Now I have enough high-quality evidence to give a thorough, nuanced verdict. Here it is:

The explanation you've quoted captures the spirit of the current scientific debate well, but contains one factual error about the piericidin experiment and oversimplifies the final conclusion. Let me break it down claim by claim.

Accurate. This is the classic textbook mechanism. Metformin is believed to reversibly inhibit complex I of the mitochondrial electron transport chain, raising the AMP:ATP ratio, which activates AMPK, which in turn suppresses hepatic gluconeogenesis. This is confirmed across multiple sources including Goodman & Gilman's and Harrison's (2025).

Accurate — and well-supported. Inhibiting complex I requires millimolar concentrations of metformin, but clinical plasma levels are in the micromolar range (~1–10 µM). This is one of the strongest challenges to the traditional model, confirmed by multiple recent reviews (PMC10072049, Drewe et al., Pharmacol Rev 2026).

One important exception: enterocytes (intestinal cells) do accumulate metformin to millimolar concentrations due to active transport — so complex I inhibition may still be relevant in the gut. The 2026 review by Drewe et al. specifically concludes that "inhibition of mitochondrial complex I in different organs appears to be an important mechanism," especially enterocytes where the drug is most concentrated.

The experiment is real, but the details are imprecise. The piericidin experiment is a genuine piece of evidence in this field. However:

So the conclusion drawn from piericidin is correct — complex I alone is insufficient — but the experimental description is slightly inaccurate in framing it as a simple liver infusion study.

Accurate, and well-supported by current evidence. This is the consensus of multiple recent high-quality reviews. Metformin's mechanism is now understood to be multifactorial, involving:

Directionally correct but incomplete. The "low-energy state" concept is real — metformin does mildly reduce ATP production, which is part of how it activates AMPK. However, the final sentence in your quote presents this as the settled, unifying explanation, which is an oversimplification. Current evidence suggests the glucose-lowering mechanism is multifactorial and that in the liver at clinical concentrations, the cytosolic redox effect (via complex IV / GPD2 inhibition) may be more important than simply "slightly less ATP." The Drewe et al. 2026 review explicitly concludes the mechanism is multifactorial across gut, liver, and muscle.

The explanation reflects a genuinely scientifically literate understanding of the debate, and the core argument is sound. The main gap is that it stops short of naming what does appear to explain metformin's effect at clinical concentrations: complex IV inhibition affecting the redox state, lysosomal AMPK activation, and intestinal GLP-1-mediated pathways.

At the cellular level, metformin was traditionally thought to act primarily by inhibiting complex I of the mitochondrial electron transport chain, thereby raising the AMP:ATP ratio, activating AMPK, and ultimately suppressing hepatic gluconeogenesis. However, this model is challenged by a critical pharmacokinetic problem: inhibiting complex I requires millimolar concentrations of metformin, whereas clinical plasma levels remain in the micromolar range — concentrations too low to meaningfully inhibit complex I in most tissues. This is further supported by experiments using piericidin A, a potent and selective complex I inhibitor, which — unlike metformin — fails to suppress gluconeogenesis or alter the cytosolic redox state (lactate:pyruvate ratio) in hepatocytes, demonstrating that complex I inhibition alone is not sufficient to replicate metformin's antihyperglycemic effect in the liver. More recent evidence points instead to inhibition of complex IV, which indirectly suppresses mitochondrial glycerophosphate dehydrogenase (GPD2), raises cytosolic NADH/NAD⁺, and reduces gluconeogenesis from lactate and glycerol — independently of AMPK activation. In parallel, metformin can activate AMPK through a lysosomal pathway involving the PEN2-AXIN-LAMTOR1 complex, again independently of mitochondrial complex I. Beyond the liver, significant glucose-lowering effects arise from intestinal mechanisms: in enterocytes, where metformin accumulates to millimolar concentrations via active transport, complex I inhibition is pharmacologically relevant and drives inhibition of glucose absorption, stimulation of GLP-1 secretion, and activation of the gut-brain-liver axis to further suppress hepatic glucose output. The net result is that metformin does create a mild cellular energy deficit that contributes to its effects, but this is not the single unifying mechanism — the overall glucose-lowering action of metformin is multifactorial, operating across the intestine, liver, and skeletal muscle through distinct but complementary pathways.