Your Company Is Paying an ‘AI Tax’ Despite Not Using AI at All

The electricity consumption of AI data centers is expected to balloon to 2 to 2.5 times its current level by 2030. Where does this electricity come from? From your local power grid.

In other words, even small and medium-sized enterprises (SMEs) that do not adopt AI are being burdened with an ‘AI tax’ in the form of rising local electricity costs. This is not just a technological issue; it is a matter of business costs.

The Reality of ‘AI Power Consumption’ in Numbers

According to forecasts from the International Energy Agency (IEA) and Goldman Sachs, the electricity consumption of AI data centers is projected to surge from 118 TWh (terawatt-hours) in 2025 to between 239 and 295 TWh by 2030. This is equivalent to about one-third of Japan’s annual total power generation.

The problem lies in the ‘concentration’ of these data centers. They are built in specific regions where power and communication infrastructure are well-established. In the United States, notable locations include Northern Virginia, Oregon, and Texas. In Europe, Ireland and the Netherlands are prominent, while in Asia, Singapore and the Inzai area in Chiba, Japan, stand out.

In these regions, cases have been reported where the Power Stress Index exceeds 0.25. This indicates that the demand from data centers is excessive compared to the local power supply capacity.

As a result, what happens? Electricity prices for the entire region go up.

A Direct Hit to SMEs — Let’s Do the Math

Let’s consider some specific numbers.

Assuming a small manufacturing or food service company has a monthly electricity bill of 300,000 yen, if local electricity prices rise by 10%, that translates to an additional annual cost of 360,000 yen. If the increase is 15%, it becomes 540,000 yen, and at 20%, it reaches 720,000 yen.

For a company with an annual revenue of 50 million yen and an operating profit margin of 5%, the operating profit would be 2.5 million yen. Losing 720,000 yen from that means that about 30% of the profit is wiped out as an AI tax.

Moreover, this cost arises even if the company does not use AI at all. Just having a data center built next door can drag them into the competition for electricity.

Why Is It ‘Invisible’?

The electricity bill does not explicitly mention an ‘AI tax.’ Instead, it quietly blends into the basic and variable charges, just like ‘fuel cost adjustments’ or ‘renewable energy surcharges.’

What makes this issue even more complicated is that it is not often discussed in the context of ‘promoting AI.’ Both the government and local authorities are pushing for the attraction of data centers as an economic revitalization strategy. Jobs are created, and tax revenues increase — that’s the public narrative. However, the number of jobs created by data centers is surprisingly low. It is not uncommon for a few dozen full-time jobs to be created from an investment of several billion yen. Meanwhile, the rise in electricity costs affects all businesses in the region.

The beneficiaries and the burden bearers are misaligned. This is the essence of the ‘invisible AI tax.’

What Is Actually Happening — Examples from Abroad

In Georgia, USA, the surge in data centers has led utility companies to apply for increases in electricity rates for households and SMEs. Similar movements are occurring in Oregon. In Ireland, a moratorium on new data center constructions is being considered.

Japan is also showing signs of this trend. The area around Inzai City in Chiba Prefecture has seen a concentration of data centers, leading to a rapid increase in local power demand. Industrial electricity rates in the Tokyo Electric Power Company area have risen by about 20% over the past three years, with data center demand cited as a contributing factor.

What Can SMEs Do?

You might think, “It’s a structural problem, so there’s nothing we can do about it.” However, there are ways to address it.

1. Immediately Visualize Electricity Costs

First, track the monthly trends of your electricity costs. Line up the electricity bills from the past three years and confirm the upward trend. This alone can transform the vague feeling of rising costs into concrete data showing an annual increase of XX yen, which can inform business decisions.

2. Review Electricity Contracts

Since the liberalization of the electricity market, SMEs can choose their electricity providers. Many cases allow for cost reductions of 5-15% by switching to new electricity suppliers. In regions where data centers are concentrated, local utility companies may have higher rates. Getting estimates from comparison sites can be done in just 30 minutes.

3. Consider Self-Generation and Storage

The costs of installing solar panels and storage batteries have been decreasing year by year. A 10 kW solar power system can be installed for around 2 to 3 million yen, with expected annual electricity savings of 300,000 to 500,000 yen. The payback period is 5 to 7 years. The higher electricity prices rise, the shorter the payback period becomes. Ironically, as the AI tax becomes heavier, the investment efficiency for self-generation improves.

4. Raise Your Voice in Local Policy Decisions

Decisions regarding the attraction of data centers are often made at the municipal level. It is the legitimate right of SMEs to request an ‘impact assessment of rising electricity costs due to data center construction’ through local chambers of commerce and business associations. If you don’t speak up, the burdens will only come down on you.

Understanding the Structure and Shifting to the ‘Using Side’ of AI

Here’s a paradoxical statement.

If you are merely on the receiving end of the AI tax, costs will only increase. Therefore, there is also the idea of using AI to increase sales and operational efficiency to absorb the rise in electricity costs.

By automating customer support with AI chatbots, you can reduce labor costs by 100,000 to 300,000 yen per month. By using AI for demand forecasting to reduce inventory loss, annual improvements of several hundred thousand yen can be expected. You can recoup the increase in electricity costs through operational efficiency gained from AI. It’s essential to think in both defensive and offensive terms.

Ending the ‘Paying Without Knowing’

The power consumption of AI data centers will accelerate further. As we approach 2030, this issue will only grow larger, not smaller.

The first thing that SME owners should do is understand the trends in their electricity costs. Then, they need to determine whether the increase is due to the ‘invisible AI tax.’

Being forced to pay without knowing is the worst scenario. Once you understand the structure, solutions will become apparent.