As summarized in the report:

"ARPA-E's ALPHA program seeks to create and demonstrate tools to aid in the development of new, lower-cost pathways to fusion power and to enable more rapid progress in fusion research and development. Assuming we achieve excess energy production from a fusion core, a next critical step is to understand the capital costs associated with a fusion power plant. An initial capital cost study was performed by Bechtel National’s power plant cost team, augmented by members of the fusion community who have published in fusion cost estimating (Woodruff Scientific and Decysive Systems). The study was based upon four conceptual designs for a fusion core and present-day standard components for the balance of plant (heat exchanger, turbines, etc). The cost study team did not attempt to compose a levelized cost of electricity (LCOE) for a fusion power plant given the juxtaposition of the conceptual nature of the fusion core designs versus the level of granular knowledge commonly used as input for calculating an LCOE.

"Across four unique fusion core approaches, the estimated cost of the core in all cases constituted less than half of the total direct cost, and, in some cases, was not even the most expensive component. Accordingly, among the four fusion approaches considered, there is no outlier approach that should be singled out for emphasis or de-emphasis.

"We found that neither neutronics nor tritium handling were major capital cost drivers. However, much engineering work remains to reach solutions that (1) appropriately account for the effects of the high energy neutrons on various components, and (2) address tritium fuel extraction, transfer, and storage, among other considerations.

"The uncertainty in the pulsed power system design and lifetime under power plant conditions should be a focus area in future work. Using a reasonable range for the cost of the power input systems, sensitivity analysis found that power systems comprised 5-20% of the total direct cost (which includes reactor core, structures and site, turbine plant, etc.).

"The up-front capital costs of a fusion power plant, as with many other power plant approaches, are likely to hinge heavily upon the scale of the plant and the balance of plant components. For example, if one treats fusion like fission by borrowing its scaling factor of roughly 0.55, these 150 MWe designs might scale to $2-6 per Watt at a 1 GWe scale.¹"

Conclusions:

"First, we conclude it is best to aggressively pursue multiple options for the fusion core in light of the cost study finding that the economics of a fusion plant are relatively insensitive to which of the four fusion approaches is chosen. Fortunately, the cost of pursuing multiple approaches does not appear to be prohibitive— the four approaches considered in this cost study are believed to follow inherently more affordable development paths than the more mature magnetic or inertial confinement approaches.

"Second, it would be prudent to link the ramp-up of the expensive engineering effort for the tritium systems and neutronics to marked progress on the fusion core. While tritium systems and neutronics will be important, their costs will not dominate the initial capital cost of a fusion power plant."

footnote:
¹ A common approach to estimating "to-be" power plant capital cost is via a scaling law based on "as-built" capital cost:
   Cost_to-be ~ Cost_as-built x (Power_to-be / Power_as-built) sf,
where sf is the empirically determined scaling factor. There is today no "as-built" cost for a fusion power plant, but one could apply a mid-range scale factor for nuclear plants (~0.55) in order to examine potential cost.