Puzzling over the term 'intra-cellular fermentation.' Literally, this means fermentation within the cell, right? How do the metabolizing organisms wend their way inside the grape cells? This seems like it would be more of a virus-y thing.
Perhaps, however, the term is used loosely to indicate fermentation within the grape, rather than its constituent cells; where bacteria, having penetrated the skin (through breakage or, say, stem holes) attack and break down the grape cells as substrate.
Browsing Goode's article (admittedly rapidly), he seems to be saying that there is a graduated sequence of fermentation processes in these conditions, where free oxygen circulation is impeded or interdicted. The initial 'intracellular' fermentation is dominated by non-yeast microbes, which use the abundant CO2 molecules to oxidize this substrate (including malic acid, from damaged/preyed-upon grape cells). When the alcoholic byproduct of this slow initial metabolic process reaches about 2%, the grape cells die, their juice is released, and large-scale sugar fermentation proceeds, thenceforth dominated by yeast (until the yeast population pollutes itself out of existence).
A relevant point would be that in this nuanced, multi-phase system, small variations in initial conditions could disproportionately affect the specifics of average, aggregated metabolic reactions at each phase, as well as the corresponding abundances of their products; which is to day, small changes in procedure may cause significant changes in the chemical composition of the final solution. I think this is, in fact, what we observe in the various wines of, say, Beaujolais, even within a single appellation.
An illustration of this idea could be that full-cluster processing, by keeping most of the grapes' stem holes plugged, significantly slows the rate of bacterial invasion of grapes' interiors, and thus of initial 'intracellular' fermentation. in this case, perhaps more grape cells would die of starvation than of bacterial assault, prior to the kick-in of yeast-dominated fermentation. The flavor contribution of bacterial waste products to the final blend would be much smaller then for a de-stemmed batch. In addition, the migration of polyphenols from grape skin inwards would be diminished, as well as ethanol production and the related production of esters giving berry aromas (stipulated - I'm leaning very heavily on Dr. Goode's description here).
In this imagining, the exclusion of oxygen in the initial stages of fermentation in vinification could play a role analogous to that of cooling dough in sourdough bread making - i.e., of promoting non-yeast microbial activity, thereby enhancing its contribution to the final flavor package.
Anyway, blah, blah, blah. I patiently await, and preemptively accept, in full, Mark Lipton's devastating critique and correction of these conjectures.