I agree that you should make sure your opamp has a supply voltage available all the time, or it will be difficult to control the startup behaviour.
However, if you decide to use PWM there are various circuits which will translate the voltage to PWM, often using a 555 timer. Some examples and explanation here :
https://electronics.stackexchange.com/questions/404793/voltage-to-pwm-circuit-need-to-understand-frequencyYou might also program a microcontroller to do the job : an arduino can measure the input voltage and generate a PWM signal quite easily, and you could perhaps add some other functions such as a startup guard time and safety interlock. Some thoughtful circuit design might still be needed to ensure it doesn't pulse the laser on at switch-on : typically, ensure that the default output state of the micro's pins (usually floating) will not turn the laser on. A light pulldown resistor might be sufficient to ensure this.
A microcontroller might seem overkill but if the pcb is ready-built it can often be cheaper or less work than constructing an analogue circuit as well as offering extra features such as safety lockout and linearisation.
It's possible that the laser PSU merely accepts the PWM input and filters it to generate an analogue voltage internally. I don't think the common CO2 laser PSUs PWM-modulate the actual laser supply voltage.
Neither the analog voltage input nor the PWM input are likely to produce a linear translation to laser power, and quite often it's not even monotonic : above a certain point power reduces instead of increasing and may reduce the tube life. So if you have a means of measuring actual laser power (eg measuring the rate of temperature rise of something exposed to the laser) you might be able to give yourself more meaningful control of cutting power by calibrating the system so that power % demanded by the controller is linear.