We have recently proposed a new model for understanding how tumors evolve. To achieve successful "Tumor-Stroma Co-Evolution", cancer cells induce oxidative stress in adjacent fibroblasts and possibly other stromal cells. Oxidative stress in the tumor stroma mimics the effects of hypoxia, under aerobic conditions, resulting in an excess production of reactive oxygen species (ROS). Excess stromal production of ROS drives the onset of an anti-oxidant defense in adjacent cancer cells, protecting them from apoptosis. Moreover, excess stromal ROS production has a "Bystander-Effect", leading to DNA damage and aneuploidy in adjacent cancer cells, both hallmarks of genomic instability. Finally, ROS-driven oxidative stress induces autophagy and mitophagy in the tumor micro-environment, leading to the stromal over-production of recycled nutrients (including energy-rich metabolites, such as ketones and L-lactate). These recycled nutrients or chemical building blocks then help drive mitochondrial biogenesis in cancer cells, thereby promoting the anabolic growth of cancer cells (via an energy imbalance). We also show that ketones and lactate help "fuel" tumor growth and cancer cell metastasis and can act as chemo-attractants for cancer cells. We have termed this new paradigm for accelerating tumor-stroma co-evolution, "The Autophagic Tumor Stroma Model of Cancer Cell Metabolism". Heterotypic signaling in cancer-associated fibroblasts activates the transcription factors HIF1alpha and NFκB, potentiating the onset of hypoxic and inflammatory response(s), which further upregulates the autophagic program in the stromal compartment. Via stromal autophagy, this hypoxic/inflammatory response may provide a new escape mechanism for cancer cells during anti-angiogenic therapy, further exacerbating tumor recurrence and metastasis.