FULCRUM: A Preference-Conditioned Hierarchical Graph Reinforcement Learning for Sustainable Multi-Data-Center Control

Geo-distributed data-center fleets must trade off electricity costs, carbon emissions, water use, and quality of service (QoS) across sites with varying prices, grids, and weather conditions. Learned controllers for this problem are trained for a single objective weighting and a single fleet size, so any change in operator priorities or footprint requires retraining. We present , a single preference-conditioned graph policy: a Graph Pattern Machine whose objective-aligned attention heads are feature-wise linear modulation (FiLM) conditioned on the operator preference \pref, followed by cross-attention from data-center nodes to preference-scaled objective tokens. All parameters are shared across nodes and edges, so the same weights act on any fleet. Under a budget-matched protocol with paired tests over 50 evaluation episodes, repeated across three training seeds, attains the best mean rank across five operator preferences among ten controllers ($1.6\pm 0.2$), matches or exceeds every preference-conditioned baseline across all seven held-out preferences never used in training ($54$ practical wins, no losses against preference-conditioned baselines), and at one-hot preferences reaches the lowest cost, carbon, water, and QoS penalty of any baseline. It steers the fleet at inference time from $-81\%$ water to $-39\%$ cost relative to its balanced operating point and, trained once at $3$ data centers, outperforms a Green-DCC controller retrained from scratch at each of $3$, $5$, $7$, and $10$ data centers by $6.6$–$8.2\%$ reward, where preference-conditioned MLP and GNN baselines cannot be loaded at all. One training run replaces five per-preference specialists at $5.0\times$ lower compute for a $1.3$ reward cost. We also report where does not lead: at intermediate preference weights, conditioned baselines use $2$–$7\%$ less water, one of three seeds emits $3\%$ more carbon at the carbon corner, and the cross-attention and FiLM components are not individually significant at this budget.