Going by total program cost, both of those are cheaper than Starship. SpaceX does this sleight of hand where they don't count their R&D cost in their calculations, but they do count it in other peoples' programs when comparing. SLS so far is about $2 billion in, and Starship seems to be at $2-3 billion. Vulcan is way cheaper than both.

SLS flight cost is $2B. Each launch. Did you forget to include the decade prior? It's so expensive that it can fly no more than once a year. The next flight is scheduled for Sep 2025, three years after the first flight. I bet you think that isn't absurd.

SLS use leftover engines from the Shuttle era. Yet it still cost 160 millions. Each. Thrown away after each flight.

Vulcan is a conservative design with no reuse. It also uses BE-4 engine, which costs them nothing to develop.

- Source for Vulcan total program cost? All I can find is their per-launch pricing. Tory Bruno himself apparently said that "new rockets typically cost US$2 billion, including US$1 billion for the main engine."

- Not sure where you're getting the SLS number, the GAO report at https://www.gao.gov/assets/gao-23-105609.pdf claims, "Since 2011, NASA has spent $11.8 billion to develop the initial SLS capability."

- I think the marginal cost is a fair question. The whole party line out of NASA was that this time Moon exploration would be sustainable. If each trip costs > $1B before you even put a payload on the rocket that's a big problem.

I completely agree with you on marginal cost, but I don't believe SpaceX to provide accurate numbers for their own rockets before they are actually built (they love wishful thinking) or their competitors' (making your competitors look bad is better for you).

Okay well respectfully I looked at the number you provided for the closest comparison rocket, SLS, and it was 6x too charitable.