Valuing data centres in mainland China

Data centres on the Chinese mainland are emerging as one of the leading alternative real estate asset classes in the region. This development is largely driven by demand for cloud computing, digitalisation, data growth and, increasingly, artificial intelligence (AI).

Government initiatives, such as the Eastern Data and Western Computing strategy, are supporting data centre development and investment in the sector. 

The strategy was launched in February 2022, before AI became widely used. At that time, the policy's main aim was to address the imbalance between eastern and western China.

Eastern China, where many major cities and economic centres are located, was generating large amounts of data and needed more computing capacity.

Western China, including areas such as Inner Mongolia, Guizhou, Gansu and Ningxia, had – and still has – more available land and abundant renewable energy resources that could support large-scale computing facilities.

Today, this strategy is also helping to meet the growing needs of AI. AI training can be better supported in western areas because it requires significant power and computing capacity, while AI inference, which is more sensitive to speed and latency, is better suited to locations closer to end users and major population centres.

Moreover, AI is increasing the need for high-density computing. This means more power is required in each rack, which, in turn, places greater pressure on cooling systems, power infrastructure and backup capacity.

These requirements have a direct impact on capital expenditure and on how far an existing facility can be upgraded to meet future technology needs.

The emergence and growth of demand for data centres have not gone unnoticed by institutional investors, infrastructure funds and developers.

Traditional demand hubs in or close to big urban centres, such as Beijing or Shanghai, can still command premium prices.

However, energy availability is increasingly influencing value. Investors are increasingly assessing megawatt capacity alongside the tangible, built, and real-world characteristics of data centres, i.e. buildings, infrastructure, equipment and location.

Power, connectivity, technical capability, tenant quality and operations are also key factors that developers and investors must consider. The importance of these features means that valuers are valuing a site's overall operating structure rather than just its buildings.

This trend is turning the traditional location narrative on its head, with areas that were once considered remote now becoming the location for important data centres. For example, western China benefits from renewable energy sources and therefore has abundant available power. 

As a result of this trend, the focus of valuation is shifting from a site's location to the availability of power and infrastructure capacity.

Investors are increasingly considering factors such as energy efficiency, renewable power availability, carbon emissions, and compliance with environmental standards when assessing assets.

Data centres that are built with sustainability in mind will attract greater investor interest, achieve stronger valuations, and face fewer risks as energy costs and regulatory requirements become more stringent.

As a result, future developments on the Chinese mainland are expected to place greater emphasis on efficient design, renewable energy integration and sustainable operations.

The DCF method captures the asset's future cash-generating potential, including factors such as lease income, occupancy, power capacity, operating costs, and growth assumptions.

Moreover, data centres on the Chinese mainland are specialised assets with limited comparable transactions, making market-based valuation less reliable. DCF therefore best reflects their value as long-term operational assets.

Built into the approach are key assumptions, including occupancy, rents, costs, power pricing and capital expenditure.

As data centres are increasingly viewed as operational assets, their value depends less on location alone and more on how effectively they operate and create cash flow.

A DCF approach reflects this by focusing on factors such as power capacity, efficiency, revenue and costs, making it better suited to capture the drivers of data centre value.

What is also apparent is that underwriting is becoming more disciplined. There is a greater focus on operational performance than on future potential alone.

This makes underwriting more disciplined by prioritising proven results over future assumptions, reducing reliance on speculation and leading to more accurate valuations and risk assessments.

Alongside this, occupancy, cash flow and asset quality are coming under closer examination, with market discussions increasingly supporting the need for infrastructure-style valuation rather than looking at the asset's traditional aspects.

Going forward, data centre valuation on the Chinese mainland is expected to become even more sophisticated. Traditional real estate fundamentals will remain relevant, but operational performance, power availability, technical capability and sustainability will continue to gain in importance.

In the future, while location will remain a key driver of data centre value, operational efficiency will become an increasingly critical factor.

Building capacity alone, therefore, is no longer sufficient when it comes to data centre valuation on the Chinese mainland. Investors are becoming more selective regarding what they are looking for in an investment.

Operational excellence is now seen as a competitive value advantage, and this is not lost on developers who are building efficient, resilient and sustainable data centre facilities.

It is therefore imperative for the data centre market to adopt a more explicit valuation framework beyond traditional property metrics.