Hydro-dominant power systems are often assumed to be inherently flexible, with their reservoirs absorbing the variability of added wind and solar. We test that assumption in the Brazilian National Interconnected System (SIN) as wind and solar capacity expand rapidly and unevenly across its four subsystems. Because flexibility in such systems is usually assessed at the national level—which hides structural heterogeneity between regions—we use eight years of hourly operational data from the system operator (ONS, 2018–2025)—energy balance, plant-level generation, interregional interchange, and reservoir energy—to resolve the flexibility constraint by subsystem, against operational adequacy criteria rather than statistical descriptors. The constraint proves to be neither system-wide nor incidental but regional and structural, concentrated in the wind-heavy Northeast. Three methodologically independent dimensions converge: (i) the Northeast is the only subsystem whose local dispatchable fleet cannot follow its own net-load ramp—ramp coverage 0.51 against the adequacy threshold of 1; (ii) it leans on interregional imports to ramp, both routinely and heavily (net-load-ramp/import correlation 0.87, and 76% of its marginal ramp met by imports, far above any other subsystem); and (iii) a simplified three-dimensional indicator ranks it lowest (p < 10^-20, Mann–Whitney), its ramp-adequacy component eroding year on year as wind and solar expand, whereas the hydro-rich subsystems' scores instead rise and fall with the hydrological cycle. Grid-driven curtailment corroborates the diagnosis: the Northeast concentrates 86.8% of the national electrically driven constrained-off energy. Flexibility planning anchored to national aggregates will therefore systematically underserve the region where the integration challenge is most acute.