Technology that removes carbon dioxide directly from the atmosphere is operating commercially. It works. The question is entirely about magnitude.
The scale gap
Global emissions are measured in tens of billions of tonnes annually. Total installed direct air capture capacity is measured in tens of thousands of tonnes.
That is a difference of roughly six orders of magnitude. Closing it is not an engineering refinement; it is building an industry larger than any that currently exists.
The energy problem
Carbon dioxide is dilute in air – a few hundred parts per million. Extracting something that dilute takes substantial energy, and that energy has to be clean or the exercise is pointless.
At meaningful scale, capture would consume a large share of global clean electricity, competing directly with using that electricity to avoid emissions in the first place.
Avoiding a tonne is cheaper than removing a tonne, by a wide margin, and will remain so.
Where it genuinely helps
Emissions with no practical alternative. Some industrial processes emit carbon dioxide as chemistry rather than as combustion, and cannot be electrified.
And historical emissions, which no amount of avoidance addresses. If atmospheric concentration is to fall rather than merely stop rising, removal is the only mechanism.
The moral hazard argument
That the existence of removal reduces pressure to cut emissions. There is evidence for this in how it is invoked in corporate commitments, where planned future removal substitutes for present reduction.
The technology is not responsible for how it is used rhetorically, but the pattern is real and worth noticing.
The reasonable position
Necessary eventually, insufficient always, and dangerous as an excuse now.