Peak Uranium by Ugo Bardi from Extracted: How the Quest for Mineral Wealth Is Plundering the Planet

Figure 1. cumulative uranium consumption by IPCC model 2015-2100 versus measured and inferred Uranium resources. Figure 1 shows that the next IPCC report counts very much on nuclear power to keep warming below 2.5 C.  The black line represents how many million tonnes of reasonably and inferred resources under $260 per kg remain source: 2016 IAEA redbook Clearly most of the IPCC models are unrealistic.  The IPCC greatly exaggerates the amount of oil and coal reserves as wellSource: David Hughes (private communication)

This is an extract of Ugo Bardi’s must read “Extracted” about the limits of production of uranium.

You can find plenty of material saying there is are a lot of uranium reserves and resources  left (EMD 2019). The problem is, uranium requires oil to be mined, extracted, and processed, and oil is likely to start declining with the next 10 years. If you read my book “When Trucks stop running”, you’ll see why trucks can’t run on electric batteries or overhead wires, and without trucks, civilization collapses, so nuclear electricity is not going to solve the energy crisis, and leaves toxic waste our descendants will have to deal with for hundreds of thousands of years (Alley 2013)

Updates on peak uranium:

July 2016 Water power. Extracting uranium from seawater. Scientific American

  • Stephen Kung of the DOE’s office of Nuclear Energy said that terrestrial sources of uranium are expected to last for only another 100 to 200 more years.
  • It takes 8 weeks to extract 6 grams of uranium from seawater, or 0.75 grams per day,  It takes 27,000,000 grams to run a 1 gigawatt nuclear power plant for one year, so it would take 98,630 years to extract enough uranium from seawater to run just one nuclear power plant.

Here are 7 other posts from this great book:

Alice Friedemann  author of “When Trucks Stop Running: Energy and the Future of Transportation”, 2015, Springer and “Crunch! Whole Grain Artisan Chips and Crackers”. Podcasts: Practical Prepping, KunstlerCast 253, KunstlerCast278, Peak Prosperity , XX2 report ]

Bardi, Ugo. 2014. Extracted: How the Quest for Mineral Wealth Is Plundering the Planet. Chelsea Green Publishing.

Although there is a rebirth of interest in nuclear energy, there is still a basic problem: uranium is a mineral resource that exists in finite amounts.

Even as early as the 1950s it was clear that the known uranium resources were not sufficient to fuel the “atomic age” for a period longer than a few decades.

That gave rise to the idea of “breeding” fissile plutonium fuel from the more abundant, non-fissile isotope 238 of uranium. It was a very ambitious idea: fuel the industrial system with an element that doesn’t exist in measurable amounts on Earth but would be created by humans expressly for their own purposes. The concept gave rise to dreams of a plutonium-based economy. This ambitious plan was never really put into practice, though, at least not in the form that was envisioned in the 1950s and ’60s. Several attempts were made to build breeder reactors in the 1970s, but the technology was found to be expensive, difficult to manage, and prone to failure. Besides, it posed unsolvable strategic problems in terms of the proliferation of fissile materials that could be used to build atomic weapons. The idea was thoroughly abandoned in the 1970s, when the US Senate enacted a law that forbade the reprocessing of spent nuclear fuel.

A similar fate was encountered by another idea that involved “breeding” a nuclear fuel from a naturally existing element—thorium. The concept involved transforming the 232 isotope of thorium into the fissile 233 isotope of uranium, which then could be used as fuel for a nuclear reactor (or for nuclear warheads). 48 The idea was discussed at length during the heydays of the nuclear industry, and it is still discussed today; but so far, nothing has come out of it and the nuclear industry is still based on mineral uranium as fuel.

Today, the production of uranium from mines is insufficient to fuel the existing nuclear reactors. The gap between supply and demand for mineral uranium has been as large as almost 50% from 1995 to 2005, though gradually reduced the past few years.

The U.S. mined 370,000 metric tons the past 50 years, peaking in 1981 at 17,000 tons/year.  Europe peaked in the 1990s after extracting 460,000 tons.  Today nearly all of the 21,000 ton/year needed to keep European nuclear plants operating is imported.

The European mining cycle allows us to determine how much of the originally estimated uranium reserves could be extracted versus what actually happened before it cost too much to continue. Remarkably in all countries where mining has stopped it did so at well below initial estimates (50 to 70%). Therefore it’s likely ultimate production in South Africa and the United States can be predicted as well.

Table 1. The European mining cycle allows us to determine how much of the originally estimated uranium reserves could be extracted versus what actually happened before it cost too much to continue. Remarkably in all countries where mining has stopped it did so at well below initial estimates (50 to 70%). Therefore it’s likely ultimate production in South Africa and the United States can be predicted as well.

The Soviet Union and Canada each mined 450,000 tons. By 2010 global cumulative production was 2.5 million tons.  Of this, 2 million tons has been used, and the military had most of the remaining half a million tons.

The most recent data available show that mineral uranium accounts now for about 80% of the demand.  The gap is filled by uranium recovered from the stockpiles of the military industry and from the dismantling of old nuclear warheads.

This turning of swords into plows is surely a good idea, but old nuclear weapons and military stocks are a finite resource and cannot be seen as a definitive solution to the problem of insufficient supply. With the present stasis in uranium demand, it is possible that the production gap will be closed in a decade or so by increased mineral production. However, prospects are uncertain, as explained in “The End of Cheap Uranium.” In particular, if nuclear energy were to see a worldwide expansion, it is hard to see how mineral production could satisfy the increasing uranium demand, given the gigantic investments that would be needed, which are unlikely to be possible in the present economically challenging times.

At the same time, the effects of the 2011 incident at the Fukushima nuclear power plant are likely to negatively affect the prospects of growth for nuclear energy production, and with the concomitant reduced demand for uranium, the surviving reactors may have sufficient fuel to remain in operation for several decades.

It’s true that there are large quantities of uranium in the Earth’s crust, but there are limited numbers of deposits that are concentrated enough to be profitably mined. If we tried to extract those less concentrated deposits, the mining process would require far more energy than the mined uranium could ultimately produced [negative EROI].

Modeling Future Uranium Supplies

Uranium supply and demand to 2030

Table 2. Uranium supply and demand to 2030

Michael Dittmar used historical data for countries and single mines, to create a model that projected how much uranium will likely be extracted from existing reserves in the years to come. The model is purely empirical and is based on the assumption that mining companies, when planning the extraction profile of a deposit, project their operations to coincide with the average lifetime of the expensive equipment and infrastructure it takes to mine uranium—about a decade.

Gradually the extraction becomes more expensive as some equipment has to be replaced and the least costly resources are mined. As a consequence, both extraction and profits decline. Eventually the company stops exploiting the deposit and the mine closes. The model depends on both geological and economic constraints, but the fact that it has turned out to be valid for so many past cases shows that it is a good approximation of reality.

This said, the model assumes the following points:

  • Mine operators plan to operate the mine at a nearly constant production level on the basis of detailed geological studies and to manage extraction so that the plateau can be sustained for approximately 10 years.
  • The total amount of extractable uranium is approximately the achieved (or planned) annual plateau value multiplied by 10.

Applying this model to well-documented mines in Canada and Australia, we arrive at amazingly correct results. For instance, in one case, the model predicted a total production of 319 ± 24 kilotons, which was very close to the 310 kilotons actually produced. So we can be reasonably confident that it can be applied to today’s larger currently operating and planned uranium mines. Considering that the achieved plateau production from past operations was usually smaller than the one planned, this model probably overestimates the future production.

Table 2 summarizes the model’s predictions for future uranium production, comparing those findings against forecasts from other groups and against two different potential future nuclear scenarios.

As you can see, the forecasts obtained by this model indicate substantial supply constraints in the coming decades—a considerably different picture from that presented by the other models, which predict larger supplies.

The WNA’s 2009 forecast differs from our model mainly by assuming that existing and future mines will have a lifetime of at least 20 years. As a result, the WNA predicts a production peak of 85 kilotons/year around the year 2025, about 10 years later than in the present model, followed by a steep decline to about 70 kilotons/year in 2030. Despite being relatively optimistic, the forecast by the WNA shows that the uranium production in 2030 would not be higher than it is now. In any case, the long deposit lifetime in the WNA model is inconsistent with the data from past uranium mines. The 2006 estimate from the EWG was based on the Red Book 2005 RAR (reasonably assured resources) and IR (inferred resources) numbers. The EWG calculated an upper production limit based on the assumption that extraction can be increased according to demand until half of the RAR or at most half of the sum of the RAR and IR resources are used. That led the group to estimate a production peak around the year 2025.

Assuming all planned uranium mines are opened, annual mining will increase from 54,000 tons/year to a maximum of 58 (+ or – 4) thousand tons/year in 2015. [ Bardi wrote this before 2013 and 2014 figures were known. 2013 was 59,673 (highest total) and 56,252 in 2014.]

Declining uranium production will make it impossible to obtain a significant increase in electrical power from nuclear plants in the coming decades.


Alley, W. M., et al. 2014. Too Hot to Touch: The Problem of High-Level Nuclear Waste.Cambridge University Press.

EMD. 2019. EMD Uranium (Nuclear minerals and REE) committee annual report.

This entry was posted in Ugo Bardi, Uranium and tagged , . Bookmark the permalink.

5 Responses to Peak Uranium by Ugo Bardi from Extracted: How the Quest for Mineral Wealth Is Plundering the Planet

  1. Ugo Bardi says:

    Thanks for your interest in my book, “Extracted”, the summary of the problems with the future uranium suppy is well done. The only problem is that the model you cite was developed by Michael Dittmar, not by me, and I think his name should be cited explicitly – thanks! UB.

  2. Robert Hasspacher says:

    I wonder why this article doesn’t discuss uranium resources in Khazakhstan, a nations with very, very large untapped mineral resources. I also wonder why it doesn’t discuss the CANDU reactor, a heavy water reactor that efficiently can ‘burn’ unenriched uranium. Also, the Russians recently deployed a fully functioning breeder reactor.

    China is currently making huge strides its quest to build a Molten Salt Reactor, a reactor that would produce very little waste and consume fuel VERY efficiently. One MSR saw 20,000 hours of flawless testing at Oak Ridge National Laboratories in the *1950s*. That’s right, over 60 years ago.

    • energyskeptic says:

      this is a book about MINING, not uranium. Enriched uranium has a NEGATIVE EROI. Only a breeder reactor could extend remaining uranium, and Russia’s breeder reactor has only been up a couple of months, too soon to tell if this will work or not. But who cares? If you can’t electrify heavy-duty transportation, as well as all the industrial uses (steel, cement, glass, etc) which often don’t have an electricity sourced process, nuclear power is irrelevant. This is a liquid fossil fuel problem.

      Breeder reactors.

      You’d need 24,000 Breeder Reactors, each one a potential nuclear bomb (Mesarovic)

      Breeder reactors are much closer to being bombs than conventional reactors – the effects of an accident would be catastrophic economically and in the number of lives lost if it failed near a city (Wolfson).

      The by-product of the breeder reaction is plutonium. Plutonium 239 has a half-life of 24,000 years. How can we guarantee that no terrorist or dictator will ever use this material to build a nuclear or dirty bomb during this time period?

      Assume, as the technology optimists want us to, that in 100 years all primary energy will be nuclear. Following historical patterns, and assuming a not unlikely quadrupling of population, we will need, to satisfy world energy requirements: 3,000 “nuclear parks” each consisting of, say, 8 fast-breeder reactors. These 8 reactors, working at 40% efficiency, will produce 40 million kilowatts of electricity collectively. Therefore, each of the 3,000 nuclear parks will be converting primary nuclear power equivalent to 100 million kilowatts thermal. The largest nuclear reactors presently in operation convert about 1 million kilowatts (electric), but we will give progress the benefit of doubt and assume that our 24,000 worldwide reactors are capable of converting 5 million kilowatts each. In order to produce the world’s energy in 100 years, then, we will merely have to build, in each and every year between now and then, 4 reactors per week! And that figure does not take into account the lifespan of nuclear reactors. If our future nuclear reactors last an average of thirty years, we shall eventually have to build 2 reactors per day to replace those that have worn out. By 2025, sole reliance on nuclear power would require more than 50 major nuclear installations, on the average, in every state in the union.

      For the sake of this discussion, let us disregard whether this rate of construction is technically and organizationally feasible in view of the fact that, at present, the lead time for the construction of much smaller and simpler plants is seven to ten years. Let us also disregard the cost of about $2000 billion per year — or 60 percent of the total world output of $3400 billion — just to replace the worn-out reactors and the availability of the investment capital. We may as well also assume that we could find safe storage facilities for the discarded reactors and their irradiated accessory equipment, and also for the nuclear waste. Let us assume that technology has taken care of all these big problems, leaving us only a few trifles to deal with.

      In order to operate 24,000 breeder reactors, we would need to process and transport, every year, 15 million kilograms (16,500 tons) of plutonium-239, the core material of the Hiroshima atom bomb. Only 10 pounds are needed to construct a bomb. If inhaled, just ten micrograms (.00000035 ounce) of plutonium-239 is likely to cause fatal lung cancer. A ball of plutonium the size of a grapefruit contains enough poison to kill nearly all the people living today. Moreover, plutonium-239 has a radioactive life of more than 24,000 years. Obviously, with so much plutonium on hand, there will be a tremendous problem of safeguarding the nuclear parks — not one or two, but 3000 of them. And what about their location, national sovereignty, and jurisdiction? Can one country allow inadequate protection in a neighboring country, when the slightest mishap could poison adjacent lands and populations for thousands and thousands of years? And who is to decide what constitutes adequate protection, especially in the case of social turmoil, civil war, war between nations, or even only when a national leader comes down with a case of bad nerves. The lives of millions could easily be beholden to a single reckless and daring individual.

      Mesarovic, Mihajlo, et al. 1974. Mankind at the Turning Point. The Second Club of Rome Report. E.P. Dutton, 1974 pp. 132-135

  3. John Weber says:

    Nuclear power has various negatives.
    The first is simply; nuclear fission would not exist without fossil fuels. I mean the research, experimentation, testing, construction, maintenance, storage and decommissioning. It is an extension of the fossil fuel supply system.

    The danger of waste is the second reason. The thousands of years required to allow some of the waste to become non-toxic requires a stability in the world that history shows no evidence.

    Chernobyl and Fukushima speak for themselves as the third reason. They are the Black Swans (Nassim Nicholas Taleb) we can perhaps imagine, but not predict. One was direct human error; one was natural catastrophe with of course indirect human error. How many more? These are some of the unintended consequences of nuclear power.

    Each of the above are premium and primary reasons why I do not support nuclear power; however, there is a fourth that is truly my main reason. We humans, as a group and as most individuals, are not smart enough, wise enough, moral enough or mature enough to handle such power. Look at the state of the world – all the angry people, all the wounded people, all the deluded people.

  4. Edward Marrs says:

    Maybe I am missing something, but I do not see how nuclear energy is anything other than another techno fantasy like wind or solar. This because technologies do not make more energy available so they do not address the problem of a diminishing energy supply. Nor do they increase the conservation of energy. But the techno fantasy imagines it does because it does not consider the energy costs of increasing complexity. Any savings from efficiency at point of use has already been spent to create the complexity required to bump up efficiency. The actual efficiency of a device or scheme must include all the associated energy required to produce, transport and maintain a device yet I never see these factors mentioned. When you look at a new Prius only the end use efficiency is mentioned. Increasing complexity most often means more and/or more unique, more complex parts and more complex relationships between parts. The creation and organizing of these parts have real energy costs.

    What high tech energy saving scheme does not utilize a computer? If it does it is game over for saving energy before it even starts. Computers are not made from fairy dust. The computer industry is mind boggling in scope, size and has an appetite for energy commensurate with its size. It is also global so there are considerable and constant transportation energy costs incurred even before manufacturing begins. The computer product line requires constant input of specific raw materials the procurement of which takes big energy as in drilling, pumping, mining, smelting and forming. Then there is the energy needed to construct, operate and maintain the infrastructure involved in materials procurement and product manufacturing. Nor does all this happen without people. They have to eat and live somewhere so there are all the associated energy costs of sustaining a constant pool of employees. These energy costs are not apparent when we go drive away in our Prius so we don’t consider them to be relevant. But those costs have all been met and that energy used is not recovered by efficiency gains at the point of end use.

    Efficient vehicles are also computerized so how can we realistically say an efficient vehicle saves energy by just looking at mileage gains. The best thing the auto and trucking industry could do would be to produce fewer vehicle models that have fewer parts. Presently there are a great many vehicles manufactures all over the world, each with different existing models and legacy models, they have different engines, body parts, transmissions, electrical components each with their own different parts which are not interchangeable. This huge inventory of parts has to be made, moved and maintained and that takes energy. Eventually the cost of increasing complexity cannot be met.