Data centres are currently dividing opinion across the country. While many people in the UK are now living a data driven life, this comes at an environmental cost and the amount of energy needed to keep AI systems and cloud computing running can be eye watering.
Data centres currently consume around 2.5% of the UK’s electricity and this is expected to quadruple by 2030.[1] This puts additional pressure on already strained grid infrastructure, and limited capacity means it can take years to connect to the national grid.
As part of this new reality, ORE Catapult has completed a study that found a potential solution in the form of tidal stream energy (TSE). The ‘Tidal stream Suitability Assessment – Data Centre Offtake’ study found that it is technically and economically feasible to use tidal stream, combined with battery storage, to meet the energy demands of data centres, due to the predictable nature of tidal energy.
The study created a hypothetical scenario where a data centre is connected to a tidal stream energy farm at Morlais, a large tidal energy demonstration zone off the west coast of Anglesey in Wales, consented to deliver 240 MW
Using the scenario, the study found an optimal set up where 81% of the data centre’s energy requirements could be covered by a combination of tidal stream and battery storage, reducing grid imported electricity and associated CO2 emissions dramatically dropped by over 100,000 tonnes. On the battery storage side, two hours provided enough back-up power to cover slack water periods when no power was generated. Without battery storage, tidal stream generation would need to increase 10 times to reach the same drop in grid imports – highlighting that battery storage is a key enabler, improving both the economics and the environmental benefits of using tidal stream energy to power a data centre.
It was found that these figures could further improve by increasing the size of the tidal farm, but even smaller ones still performed well, with at least 60% of energy requirements being met in most of the cases considered.
For an 80 MW data centre, in this analysis, the point at which the tidal farm produces surplus energy that can be sold to the grid is 228 MW, which would also create an additional revenue stream for a project of this scale
Alternatively, the data centre could function without a grid connection if demand flexibility was incorporated into its operation – temporary caps could be implemented on demand during ‘slack water’, when tidal stream generation is at its lowest. Increasing TSE farm size and capacity would make off-grid operation more achievable, with some scenarios showing that the data centre would need to reduce its annual energy demand by less than 10%. This could significantly reduce the time and cost associated with establishing a grid connection.
It can take up to 10 years for an energy project to get a grid connection in the UK, although it is hoped that recent reforms will accelerate this timeframe. Data centre developers should investigate the cost benefit of incorporating battery storage versus incorporating demand flexibility on a case-by-case basis, considering costs and timescales for grid connection, specific capacity factors for each tidal technology and location, and power purchase agreements with tidal stream developers. Removing the grid connection would, however, prevent excess generation from the TSE farm being exported, requiring curtailment or another offtake. This creates a dilemma where increasing tidal farm size reduces grid dependence for import but is needed more for export. It would also make the data centre more vulnerable to periods of reduced output from the tidal farm, including maintenance and outages due to weather.
Another way in which off-grid operation of the data centre could be achieved is via other renewable generation sources, such as solar and wind. This would diversify the energy mix and make the data centre less vulnerable during periods of reduced energy production. Future work could evaluate how energy produced for the data centre changes as other renewable sources are added – identifying energy configurations that are less likely to require grid imports. Based on the findings in this study, tidal stream would still be expected to provide most of the data centre’s energy, even if its capacity was reduced to introduce other energy sources generation.
Building on its findings, the study recommends that further work should be done to fully understand the practicalities of co-locating tidal stream with data centres and recommends engaging with data centre operators to understand their willingness to consider tidal stream as a primary energy source.
The 240MW tidal farm considered in this study is significantly larger than any individual UK project in development that has secured a seabed lease and consent. So, the opportunities immediately on offer would be smaller in size, but they do represent an appropriate scale to trial tidal energy co-location with data centres.
The study recommends a demonstration project which would help reduce perceived technical risks, and the development of a consenting pathway for co-located projects.
Another challenge for grid connected projects is the UK Government’s Contracts for Difference budget currently available for tidal stream energy. It has been limited in previous allocation rounds, resulting in less than 50MW being awarded, and this could be a problem for grid connected projects that need to export excess generation.
If this continues to be the norm, there will not be enough budget for large projects, unless they are built in phases. However, this is a wider issue for the industrialisation of TSE, and larger budgets are expected to be introduced once further commercialisation has taken place, and more eligible capacity becomes available.
This ORE Catapult study demonstrates there is a credible pathway to developing tidal stream energy to power data centres. The predictable energy source could provide a crucial balance for electricity grids, particularly as more intermittent renewable energy sources are added to networks.
Previous UK studies have shown that tidal stream energy could lower the costs of operating the energy system by up to £1bn a year, due to reduced requirements for grid upgrades, storage systems and electricity imports. The rising need for data centres could be the perfect test case for demonstrating how this renewable energy could sustainably power our data needs in the future.
[1] https://commonslibrary.parliament.uk/research-briefings/cbp-10315/