28 November 2017
Uranium, and the nuclear power industry, have both gone through some lean years of late. The uranium market has been oversupplied, with end-user inventories rising steadily and prices halving from around US$40/lb in 2015 to around US$20/lb in 2017. Downstream of uranium mining, an increase in anti-nuclear sentiment has been evident in many traditional markets since the Fukushima tragedy.
Our belief is that the long run future of uranium is more likely to be in either the “left hand tail of the distribution” (i.e. a low case world where nuclear generation dwindles in importance) or the “right hand tail of the distribution” (i.e. a green world where nuclear generation increases in importance), than a traditional distribution scenario (i.e. the average of tomorrow will look similar to the average of the past). Hence, the uranium story is a “tale of two tails”.
Why so? On the one hand, following the Fukushima tragedy, with the nuclear power sector facing challenges with respect to public confidence, and the reactor fleet ageing, there is a plausible long run story where the uranium industry faces grave difficulties as its downstream consumer fades into obscurity. On the other hand, as the world moves towards greater decarbonisation of stationary energy generation, one can just as easily tell the story of nuclear as a ‘green knight’, providing a competitive source of low net emissions baseload power to pair with renewables, with the resulting power mix a clear win for the environment. Finding a plausible narrative that falls neatly between those two in terms of uranium demand and price, and is equally compelling, is hard to do.
Over the years, we have developed analytical tools that enable us to wrestle with this sort of uncertainty. That is why we can confidently say that our future plans for Olympic Dam remain comfortably robust to a “low case world” for uranium. Of course, we will continue to scrutinise the plausibility of our low case(s), and the appropriateness of our high(s), as the relevant signposts presented by national energy, climate, nuclear power specific and uranium mining specific policies are revealed over coming years.
The global uranium (U308) market was around 73Kt in 2016.1 More than half that demand came from the Americas and eastern and western Europe combined. China was next with around one-seventh of global demand. The average age of the 129 reactor units in the Americas fleet is 37 years; western Europe’s 134 units are slightly younger (e.g. France 31 years, UK 33 years); and eastern Europe’s 52 units are slightly younger again at 28 years.2 China’s 33 units, by contrast, are 7 years old on average, while India’s 22 units are a more mature 21 years.
Why does the age of the fleet matter? It matters because when reactors (or any form of power generation capacity for that matter) come to the end of their effective life3, a decision is forced upon the owner. Should they extend or decommission? And if the latter, should they replace, and if so, with what?
The longer the forecasting horizon, the more important age of the fleet and its effective lifetime becomes. As our forecasting horizon extends to 2050, fleet lifetime is one of the major assumptions used in our range analysis. In our low case for uranium demand, we assume a reactor life of slightly less than half a century as a global average. In a high demand world, we stretch that to an average of almost 60 years.
The longer the forecasting horizon, the more important age of the fleet and its effective lifetime becomes.
The fleet ages above indicate the traditional consumers of uranium are likely to face ‘extend or decommission’ decisions in scale within the next decade. After the 2020s, the number of decision points in traditional markets would accelerate. While pure economics would be the starting point for those decisions, policy and politics will never be far from the debate.
...the traditional consumers of uranium are likely to face ‘extend or decommission’ decisions in scale within the next decade. After the 2020s, the number of decision points in traditional markets would accelerate.
Our estimates of the competitiveness of nuclear generation on a “levelised cost of electricity” basis (LCOE)4, which measures the present value of both upfront capital and operating costs over the lifetime of the asset per unit of energy produced, indicates that in the US, new nuclear plants are unlikely to be competitive with wind, coal or gas, so not only will ageing American plants likely be decommissioned, their lode could be picked up by other generation technologies. The economics look quite competitive in China, however, and that is where we estimate the most notable growth in nuclear capacity emerging. Nuclear is in middle of the LCOE pack in India right now, but it could be eclipsed by solar and then wind over the course of the 2020s, and it never catches coal.
It is important to note the LCOE measure is relatively flattering for nuclear competitiveness, as it captures the low operating costs and lengthy operating lifetimes that combine to offset the very large upfront costs associated with reactors. However, not every decision maker is able or willing to think in such terms. Some will require a quicker payback; others may be unable to fund the upfront capex. That is one of the reasons why even today the global nuclear fleet is dominated by the OECD, almost two decades into the “Asian century”.
While the 10 GW of nuclear power capacity that came on line in 2016 was the highest in many years5, only 3 GW of nuclear capacity began construction - 60% lower than the average of the previous decade. That was predominantly in China. Less investment does not necessarily mean substantial capex cost deflation. Indeed, we have observed that average build times for plants in recent times have been extending rather than decreasing6, which almost inevitably sees project economics deteriorating versus plan.
In our low-case we see global uranium needs having already peaked in the early 2000s, with demand plateauing for the next two decades. Growth in China, India, Latam and the Middle East is expected to be more than offset by declining demand in the traditional markets. However, in our green case, the demand peak would not occur until the late 2030s.
The difference between our low-case and a green world where environmental policies come to the fore – in other words, our tale of two tails is more than 300 operating reactors in 2050 (ranging from less than 200 to more than 500, from a base of around 400 today). In terms of compounded annual growth in uranium demand, in the ten years to 2025, the gap between the low and the high is around 6 percentage points. Between 2015 and 2050, the gap is 3 percentage points of compounded annual growth. It is also worth noting that there are some who see an even more important role for nuclear in any effective response to climate change.7
The difference between our low-case and a green world where environmental policies come to the fore is more than 300 operating reactors in 2050.
We reiterate that future plans for Olympic Dam are comfortably robust in the low demand case, with considerable upside if we end up in the right hand tail of the distribution after all.8
As a final observation, we note that when assessing the probability of a green case for uranium emerging, it is instructive to consider:
1. The share of total primary energy demand coming from electricity is rising inexorably.
2. The electrification of transport is a notable element in that trend.
3. Electric vehicles are only as green as the mix of power generation technologies that enable them.
The clear inference is that the challenge of greening the world’s energy appetite and moving towards a more favourable long run climate outcome cannot exclude how we generate the planet’s electricity. Nuclear generation is a well-established technology that can provide affordable life-of-asset base load power in a carbon conscious fashion. That is not to say that EVs and uranium are intrinsically linked: merely that from a climatic perspective, a prima facie carbon saving (moving to EVs) is only realised if the power they use is greener than the internal combustion method being displaced.
2. Japan has 54 reactors, the vast majority of which have been shut down since the Fukushima tragedy. The average age is similar to the eastern European fleet.
3. Estimates of plant lifetimes range from a conservative 40 years (equivalent to the original licensing period in the United States) to beyond 70 years in the views of some experts. In our estimates, we assume a minimum 40 year life that is extendable by 10 to 30 years, dependent on the generation (I, II or III) of the technology. See https://www.scientificamerican.com/article/nuclear-power-plant-aging-reactor-replacement-/
4. Estimates of LCOE are quite sensitive to input assumptions on long run discount rates and capacity factors, which means that major organisations and experts will disagree on precise levels. Indeed, some researchers feel that in the case of nuclear, the degree of uncertainty surrounding these input assumptions is so high that LCOE becomes an unsuitable metric. We acknowledge the imperfections of the technique, but in the absence of a superior alternative, LCOE can be appropriate as a rough guide to competitiveness.
6. There are exceptions to this trend of course, with some Chinese and South Korean reactor projects performing well versus plans. If we exclude these two countries, average build times are up around four years from a decade ago.
Bill Gates, George Soros and Li Ka-Shing have all either invested in uranium miners or nuclear technology since the beginning of 2016, backing a turnaround.
8. The variation in uranium grades across the Olympic Dam geology offers optionality for extracting value via the combination of a flexible mine plan and our marketing intelligence, which could also help buffer returns against a low price, as opposed to a low demand, world.