[photochemsyn]

Preconcentration of CO2 from air makes a lot of sense as you can then do various types of industrial chemistry on a pure CO2 stream at greater efficiency. From here you can go towards either methane and long-chain fuel hydrocarbons (Sabatier, modified Fischer-Tropsch, etc) or, as in this paper, towards solid carbon forms.

Making something like graphite from pure CO2 has certain advantages as well (easier to get purity) and graphite electrodes are used in scrap steel recycling and other industries.

For comparison, see the ISS use of Sabatier reaction and some issues they had with catalyst poisoning:

https://ntrs.nasa.gov/search.jsp?R=20140002591

This indicates that power plant emissions, typically contaminated with sulfur / arsenic / mercury / nitrogen etc. , at about 10% CO2 as I recall, would be a very poor option relative to direct air capture.

As far as scaling, even existing systems (see ISS) could be scaled fairly rapidly and would be able to produce enough fuel for specialized uses, i.e. plausibly supplying SpaceX / ULA/ etc. rocket launches as a first step, then moving to supply airports with jet fuel for long-distance travel at a much larger scale.