The paper introduces FissionReady, a scheduler for data centers supplied by a multi-module small modular reactor (SMR) plant. It jointly manages the output of each reactor module, deferrable batch workloads, and grid purchases while accounting for fuel age, iodine-xenon dynamics, and reactivity margin. In simulations, the approach eliminates forced outages and missed batch-job deadlines while reducing water use and grid-electricity costs relative to a comparable baseline configuration.
Key Takeaways
SMR load following cannot be planned around aggregate plant output alone: each module’s feasible operating trajectory depends on fuel age, load history, and the anticipated xenon peak.
Joint optimization of generation, deferrable compute workloads, and grid purchases can avoid both surplus nuclear generation and reactive grid procurement during forced outages.
In the reported experiments, FissionReady delivers zero forced module outages and zero missed batch-job deadlines.
Compared with the equivalent Baseline-2 configuration, the system reduces water consumption by 31%; reported water intensity is 346 versus 475 L/MWh.
Fixed load following that does not account for the condition of individual reactors results in 11 forced outages and 212 hours of cumulative downtime in the scenario studied.
Key Figures
Water consumption reduction
31 %
FissionReady relative to the comparable Baseline-2 configuration without load following.
FissionReady water intensity
346 L/MWh
Compared with the equivalent configuration in the water-intensity analysis.
Comparable configuration water intensity
475 L/MWh
Baseline configuration used in the water-intensity comparison.
Reduction in SMR energy losses
29.8 до 1.7 %
Reported FissionReady result.
Reduction in grid-electricity costs
примерно в 2 раза
Compared with a baseline configuration of the same size, according to the abstract.
Average data-center demand
7 MW
Scenario based on the Alibaba cluster trace.
Peak proposed demand
16.1 MW
Scenario based on the Alibaba cluster trace.
Baseline-2 capacity
10.2 MW
SMR fleet sized to P90 demand; the primary comparable baseline.
Number of servers in the modeled cluster
30000 серверов
Cluster based on Xeon Platinum 8175M servers.
Scheduler decision granularity
1 минута
Interval used for job-scheduling decisions.
Maximum batch-job wait before deadline violation
12 часов
Permitted waiting time for batch workloads.
Fixed Load Following forced outages
11 остановок
For the F-LF scheme; FissionReady reports zero.
Fixed Load Following cumulative downtime
212 часов
For the F-LF scheme as a result of forced outages.
Fixed Load Following relative grid cost
30 раз
F-LF relative to FissionReady, due to unplanned purchases during outages.
Recovery time after a xenon transient
20 и более часов
At the end of the fuel cycle following a forced module shutdown.
Average optimizer computational overhead
35 мс
Six SMR modules and a 48-hour horizon; less than 0.006% of a 10-minute scheduling interval.
Default safety margin
300 pcm
Used to limit risk from uncertainty in the reactor model.
Water savings with forecast noise
28.8 %
Relative to Baseline-2 under the 50% noise scenario.
Average batch-job waiting time with forecast noise
1.31 часа
Noisy 50% scenario.
Practical Value for Data Center Owners
For data-center owners and project teams, the study establishes an operational architecture requirement for an SMR-powered site: the nuclear plant cannot be treated as a single aggregated dispatchable resource. It requires module-level telemetry on fuel and iodine-xenon conditions, an interface linking generation dispatch, batch-workload shifting, and grid procurement, as well as a backup grid interconnection. The approach is useful as a framework for evaluating co-located SMR-plus-data-center concepts and for defining energy-management-system requirements; its results are model-based rather than evidence of industrial operation.
Wärtsilä’s presentation examines the use of modular gas and dual-fuel power plants alongside energy storage to supply data centers, including AI workloads. It links constraints on grid connections to growing demand for primary onsite and off-grid generation, and cites several projects in Ireland and the United States. The material emphasizes capacity availability, modularity, part-load efficiency, lower water consumption, and the potential to transition to sustainable fuels over time.
Data center power supplyAI workloadsOnsite generationOff-grid data centers
This AVK and Wärtsilä Energy white paper examines constrained grid capacity as a major barrier to the expansion of European data centres, particularly facilities serving AI workloads. It proposes on-site microgrids combining dispatchable generation, BESS and renewables to accelerate site energisation before—or independently of—a grid connection. The paper compares power-system scenarios and four supply configurations for a representative AI data centre in terms of cost, reliability, land use and carbon footprint.
Data Centre Power SupplyAI Data CentresMicrogridsDispatchable Generation
This Cummins white paper examines the shift in data center onsite power systems from traditional emergency backup toward an active-standby model. It identifies decarbonization and shortages of grid capacity and transmission capability—particularly acute in Europe and North America—as the key drivers. The paper considers cleaner fuels and generation technologies, energy storage, fuel cells, and microgrid controls that could enable data centers to support the grid without moving to continuous onsite generation.
Data Center Power SupplyOnsite GenerationActive StandbyGrid Constraints