How AI will revolutionise energy storage investment

• Storage is a nascent technology that has traditionally spooked lenders due to it having an insubstantial track record
• Such lender uncertainty means debt sizing is limited to 20-30 per cent in some markets
• However, new report shows AI can be used to better predict returns on storage projects
• Few banks currently use AI for forecasting and business modelling, but it’s expected nearly half will do so within three years
• Greater use of AI to forecast returns on energy storage investment is set to make financing much more accessible to the sector

Convincing lenders to finance energy storage projects has traditionally been challenging for a number of reasons. These barriers include lithium-ion battery capacity fade, which reduces lifespan and efficiency, meaning more capital investment is needed to maintain and augment batteries in order to maintain peak performance. Another issue that can spook lenders is the unpredictability of merchant revenue streams – revenue sources such as arbitrage, ancillary services and capacity markets make it troublesome to predict profitability in the longer-term.

Securing debt financing has also been challenging for the energy storage sector, largely because storage is a nascent technology. When providing finance, lenders want to see a long history of performance data, which is precisely what storage lacks. Consequently, storage is seen as being more risky than a lot of other investments. For instance, in the ERCOT market in the US, debt sizing for standalone storage is reportedly often limited to 20-30 per cent of total project cost, which gives a clear indication of the extent of lender uncertainty regarding the returns achievable when seeking to tap merchant revenue streams. This lack of clarity means that energy storage projects may secure less favourable loan terms in comparison to more mature technologies.

How lenders size debt

Two of the key debt sizing techniques used by lenders when providing finance for energy storage projects are:

• Debt Service Coverage Ratio (DSCR), which is a key metric used to assess a project’s ability to meet its debt obligations. Calculating the DSCR involves dividing the project’s operating cash flow by its debt service payments. A higher DSCR is perceived as a healthier ability to cover debt, and therefore the potential for securing a higher debt sizing increases.
• Loan Life Coverage Ratio (LLCR), which is similar to DSCR, but LLCR is concerned with the project’s ability to cover debt service payments over its entire loan term. It considers the net present value of future cash flows, incorporating potential fluctuations and uncertainties.

However, as is clear, calculating such ratios involves factoring in a high degree of uncertainty such as the volatility of merchant revenue streams and battery degradation, for example. Yet the wider use of artificial intelligence in the financial industry is set to eliminate a significant amount of risk for lenders when investing in energy storage.

AI can be used to better predict returns on storage investment

A new report from the London School of Economics (LSE), which focuses on the role of AI in the climate transition, concludes that AI can be used to “better predict investment risks and returns, improving financial decisions where information is scarcer, particularly in emerging markets where perceived risk is high, often due to limited and asymmetric information”. How does AI do this? AI aggregates diverse data sources on “realised project risk”, providing more accurate risk assessment and prediction and, consequently, making financing more accessible, the LSE report says. The report adds that AI is being used to optimise renewables investment portfolios by helping investors to decide the best locations for renewables projects. For example, lenders can use satellite data to estimate solar power generation potential in different regions.

In many respects, lenders are only scratching at the surface when it comes to making the best use of AI to reduce the risk associated with energy storage investments. For example, if we take the UK, a survey by the Bank of England showed that 41 per cent of UK financial services firms were using AI for the optimisation of internal processes, but only approximately 20 per cent are using AI for forecasting and business modelling, for example, while only around 5 per cent are using AI for risk modelling.

But it is only a matter of time before the majority of banks are using AI to make more accurate forecasts on the return on investment in energy storage projects. Indeed, within three years, it is estimated that nearly half of UK financial institutions will be using AI for forecasting and business modelling. The expectation is that the wider use of AI to assess investment risk will make financing more accessible for energy storage developers as it will enable lenders to get a more accurate picture of the potential returns on investment for storage projects. It’s a scenario that will do much to fuel the wider deployment of energy storage around the world.