The classical view of the AMP-activated protein kinase (AMPK) system is that it is a sensor of energy that monitors the cellular concentrations of AMP and ATP, and which regulates energy balance by stimulating catabolism and inhibiting anabolism whenever the cellular AMP:ATP ratio rises [1]. The kinase is a heterotrimeric complex of a catalytic α subunit and regulatory β and γ subunits. The γ subunits contain two tandem domains that bind the regulatory nucleotides, AMP and ATP, in a mutually exclusive manner. The kinase is only active after phosphorylation at a conserved threonine residue in the α subunit (Thr-172) by upstream kinases, of which the most important is the tumour suppressor LKB1. LKB1 appears to phosphorylate Thr-172 continually, but binding of AMP (but not ATP) to the γ subunit inhibits Thr-172 dephosphorylation. Since any fall in the cellular ATP:ADP ratio is amplified by adenylate kinase into a much larger rise in the AMP:ATP ratio, this mechanism acts as a sensitive switch that converts the kinase into its active, phosphorylated form. In addition, the phosphorylated kinase is allosterically activated 10-fold by AMP binding; the combined effects of phosphorylation and allosteric activation result in >1000-fold activation. In some cells, Thr-172 can also be phosphorylated by CaMKKβ, a calmodulin-dependent protein kinase. This occurs in response to a rise in intracellular Ca2+ and does not require any increase in AMP. The β subunits of the AMPK heterotrimer contain a central glycogen-binding domain (GBD) that is conserved in all eukaryotic orthologues. Although this domain is known to cause binding of AMPK to glycogen in intact cells [2,3], its physiological function has been unclear. We have now shown that it is a regulatory domain and that binding of glycogen to it causes allosteric inhibition of AMPK. Glycogen is a polymer of glucose units joined by α1-4 linkages, with occasional branches formed by α1-6 linkages. A major problem with the study of glycogen as a regulatory molecule is that it does not have a defined structure, but varies both in size and branching content. Based on findings suggesting that the degree of branching affected its inhibitory potency, we have synthesized small α1-4 linked glucose oligosaccharides containing single α1-6 branches, and have shown that these are potent allosteric inhibitors of AMPK. The most potent oligosaccharide gives half-maximal inhibition at 90 μM, and also markedly inhibits phosphorylation of Thr-172 by LKB1 and CaMKKβ. Inhibition is due to binding to the GBD, because point mutations in the latter than abolish glycogen binding also abolish inhibition. One of the physiological targets of AMPK is muscle glycogen synthase (mGS), which is inactivated by phosphorylation at Ser-7 [4]. A curious paradox is that although AMPK is activated by ATP depletion during exercise, mGS is usually found to be dephosphorylated and activated following exercise, as long as the exercise had been sufficiently prolonged to cause significant glycogen depletion. We believe that the regulation of AMPK via the GBD may explain this paradox. The outer tier of glycogen (representing the glucose that can be released by glycogen phosphorylase without the need for the action of debranching enzyme) contains up to one third of all of the glucose units in a single molecule. Theoretical studies suggest that in a fully synthesized molecule of glycogen the outer chains are so tightly packed that the branch points would not be accessible. We suspect that under these conditions AMPK binds to glycogen but is not inhibited, so that it would phosphorylate mGS, thus exerting a feedback inhibition of further extension of the outer chains of glycogen. We also propose that when exercise commences, phosphorylase removes some of the outer chains, exposing the branch points which would then bind AMPK and cause inhibition. mGS would now be dephosphorylated and activated by the glycogen-bound forms of protein phosphatase-1, so that it was ready to replenish glycogen stores as soon as exercise ceased. This hypothesis may also explain why insulin-stimulated glucose uptake and glycogen synthesis is enhanced following a single bout of exercise, as long as the exercise bout was sufficient to cause significant glycogen depletion. It can also answer another unsolved question: how do cells “know” when their glycogen stores are sufficient, and conversely how do they “know” when they are insufficient and need replenishing? These are important questions because insulin resistance, the primary cause of type 2 diabetes and the metabolic syndrome, can be viewed as a mechanism to limit the amount of nutrient that cells can store.