No. Comparing to a control group is required, but it doesn't have to be a placebo control group. It's 100% acceptable, even required, to compare a new treatment to the standard-of-care treatment.
For example, if I were testing a new oral medication for Type 2 DM, I might randomize the participants to receive either the new med or metformin, which would act as the control. It would be unethical to randomize them to a placebo when they have a serious disease with a known effective treatment. This is my argument for GLP-1Ras, too.
It's tempting to ask for this, but it's not generally a good idea across the board.
When it comes to establishing efficacy (how well a drug works), a non-inferiority trial (what you're asking for) is suitable for that.
When it comes to establishing safety, it can be risky to allow for that. Here's a completely fictitious example:
Let's pretend Ozempic had a side effect where it deteriorated vision in 5% of people, while suffering from diabetes also had a 5% rate of vision deterioration. In conducting the original trial in diabetics, that safety signal would have been completely missed, since the people Ozempic saved from diabetes-related vision problems would now experience Ozempic-related vision problems. Perhaps those problems also take a couple years to show up. Approval is granted and nobody knows about that side effect, since just as many people in the control group as the active group had vision problems.
Next Wegovy (same active ingredient) is approved for weight loss (perhaps at a lower dose) based on a shorter trial in obese (but not diabetic people). Since it's a shorter trial length the vision problem isn't detected.
After that Zepbound is trialed (and we'll pretend it has that same side effect, but affects 8% of people). If Lilly ran Zepbound against a placebo then that risk would be discovered. Instead, Zepbound uses Wegovy (at its highest dose) as the control group. The small increase in vision problems isn't statistically significant and Zepbound is approved.
Next up is Retatrutide (let's say vision problems affect 10% of people). That would definitely be caught if the control group received a placebo, but since they receive Zepbound, it isn't noticed there either.
To be clear, I don't think any of these drugs cause vision problems. I'm just explaining why it's easy to miss safety signals in drug trials when the control group isn't receiving a placebo.