Nanocelluloses are seen as the basis of high-performance materials from renewable sources, enabling a bio-based sustainable future. Unsurprisingly, research has initially been focused on the design of new material concepts and less on new and adapted fabrication processes that would allow large-scale industrial production and widespread societal impact. In fact, even the processing routes for making nanocelluloses and the understanding on how the mechanical action fibrillates plant raw materials, albeit chemically or enzymatically pre-treated, are only rudimentary and have not evolved significantly during the past three decades. To address the challenge of designing cellulose comminution processes for a reliable and predictable production of nanocelluloses, we engineered a study setup, referred to as Hyper Inertia Microfluidizer, to observe and quantify phenomena at high speeds and acceleration into microchannels, which is the underlying flow in homogenization. We study two different channel geometries, one with acceleration into a straight channel and one with acceleration into a 90°bend, which resembles the commercial equipment for microfluidization. With the purpose of intensification of the nanocellulose production process, we focused on an efficient first pass fragmentation. Fibers are strained by the extensional flow upon acceleration into the microchannels, leading to buckling deformation and, at a higher velocity, fragmentation. The treatment induces sites of structural damage along and at the end of the fiber, which become a source for nanocellulose. Irrespectively on the treatment channel, these nanocelluloses are fibril-agglomerates, which are further reduced to smaller sizes. In a theoretical analysis, we identify fibril delamination as failure mode from bending by turbulent fluctuations in the flow as a comminution mechanism at the nanocellulose scale. Thus, we argue that intensification of the fibrillation can be achieved by an initial efficient fragmentation of the cellulose in smaller fragments, leading to a larger number of damaged sites for the nanocellulose production. Refinement of these nanocelluloses to fibrils is then achieved by an increase in critical bending events, i.e., decreasing the turbulent length scale and increasing the residence time of fibrils in the turbulent flow.