Certainly you can generate your own clone. But the issue is, if you are using different expression systems (ex.mammalian) you need to codon optimize and synthesize your construct (note: cas9 is more than 3 kb). Further you might need to clone this construct in your plasmid of interest. hence this is waste of money and time.
Moreover addgene plasmids are cheap, sequence verified and the related protocols are discussed in the associated publications. hence it will be a safe bet to go with addgene.