Depends in the goal.
If they are interested in Hardware, there is no better solution, than buying a hand full of cheap, "knock-off" Arduino Nano or RP2040 boards, a couple of LEDs and resistors and then let them Experiment with Arduino C, which teaches C programming in a fun and meaningful manner.
If they should get actual useful skill, start with Python. It's easy to learn. It's widely used for small, single-person projects, e.g. for gathering and analyzing data, running simulations, organizing something, interacting with APIs...
For any project, where I'd say, that it can be done quickly by a single person quickly on the side, Python is usually the best option. Even for experienced programmers. But especially for people that are new to programming or learning. It's easy, approachable and (because everything is handled by some library) quite safe to work with.
It's by far the most useful language for anyone that isn't a full software developer. And it's among the easiest to learn.
Also, when teaching Python, consider teaching with Jupyer documents. It immediately shows, what the short block of code is doing, instead of writing a huge and long script file, then executing it and struggling with debugging where students might not even know, how to start. An easy to host solution to supply Jupyer to a whole class is JupyterHub. If you aren't concerned about your students data, Google Colab is a JupyterHub for free.
And now, finally, if you want a course that has some practical use (e.g. data analysis and math stuff) while also doing fun things and being incredibly interactive because of the use of hardware... Well... It's python again.
Let them write a short branching story using Ren'Py. It's easy and will get them engaged. Go on and switch to Jupyter/Colab and do a bit of data scraping, data analysis/math or image filters. Basically introduce numpy, then pandas, then scipy, beautiful soup or openCV. Pure hard skills, that are applicable in many jobs. But realistically, due to time constraints in normal courses, all you want to do here, is numpy to get into it and automate math homework, followed by pandas, to load and automate excel files. And finally, let them do something with hardware, by programming a RP2040 in microPython. It's not as widely adopted as the Arduino ecosystem, but it's good enough for the simple stuff students will be able to learn, and you don't have to introduce a completely different language.