In　this　paper　we　consider　a　risk-averse　newsvendor　to　assemble　a　final　product　that　is　made　up　of　multiple　complementary　components,　where　both　the　product　demand　and　the　supply　capacity　of　each　component　are　random.　A　novel　approach　is　proposed　to　derive　the　first-order　condition　for　the　optimal　order　quantity　under　the　Conditional　Value-at-Risk　(CVaR)　criterion.　A　comprehensive　comparative　statics　analysis　is　conducted.　The　basic　model　is　also　extended　to　allow　dependence　among　the　random　demand　and　supply　capacities.　We　show　that　the　objective　function　remains　to　be　quasi-concave　if　those　random　variables　are　negatively　dependent,　and　that　the　positive　(negative)　dependence　of　those　random　variables　increases　(decreases)　the　optimal　order　quantity　and　objective　value.　Numerical　examples　are　provided　to　illustrate　some　of　the　main　results.　©　2021　Elsevier　B.V.