Research collaboration areas
Specific open questions, rather than general areas of interest. Each of these is a genuine gap in CoralDC's own research programme, and each is stated as a research question because that is the form a principal investigator can actually respond to.
Thermal engineering and validation
Long-term thermal interface material behaviour in dielectric service. There is effectively no published data on how current liquid-metal and graphite interface stacks behave over years of immersion in hydrocarbon or ester fluids. This is an unanswered question for the entire industry, not only for CoralDC.
Annual-cycle performance of dry and dew-point heat rejection at elevated loop temperatures, measured rather than modelled, across varied climate zones.
Independent measurement of pod-scale efficiency, water and thermal performance to a standard that public procurement will accept.
Materials science
Heat spreading and thermal interface performance at the package level, and the relationship between package-level improvement and achievable system density, which is not linear and is poorly characterised at modular scale.
Fluid chemistry: ageing, additive depletion, water ingress behaviour and material compatibility in dielectric coolants over multi-year service.
Controls and machine learning
Learning across small heterogeneous fleets under data-sharing constraints. CoralDC's structural position is many small sites across different climates, buildings, hardware generations and workloads, with customers who will contractually restrict data pooling. Whether useful cross-site learning is possible under those constraints is an interesting research problem independent of CoralDC's commercial interest in the answer, and CoralDC does not know the answer.
Predictive thermal control from electrical and workload telemetry rather than from coolant temperature. Which signals genuinely precede thermal events.
Power systems
Flexible load behaviour of AI workloads, and whether distributed small-site flexibility is materially different from single-campus flexibility as a grid resource.
Conversion and distribution efficiency at pod scale treated as a thermal problem rather than only an electrical one.
Water and thermal recovery
Heat-driven water recovery at single-deployment thermal scale rather than campus scale. Site water balance methodology for distributed infrastructure.
Heat reuse economics
Whether heat offtake from modest, distributed deployments can be made financeable in markets without established district heating, which is an economics and policy question at least as much as an engineering one.
Start a research conversation
The most useful first message identifies a specific research question, names the funding programme or grant mechanism involved if there is one, and proposes what the collaboration would need from CoralDC specifically.