florence needed a slurry circuit working. Bentonite fluid was pumped to the cutter face to hold the ground stable, and a treatment plant on the surface was there to separate slurry from the spoil and send the fluid back down. But the treatment plant was not fully built on site by the time the machine began drilling. And so poor old Florence was sent forward in a configuration suited to harder ground than what the probe was already drilling. These were exactly the conditions that Florence was least equipped for. Saturated soft material and large inflows of water. And so 8 weeks later and only about 100 m in, Florence stalled. Mud simply flowed in around her, causing progress to collapse to crawling rates. This was something like half a meter of advancing followed by days of pumping out mud and cleaning the conveyor system. Not brilliant. And keep in mind, this was now the summer of 2022 in the northern hemisphere, which is winter in Australia. So, what were already poor conditions were understandably every bit worse. Even so, Snowy Hydro's public statements that Australian winter described steady progress on the Tatangara Drive. Yet, underground, the operation was effectively stalled. Workers were spending more time clearing the spoil out of the back of the alien machine than the cutter was spending grinding rock at the front.
In December 2022, the consequences reached the surface. Above Lawrence's position, the ground gave way and a sinkhole appeared in the alpine landscape of the national park. The cavity was the unmistakable signature of soft ground that had not been properly supported. Material flowing into the tunnel faster than it could be replaced or sealed. After 8 months of work, Florence had advanced 150 m out of her intended 15 kilometer drive. The repair operation that followed ran for most of 2023. Crews pumped grout, cement, and foam into the ground around and above the machine, trying to stabilize the saturated material enough to free Florence and seal the void. The volumes involved were enormous.
But in July 2023, again, Australian winter, one of those grouting operations went rather wrong. A reaction in the injected polyther foam produced isocyanate gas, a toxic compound that attacks the lungs and airways. Workers needed to be immediately evacuated as government hazard agencies issued notices over serious risk to the health of staff. A work site that had already lost a year to mud had now lost workers to fumes. Malcolm Turnbull, the prime minister who had launched the project back in 2017, watched the unfolding mess and offered a blunt assessment. Somebody, he said, had screwed up. And it was true that sending a dual mode machine into known soft ground without a slurry system ready was the kind of decision that became harder to defend with every meter of grout pumped into the hillside. Admittedly, this was somewhat easier for Turnbull to say. He'd stepped down as PM in 2018 following a leadership challenge from within his own party, partly stemming from Turnbull's chronically low opinion polling. By the time Snowy 2.0's woes were gaining pace, Turnbull had resigned from politics entirely and was now working in the private sector. Interestingly, he later sold two pumped hydro storage projects to Australian energy company AGL.
Eventually, Florence was freed though and ready to restart. But the cost of her stalled drive had caused the project's cost to grow to around 8 billion Australian dollars. One tunnel boring machine stuck behind one badly chosen launch configuration had become the most expensive hole in Australian infrastructure history.
The blame game. Any remaining optimism surrounding Snowy 2.0 did not outlive 2022. In August, Paul Broad resigned as Snowy Hydro's chief executive after years of insisting the project was on track. To be fair, his departure came as the picture underground at Tatangara was getting impossible to spin. Not long after Broad's departure, Clough Group, the Australian half of the contractor consortium driving the works collapsed into administration. Another company, WeBuild, absorbed Clough's Australian operations to keep crews on site and concrete moving. But the disruption rippled through every subcontractor on the project.
In February 2023, Dennis Barnes took over as Snowy Hydro's chief executive. His brief was straightforward. Open the books, work out how bad the situation actually was, and tell the Commonwealth shareholder the truth. This led to the so-called reset, wherein internal facts would be established, and as could be predicted, the final costs and deadlines would be reconfigured. And so it was. 6 months later, Snowy Hydro was ready to unveil the new estimated cost of the project, $12 billion. The first power deadline was pushed to late 2027, and full operation slipped to 2028, 7 years after its initial target date. At least Snowy Hydro's own framing was unusually candid. The company conceded that the design at the point of contract signing had been immature and that construction was proving more complex and much more expensive than anticipated. Even so, this was still a spectacular understatement.
Australia's new energy minister Chris Bowen was more direct. In a joint media release with Snowy Hydro at the end of August 2023, he stated that the site conditions should have been recognized as potential causes of cost escalation and that a second tunnel boring machine should have been ordered from the outset. The release made Snowy Hydro's six-month assessment look questionable. In 2022, the Australian National Audit Office had revealed the company's early governance arrangements and found them effective with no recommendations for change. But within months of that finding being published, Florence was bathed in mud and a sinkhole was opening above her. The audit had seemingly just inspected the paperwork around a project whose underground reality was already coming apart.
And then there was the cost situation. At $12 billion and climbing, Snowy 2.0 was on track to exceed the inflation-adjusted cost of the entire original Snowy Mountain scheme. 16 dams, seven power stations, and 145 km of tunnels built between 1949 and 1974. Snowy 2.0 wasn't even building any new dams. It was adding one power station and 27 km of tunnel to infrastructure that already existed. For that, the country was now committed to spending more than it had spent on the founding infrastructure. Inflation adjusted.
The accountability question sharpened further when Snowy Hydro's executive bonus arrangements came under scrutiny. Senior staff had continued to receive performance payments through years in which the project missed targets, blew its budget, and delivered a sinkhole. Critics in Australia's new administration found the defense harder to accept each time the cost estimate moved. By late 2023, the numbers were public and well-known. Ownership of the project had transferred from people who'd launched it, the Turnbull government, to the people now stuck delivering it under Prime Minister Anthony Albanese. Snowy Hydro's chief executive had swapped out and the project had passed fully into the hands of the national government rather than regional governments. A lot of changing hands and still the project was not even close to being complete.
More trouble underground. The 2023 reset at least brought Snowy Hydro something it had not had for a while. A few months of cautiously good news. Kirsten, a smaller boring machine working on an emergency egress and ventilation tunnel, completed her 2.9 km emergency segment in early 2024. The Kirsten Tunnel only served to provide supporting infrastructure, but at least it proved the geology was manageable this time and offered Snowy 2.0 its first construction win following 5 years of flailing around. Meanwhile, the Lady Eileen Hudson, another boring machine, and also who's inventing these names, was making useful progress on her own assigned tunnel. By mid-2024, excavation of the vast underground cavern that would house the power station was roughly halfway done. Crews were drilling and blasting through hard granite at the heart of the mountain, and the rock was behaving.
Then Florence got stuck again. After her year-long extraction from the mud at Tatangara, she had been restarted and pointed deeper into the mountain. The route required her to steer through a curved section of tunnel. And somewhere in that curve, the rock began to close in on her shield. Tunnelers call this rock pinching, when surrounding stone deforms under pressure and grips the cylindrical body of the machine, locking her in place. Florence's first immobilization had been soft ground flowing in. Her second, it was the opposite. It was hard rock squeezing down. And so in August 2024, Florence ground to a halt once more. Freeing her required specialist crews armed with ultra high pressure water jets cutting the granite away from her shield millimeter by millimeter until she could move forward again. And so once again, it fell months and months behind.
While that work was underway, a separate incident hit the main access tunnel. In April and May of 2024, during a drill and blast operation, a charge brought down rock into an adjacent tunnel one blast earlier than the plan had called for. Happily, no workers were underground when it happened. Internal accounts that later reached reporters alleged that larger blast rounds than standard were being fired in an effort to speed excavation. The breakthrough into the neighboring tunnel was by that account the predictable consequence of pushing the explosive charge beyond conservative limits.
All this resulted in a fourth tunnel boring machine purchased in 2024. The plan was to launch her from the Talbingo end and drive the Headrace tunnel, the one Florence had been supposed to be working on from the opposite direction. Basically, they got the second machine that Chris Bowen had said they should have got at the beginning. Reducing the project's dependence on Florence's single drive was now treated as little more than a basic requirement if Snowy 2.0 was ever going to be finished at all.
But incidents kept coming through 2025. In February, during underground testing, one of the project's massive ventilation fans imploded under pressure conditions that it was specifically designed to withstand. Metal shrapnel tore through the fan casing and flew across the construction site, narrowly avoiding nearby workers. The accident forced an immediate shutdown of underground operations and triggered an emergency inspection by Safe Work NSW, which in turn caused inspectors to issue six prohibition notices after discovering loose and missing bolts on multiple other critical ventilation systems. Investigators and union representatives called the implosion a near fatality event. And all this raised serious questions about maintenance standards and management oversight on the already deeply troubled project.
Amidst all these accrued problems, industrial relations worsened. Workers began walking off the job in a series of strikes and stoppages, protesting what unions described as poor pay agreements, exhausting shift schedules, understaffing, and worsening site conditions deep inside the tunnels. The latter, at least, would seem to be rather credible. Each such stoppage brought critical construction work to a standstill, delaying progress for days at a time and adding further pressure to a project already running many, many, many billions of dollars over budget. And so workplace safety issues which had already halted work during the COVID pandemic came to undermine the project once again. Safe Work NSW shut down a concrete testing laboratory at the site after finding silica dust controls inadequate, raising fresh questions about how the project was managing the health of its workforce.
By October 2025, the project was in dire straits. But Snowy Hydro clung to the hopeful news that the project was roughly 67% complete by scope. Tunnels were advancing, the cavern was taking shape, and the fourth machine was preparing to launch. Yet, in the same breath, the company acknowledged what almost everyone watching had already concluded. The $12 billion figure agreed in the 2023 reset was not going to hold. Another revision was definitely coming.
The environmental price. All right, let's leave the tunnels aside for just a moment and move above ground and into the present. As you know, Snowy 2.0 is being built inside Kosusco National Park, which happens to be a protected alpine landscape, home to some of the rarest habitats on the Australian mainland. So although the project's goal is to support renewable energy, a noble cause, its construction site is one of the most ecologically sensitive places in the entire country. Snowy Hydro's answer to that tension is a single number. The direct construction footprint, the company claimed in a 2023 status report, would cover roughly 0.01% of the park. By that measure, the disturbance is vanishingly small against the broader natural wilderness. But critics argue the footprint figure captures the bulldozed area above ground and not much else. Not the water systems being reshaped underground, not the spoil dumps, not the species movements the project will set in motion.
The spoil alone is significant. Excavating 27 km of tunnel and a power station cavern is generating roughly 9 million cubic meters of rock that has to go somewhere. Some is being used on site for road and pad construction. The rest is piled in designated imp placement areas within the park. If containment fails during heavy rain, sediment and dissolved minerals can wash into the Tutt and Morumba river systems downstream. The water itself raises a deeper concern. The Talbingo and Tantangara reservoirs sit in catchments that in ecological terms have been kept apart for a long time. But pumping water back and forth between them every day for the life of the scheme biologically connects the two. Fish, eggs, larvae, plant fragments, and pathogens are going to travel with the water. And you might be wondering, well, okay, but why does that matter? Well, the issue is that Talbingo, the lower reservoir, holds species that Tatangara's catchment does not. Redfin perch, an introduced predator, lives in the lower system. So does the small invasive fish gambusia, alternatively known as the mosquito fish, which is a notorious pest. And with a name like that, none of us are surprised. Both are precisely the kind of passenger that pumped storage is well suited to redistribute. Ecologists have warned that the critically endangered stock galaxius, a small native fish surviving in only a handful of high country streams, sits squarely in the path of any such new arrivals.
The legal accommodation around all of this is rather interesting. To proceed, Snowy 2.0 requires exemptions under New South Wales biosecurity rules covering the movement of species that would normally trigger control measures. The state government approved the project in May 2020, thereby effectively exempting Snowy Hydro from biosecurity restrictions that would otherwise prohibit that species transfer. Scientists argued this was unprecedented. And there was an obvious irony to all of this. A project sold as climate infrastructure needed special permission to move organisms that the same state's laws were designed to keep in place.
Meanwhile, Humlink, the high voltage transmission line being built to carry Snowy 2.0's power to the grid has generated its own separate fight over compensation, easements, and overhead pylons through farmland. Of course, none of this means Snowy 2.0 should not be built necessarily. It does mean the brochure version though, clean energy, minimal disturbance, and national park essentially untouched is uh quite the opposite of the spoil heaps, the connected catchments, and the biosecurity paperwork required to make the scheme, you know, legal.
By early 2026, the underground picture had finally begun to look less like a crisis and more like a construction site. In May, Lady Eileen Hudson, who can forget that name, this guy, broke through into the main cavern, joining the network of completed tunnels. The fourth boring machine, the one ordered late to attack the headrace from the Talbingo end was now part of the recovery plan rather than being part of the future. Crews shifted attention from raw excavation toward the long task of fitting out the power station itself. Installing turbines, generators, transformers, and the thousands of kilometers of cabling that turn a hole in a mountain into a generator, or a very big battery, I suppose. The official targets have not moved since the 2023 reset. First power is still scheduled for late 2027, and full commercial operation is still scheduled for 2028. A fresh cost review is underway at the time of writing. Although I wonder which direction that's going to go. Although Snowy Hydro has already conceded that the $12 billion figure agreed at the reset is not going to be the final number. So how far past it the project lands is now the central question for the Australian government.
And what you might ask does this mean for the central questions that Snowy 2.0 was meant to resolve? Well, unsurprisingly, it's not exactly brilliant news. Australia's remaining coal plants are scheduled to retire through the late 2020s and into the 2030s. If Snowy 2.0's 2200 megawatts arrive late, the buffer has to come from somewhere else. Possibly more grid-scale batteries, more gas peaking plants, more demand response programs, or by somewhat humiliatingly extending the life of coal stations the country has already announced it will close. Each option carries its own cost and its own political fight.
If the project does land, the upside is real. Pumped hydro at this scale is designed to operate for a century or more. Down moles, tunnels, and underground caverns have lifespans measured in generations, and the turbines can be replaced inside the same civil shell. A renewable heavy grid in 2050 would have genuine use for 350 gigawatt hours of dispatchable storage sitting under the Snowy Mountains. But Snowy 2.0's critics already have their verdict ready. The project is little more than a white elephant oversold by politicians, mismanaged by its operator, and propped up by only the impossibility of walking away. Interestingly, some critics had this verdict sealed and stamped from the very outset of the project, such as energy expert Bruce Mountain, great name, who wrote something to that effect as early as October 2019. Bruce Mountain, god damn.
Yet, the project supporters have their own stance. Snowy 2.0 is now a painful but necessary delivery of infrastructure the country will be glad to own once it is finished. The honest answer will not be available though until those turbines start spinning. Snowy 2.0 may yet power Australia through the years when coal retires and the wind does not blow. The tunnels will be there and the turbines will spin, in theory. Whatever the final cost, the country will eventually own a crucial working asset in the mountains. But the legacy of how it was built is already settled. A $2 billion announcement made during a brief administration has since devolved into a $12 billion open hole. A single tunnel boring machine best exemplified this, being delivered into ground its equipment was not ready to handle. Then there was the audit that somehow found governance effective only months before a sinkhole opened in a national park. And there were the public statements of steady progress while crews underground pumped mud for days at a time. Tunneling through the mountains was going to be hard, but it didn't have to be this hard. Thank you for watching.