I'm in the very early stages of designing a low cost audio device. I will be using PCM (probably 16-bit 44.1khz) audio. I haven't settled on a microcontroller yet, but the stm32f74 arm cortex m7 looks promising because it has audio capabilities built in like SAI, I2S, etc. My question is, do I need external components, other than maybe an amplifier and obviously a speaker when using SAI and PCM? Or can I output directly to a speaker from the stm32?
SAI and PCM are digital audio formats. Typically, you would need both a digital to analog converter and an amplifier to drive a speaker from such a signal (speakers are analog creatures). There are many, many parts that can be used for this.
It is also possible to use a class-D amplifier that accepts a PCM signal directly, as such amplifiers can integrate the interpolator, DAC, and modulator in one unit, thus eliminating the need for a separate DAC. An example would be Maxim's MAX98357 which can be connected directly to an I2S or left-justified TDM audio interface, but there are also many other options available.
In short, yes, you need extra components, but the exact components will be determined by you and your requirements.