Nonalcoholic fatty liver disease (NAFLD) is the most common liver pathology. Here we propose tissue-cooperative, homeostatic model of NAFLD. During early stages of NAFLD the intrahepatic production of miR-122 falls, while the secretion of miRNA-containing exosomes by adipose increases. Bloodstream carries exosome to the liver, where their miRNA cargo is released to regulate their intrahepatic targets. When the deterioration of adipose catches up with the failing hepatic parenchyma, the external supply of liver-supporting miRNAs gradually tapers off, leading to the fibrotic decompensation of the liver and an increase in hepatic carcinogenesis. This model may explain paradoxical observations of the disease-associated decrease in intrahepatic production of certain miRNAs with an increase in their levels in serum. Infusions of miR-122 and, possibly, some other miRNAs may be efficient for preventing NAFLD-associated hepatocellular carcinoma. The best candidates for exosome-wrapped miRNA producer are adipose tissue-derived mesenchymal stem cells (MSCs), known for their capacity to shed large amounts of exosomes into the media. Notably, MSC-derived exosomes with no specific loading are already tested in patients with liver fibrosis. Carrier exosomes may be co-manufactured along with their cargo. Exosome-delivered miRNA cocktails may augment functioning of human organs suffering from a variety of chronic diseases.