At the 14th IAEA Conference on Plasma Physics and Controlled Nuclear Fusion Research, Wurzburg, Germany, several senior experts on magnetic fusion development shared with me their view that finally, the world fusion effort has accumulated enough understanding of the requirements and conditions for ignited long-pulse burn, that a tokamak experiment to achieve that objective could be built and expected to work with high confidence. Although this is not yet a unanimous view of all the IAEA participants, I do agree with this majority view. As a reflection of this growing confidence, the governments of the United States, the European Community, Russia and Japan have recently commissioned the Engineering Design Activity (EDA) phase of the International Thermonuclear Experimental Reactor (ITER). The technical vision of ITER presented at that conference by the new ITER director, Paul-Henri Rebut, was widely accepted as a most practical and workable approach to achieving the above objective, assuming a successful demonstration of the proposed gas-target divertor during the EDA R&D phase. Having pioneered an early version of the gas-target divertor [l], I also agree with this assessment of the technical viability of Rebut’s proposed design. At the end of his ITER presentation, however, Rebut touched on an issue of much lower confidence among the fusion community, namely, that ITER would demonstrate “a reasonable cost per unit of fusion power”. Estimating myself that the magnet set (TF + PF) of his proposed design would store about 80 GJ, would have a cold mass exceeding 13000 tons, including magnet support structure, and would alone cost more than 5 B $ at 4OO$/kg (for 13 T max, Nb,Sn), I take a charitable interpretation of Rebut’s remark to mean that ITER demonstrate a reasonable unit cost, not by itself alone, but viewed as a significant step toward a practical power reactor. This then shifts the confidence issue on economics to a judgment on how much improvement in unit cost can be reasonable expected in tokamak reactors beyond ITER. Improvement beyond ITER was an important motivation behind the recent recommendation by the US Secretary’s Energy Advisory Board (SEAB) to proceed with the tokamak Physics Experiment (TPX) at the Princeton Plasma Physics Laboratory. This prompts the question as to how much improvement in unit cost would be needed from the present embodiment of ITER to an economically-competitive power reactor, assuming a 1 GW, power reactor would be at least no larger in size or mass than ITER. I would estimate the present expectation of ITER’s unit cost, based on a “median” performance with operation at prudent margins below accepted limits, to be = 10 B $/l GW,,, or, equivalently, 25$/W,, as if hot blankets and a 40% efficient conversion system were included. For a “competitive” commercial power reactor, I would take 2.5$/W, (based on total capital as in the ITER number). Thus, by this figuring, I estimate than an improvement factor of about 10 is needed (I wouldn’t quibble between 7 and 14).