Loading of fructose (F) in mammals prevented ATP synthesis, and as a consequence, increased feed intake; phosphate (Pi) loading prevented the effect of F. The objective of this work was to link hepatic ATP synthesis and feeding behavior. Accordingly, 4 ruminally cannulated multiparous cows (15–26 d postpartum) were used in a 4 × 4 Latin square. Periods were 24 h, including 1 h for infusions and 23 h for recovery. Cows were infused to the abomasum with 0.6 mol/h of F or glucose (G) and 0.3 mol/h of Pi (as NaH₂PO₄) or NaCl (Cl). Liver biopsies were taken 30 and 60 min post-infusion and examined for Pi content and expression of genes involved in the TCA cycle (α-ketoglutarate dehydrogenase; KD), fructolysis (Fructokinase), oxidative phosphorylation (ATP5b) and gluconeogenesis (cytosolic Phosphoenolpyruvate carboxykinase; PEPCK). Ribosomal protein S9 was used as a housekeeping gene. Effects of treatments were analyzed by ANOVA, the fixed effects were F infusion, Pi infusion, period, and their interactions. Cow was included as a random effect. The F+Cl treatment increased DMI in the first 4 h after infusion began by 28–47% (F-Pi interaction, P = 0.003; 5.14, 3.50, 4.00 and 3.59 kg for the F+Cl, F+Pi, G+Cl and G+Pi, respectively), demonstrating that Pi infusion prevented the effect of F. Moreover, the G+Pi treatment increased hepatic Pi 30 min into the infusion by 49–89% (F-Pi interaction, P = 0.07; 8.19, 10.4, 10.1 and 15.5 μmol/g for the F+Cl, F+Pi, G+Cl and G+Pi, respectively). F also increased hepatic RNA expression of KD 60 min into the infusion by 2.25-fold (P = 0.01). Pi tended to increase Fructokinase expression 30 min into the infusion by 2.9-fold (P = 0.09). F increasing DMI while Pi preventing the effect of F, support a connection between ATP synthesis and DMI in cows postpartum. The increase in hepatic KD expression may suggest a shift toward ATP producing pathways (e.g., TCA cycle) as a response to a reduction in ATP rather than reducing ATP consuming pathways (e.g., gluconeogenesis). The increase in Fructokinase expression suggest that fructolysis is regulated by the availability of Pi rather than the availability of F. These hypotheses needs to be further examined.