Jeffrey LewisChina Exports SWU

Did you know that China occasionally exports enrichment services?

I didn’t.  But, sure enough, they do from time to time.

Apparently China used to export enrichment services to the United States, but those exports stopped in 2007.  They began again last year.  Here is a chart created from data released by the Energy Information Administration in 2010 and 2012.

Purchases of enrichment services by owners and operators of U.S. civilian nuclear power reactors by origin country and year, 2007-2011 
Thousand separative work units (SWU)
Country of Enrichment Service (SWU-origin) 2005 2006 2007 2008 2009 2010 2011
China W W 0 0 0 0 W
France 1,830 2,154 613 556 895 W W
Germany 583 818 681 468 1,059 681 1,539
Netherlands 581 960 1,703 1,038 1,345 2,292 1,506
Russia 5,059 4,724 6,176 4,793 5,478 5,055 5,308
United Kingdom 1,379 2,000 1,939 2,195 2,940 2,119 2,813
Europe1 W W W W W W 670
Other 2 W W W W W W 0
Foreign Total 10,343 11,808 12,729 10,709 13,115 11,526 12,395
United States 1,052 1,630 1,473 1,890 4,102 2,251 2,434
Total 11,394 13,437 14,202 12,599 17,217 13,776 14,829
Average Price (US$ per SWU) 106.57 114.58 121.33 130.78 136.14 136.12

The “W” is for “withheld”– “W = Data withheld to avoid disclosure of individual company data.”

As it turns out, however, the data is recoverable.  The Convention on the Physical Protection of Nuclear Materials requires that states party make “prior arrangements …specifying time, place and procedures for transferring transport responsibility.”  In the United States, that means a declaration to the NRC, which looks like this.

Westinghouse imported 118 tons of UF6 enriched to ~5 percent in May and November 2011.

If that is the whole shipment, then that should work out to something like 580 tons of Seperative Work Units (SWU) — a surprisingly large order.  (Someone should check my math! I am notoriously unreliable when it comes to arithmetic.  And I used the handy new URENCO SWU Ap on my iPhone.  So ….)

There are two interesting things about this import.

First, the China Nuclear Energy Industry Corp (CNEIC) needs to get better at packaging UF6 for transport.  The NRC’s ADAM database contains a number of documents detailing the violations that occurred with the shipments.   (Recommended search string: CNEIC UF6 Westinghouse) Generally speaking, the problems related to the valves, although some of the cylinders were not properly marked.

Second, URENCO and other enrichers have to be terrified that this is a sign of things to come.  Although much of the uranium trade is based on long-term contracts (rather than spot sales), over time China might very well muscle into this business with the indigenous centrifuge plant it has constructed at Lanzhou.  I wouldn’t be surprised to see words like “dumping” get thrown around in the next decade or so.


  1. krepon (History)

    Hasn’t China provided fuel for Tarapur, as well?

  2. Cthippo (History)

    UF6 is, basically, a fungible commodity, and hence anyone with the capability to enrich can sell it. Why not China.

    As for URENCO’s concerns, I wonder how much of the cost of enrichment is in labor costs, which is where China’s primary advantage lies. Seems like the capital costs of the plant and market factors would be more important than labor costs.

    • rwendland (History)

      I suspect, as for nuclear power plants, the dominant cost factor for a new enrichment plant is cost-of-capital. I believe private financiers in the west are looking for around a 12% return (including the risk premium) for nuclear power finance, and I’d guess it would be similar for a new enrichment plant. I suspect large Chinese companies can get capital very much more cheaply from the state-backed chinese banks – where essentially the state is carrying much of the risk.

      So, if there is a global need for expansion or replacement of enrichment capacity, new plant would have lower costs in China unless the western companies can tap into cheaper military/govt finance somehow. But as the western “nuclear renaissance” looks utterly stalled, it doesn’t seem obvious we need much new enrichment capacity – does anyone have a pointer to a paper estimating the lifetimes of existing enrichment plant globally?

  3. Steve (History)

    CNEIC has been supplying LEU since the early 1980s:

    From China Today: “Starting in 1981, Chinese low-concentration uranium products entered international markets and the quality of the products was well received.”

    The buyers? Could be Argentina and South Africa:

  4. Jon (History)

    Yes, UF6 is a fungible commodity. However, since Westinghouse will be supplying the first core loads for the AP1000s they are currently building in China, it would seem economically advantageous for the Chinese to supply their own UF6. 118 tons of 5% enriched UF6 would equate to 79.8 tons of 5% enriched U. About right for a 1000MW core?

  5. Miles Pomper (History)

    Capital costs are by far the dominant costs in enrichment plants–roughly two-thirds of the cost– according to Geoffrey Rothwell’s 2009 Market Power in Uranium Enrichment analysis in Science and Global Security which is the best out there. Labor, electricity, and materials split the rest.

    Rothwell also noted that China had about a 1-2 percent of the US market from 1995-2008.

    It is not clear what would account for the recent upswing in China’s market share but in a purely free market I doubt the Chinese could compete with the Russians who have cheap finance, efficient technology, etc. However, the Russians are currently limited by the terms of the megatons-megawatts agreement….

    • Jeffrey (History)

      The Russians believed that they could maintain a technological lead over the Chinese while selling them centrifuges that were one generation behind. I wonder about the wisdom of the decision.

    • rwendland (History)

      Thanks for link to that interesting paper.

      Rothwell is only using 5% as the cost of capital in his model (see table A.3), that implies cost-of-capital is about 66% of levelized SWU cost. If new western plant had to use private sector capital at a more plausible cost-of-capital in the 10% to 12% region, cost-of-capital would likely be around 80% to 90% of levelized SWU cost.

      NB Amused that some of Urenco (Capenhurst) centrifuges had been running non-stop since 1982 (as at 2005), and “If machines are stopped, risk is they will not start again”! (pdf page 12)

Pin It on Pinterest