Singapore Data Center REITs Bet on Japan: Is Power Scarcity Creating a New Growth Trade?

Singapore-listed data center REITs are increasing their exposure to Japan as AI and cloud demand collide with a shortage of power-ready facilities.

Keppel DC REIT recently proposed buying two Tokyo data centers, while Digital Core REIT increased its stake in an Osaka facility.

The opportunity looks attractive.

But the same power shortage supporting asset values could also restrict future growth.

That creates an important question:

Is Japan becoming a new data-center growth trade—or an increasingly expensive one?

Educational research only. This article is not investment advice.

Why Singapore REITs Are Buying in Japan

Japan is one of Asia-Pacific’s largest and most mature data-center markets.

It currently has more than 1.8 gigawatts of operational capacity and a vacancy rate of about 6.6%.

Demand is being supported by:

  • cloud adoption;
  • AI workloads;
  • domestic enterprise computing;
  • hyperscale expansion.

Hyperscalers already account for roughly 60%–70% of total capacity.

That combination makes operational Japanese data centers attractive to Singapore REITs seeking growth.

Keppel DC REIT Makes a Big Tokyo Bet

Keppel DC REIT has agreed to acquire major interests in two fully fitted data centers in Greater Tokyo.

The transaction values the assets at roughly US$1.1 billion.

After completion, Japan’s contribution to Keppel DC REIT’s rental income is expected to rise from about:

9% → 23%.

Both facilities are fully occupied.

Keppel DC REIT also raised S$625 million through a private placement to help fund the acquisition.

The placement was around 3.4 times covered, suggesting strong investor demand.

Digital Core REIT Is Expanding Too

Digital Core REIT has also increased its Japan exposure.

It agreed to acquire another 25% stake in an Osaka data center, raising its ownership from:

20% → 45%.

The Osaka asset is also fully occupied.

This tells us something important.

Singapore REITs are not simply betting on future construction.

They are targeting operational facilities that already have tenants and power access.

That reduces some development risk.

Why Power Is Becoming So Valuable

A data center is only useful if it can access enough electricity.

That is becoming increasingly difficult.

AI servers require enormous power for:

  • GPUs;
  • cooling;
  • networking;
  • backup systems.

In Japan, access to high-voltage electricity is becoming one of the most valuable features of a data-center asset.

This is especially true in Greater Tokyo, where available grid capacity is constrained.

Osaka currently has somewhat better power availability.

The result is simple:

Power-ready assets become scarcer → competition increases → valuations can rise.

Scarcity Can Help—and Hurt

Power scarcity creates two opposing effects.

Positive for Existing Assets

Operational data centers with secured electricity become more valuable.

They can benefit from:

  • high occupancy;
  • stronger rental demand;
  • limited competing supply.

Negative for Future Growth

The same electricity shortage can make expansion difficult.

New projects may face:

  • long grid-connection delays;
  • higher development costs;
  • limited available sites;
  • expensive power infrastructure.

So scarcity can support existing assets while making future acquisitions harder.

Japan Still Has More Supply Coming

Japan has roughly 3.5 GW of additional data-center capacity planned or under development through 2030.

But much of that future supply is already spoken for.

Around:

40% of capacity under construction

and:

26% of planned capacity

is already pre-leased.

That suggests demand remains strong.

Japan also expects electricity demand to rise about 5.3% over the next decade, partly because of data centers.

The government is planning to strengthen the grid and build or upgrade around 30 substations.

How the TSL Macro Model Fits

TradingSimuLab’s Macro Model helps frame the broader environment.

Important questions include:

Net Score
Is AI infrastructure spending supporting economic growth?

Confidence
Are cloud demand, capital investment and power demand moving together?

Scenario Probabilities
Could strong infrastructure investment continue—or do higher rates and electricity constraints become a drag?

We are not assigning a live TSL Macro score here.

The purpose is to connect the REIT story to the wider economic environment.

Why Risk Simulation Matters

A high-quality data center can still be a poor investment if too much is paid for it.

TradingSimuLab’s Risk Simulation would focus on:

Expected Return
Does future rental growth justify the acquisition price?

VaR and CVaR
How severe could downside become if valuations fall?

Max Drawdown
How much could unit prices decline during a REIT selloff?

Probability of Gain
How often do simulated outcomes remain positive?

The central lesson is:

Strong demand does not eliminate valuation or financing risk.

What Could Strengthen the Theme?

Watch for:

  • rising AI workloads;
  • higher cloud demand;
  • strong occupancy;
  • positive rental reversions;
  • continued pre-leasing;
  • successful grid expansion.

What Could Weaken It?

Risks include:

  • higher interest rates;
  • expensive acquisitions;
  • unit-holder dilution;
  • grid delays;
  • slower AI spending;
  • rapid technological obsolescence.

Data centers can become outdated faster than traditional property, making capital expenditure an ongoing concern.

Final Takeaway

Singapore data-center REITs are increasingly using Japan as a major growth market.

The opportunity is driven by:

AI demand → cloud growth → data-center demand → electricity scarcity → higher value for power-ready assets.

But scarcity cuts both ways.

It can make existing assets more valuable while making new capacity harder and more expensive to build.

The better question is not:

“Is Japan’s data-center market growing?”

It clearly is.

The question is:

“Can Singapore REITs acquire power-ready assets at prices that still generate attractive returns?”

For more Singapore market research, macro analysis and risk simulations, sign up to TradingSimuLab and explore the platform.

Continue exploring TradingSimuLab.

  • Monte Carlo Simulation for Stocks: How Thousands of Price Paths Help Measure Risk

    Educational research only — not investment advice. A Monte Carlo stock simulation does not try to predict one exact future price. Instead, it creates hundreds or thousands of possible price paths. The goal is simple: Rather than asking “Where will this stock be?” ask “What range of outcomes is possible?” That makes Monte Carlo simulation…

  • CVaR Explained: How to Measure the Losses That Happen Beyond VaR

    Educational research only — not investment advice. CVaR explained simply means measuring the average loss when things go worse than your Value at Risk threshold. CVaR is also called Conditional Value at Risk or Expected Shortfall. It answers a question that VaR cannot: If a bad outcome happens, how bad could the average loss be?…

  • Value at Risk Explained Simply: What VaR Can—and Cannot—Tell Investors

    Educational research only — not investment advice. Value at Risk explained simply means estimating how much an investment could lose over a specific period under normal market conditions. VaR tries to answer: How much could I lose before the outcome becomes unusually bad? It is useful—but only if you understand its limits. What Is Value…

  • What Is Maximum Drawdown? How to Measure the Real Risk of an Investment

    Educational research only — not investment advice. Maximum drawdown measures the largest decline an investment experiences from a previous peak to a later low. It answers a very practical question: How bad did the investment get before recovering? That makes drawdown one of the most useful ways to understand investment risk. What Is Maximum Drawdown?…

  • Expected Return vs Risk-Reward: Why They Are Not the SameThing

    Educational research only — not investment advice. Expected return vs risk reward sounds like the same idea. It is not. Both help investors evaluate an opportunity, but they answer different questions. Expected return asks:What is the average outcome after considering different probabilities? Risk-reward asks:How much could I gain compared with how much I could lose?…

  • Probability of Profit Explained: What Does a 60% Chance of Gain Really Mean?

    Educational research only — not investment advice. A probability of profit tells you how often an investment or trade is expected to finish with a gain under a set of assumptions. If a model shows a 60% probability of profit, it means: about 60 out of 100 simulated outcomes finish above the starting point. It…

  • How to Measure Whether a Stock Trend Is Getting Stronger or Weaker

    Educational research only — not investment advice. A stock can be in an uptrend and still be losing strength. That is why a trend strength indicator can be more useful than simply asking whether price is going up or down. The real question is: Is the trend becoming more persistent—or starting to weaken? Start With…

  • Market Timing Explained: Why a Good Stock Can Still Be aBad Entry

    Educational research only — not investment advice. Market timing is often misunderstood. It does not simply mean trying to predict the exact top or bottom of the market. A more useful idea is: A good company can still be a bad trade if you enter at the wrong time. That is because stock quality and…

  • How to Tell if a Stock Is Trending or Just Moving Sideways

    Educational research only — not investment advice. Good stock trend analysis starts with one simple question: Is the price actually trending—or is it just moving around inside a range? The difference matters. A strategy that works well during a strong trend can perform poorly when a stock is moving sideways. That is why identifying the…