100lbs is obviously not a lot of mass for refueling to mean much. But, if Apollo needed to launch 3000t to land 16t on the Moon, a Starship with 100t of payload

100lbs is obviously not a lot of mass for refueling to mean much. But, if Apollo needed to launch 3000t to land 16t on the Moon, a Starship with 100t of payload and say, 300t of propellant, would need a ~75000t rocket from Earth. Efficiency improvements reduce this a bit, but not meaningfully. Hopefully you can see how that's kind of unrealistic and unscalable to do in one launch. Orbital refueling is a core technology necessary for sending useful amounts of payload around the solar system because the rockets themselves can't really get much larger without becoming completely uneconomical. Congress understood that orbital refueling would make their obscenely expensive SLS program unnecessary, which is why they effectively banned NASA from using the word "depot" (until Senator Shelby retired) under threat of canceling the entire space technology program, and strongarmed ULA into not pursuing an upper stage capable of carrying Orion and performing orbital refueling (which again, would've replaced a $2B per launch rocket with a couple of launches of a ~$150M rocket). The number of launches sounds like a lot, but consider that F9 is already launching 2-3 times each week, and Starship is being designed explicitly to not have the factors that are limiting F9 cadence from increasing further. The engines are better designed for reuse, the launch infrastructure is easier to build many copies of, the boosters always come back to land, the second stage integrates onto the booster at the launch pad etc. Also, as a side note, they'll need roughly the same number of refueling flights for Mars. The fuel requirements for going to the Lunar surface and Martian surface are actually pretty similar, since on Mars you can save a lot on using the atmosphere to slow down, while you can't on the Moon.