How should electricity pricing be designed to support data centers? Lessons and approaches from the US to Thailand

• The rapid expansion of data centers presents a significant opportunity for Thailand's digital economy and AI development. However, their massive and continuous electricity demand will inevitably increase cost burdens on the electrical grid. The draft Power Development Plan 2026 (PDP2026) estimates that electricity demand from data centers could reach approximately 42,500-55,700 GWh per year by 2050. This could necessitate further investment in generation capacity, substations, transmission lines, and grid reliability reserves. Consequently, it poses cost-allocation risks, particularly if the state invests in infrastructure ahead of demand, only for projects to be delayed, downsized, or canceled, leaving the built assets underutilized and shifting the cost burden onto other electricity consumers.
• International precedents, such as those in the US and Singapore, reflect a shared principle: the growth of data centers should be accompanied by responsibility for the costs and resources they consume. In the US, the Ratepayer Protection Act proposes that large-scale data centers bear the incremental costs associated with generation capacity, while also requiring financial assurance mechanisms to reduce the risk of these costs being shifted to existing electricity consumers. Singapore, meanwhile, focuses on selectively allocating electricity capacity to data centers that demonstrate high energy efficiency, low carbon emissions, and strong economic value creation.
• Thailand has already begun moving in a similar direction. On July 15, 2026, the National Energy Policy Council (NEPC) approved, in principle, the establishment of a dedicated electricity tariff for data centers to reflect the incremental costs of electricity procurement and grid expansion. The resolution also requires large-scale data centers to provide financial assurance before the state invests in supporting infrastructure. In addition, on October 1, 2026, the Data Center Board approved the classification of data centers by electricity consumption level in preparation for regulatory oversight.
• Looking ahead, Thailand should establish greater clarity in three key areas: (1) distinguishing costs attributable specifically to data centers from system-wide costs that deliver shared public benefits; (2) defining the obligations and liabilities of data center operators if actual electricity consumption falls short of projected demand; and (3) incentivizing projects that maximize energy efficiency while aligning with local grid capacity. If designed effectively, Thailand could attract data center investments while preserving competitive electricity tariffs and shielding ordinary citizens and local businesses from the financial burdens associated with the growth of the digital industry.

 
How Thailand can attract data centers without raising electricity costs for the public
Investment in data centers presents a significant opportunity for Thailand to support the growth of its digital economy and AI ecosystem. However, these facilities consume large amounts of electricity on a continuous basis, raising an important question: Who should bear the cost of the additional grid infrastructure required to support such investments?
If these costs are not clearly allocated, investments intended to drive the broader economy could end up forcing citizens and local businesses to shoulder higher electricity bills, even if they receive no direct benefit from the projects.

Where Thailand faces risks: Investing in advance when projects may consume less electricity than projected
According to figures in Thailand's draft PDP2026, electricity demand from data centers in Thailand could reach approximately 42,500-55,700 GWh per year by 2050. This indicates that the country must expand its electrical grid to accommodate this growing demand.
The risk cuts both ways. On the one hand, if the existing grid lacks sufficient capacity, electric utilities may need to construct substations, expand transmission lines, or procure additional generation. If these capital costs are incorporated into electricity tariffs charged to all users, ordinary ratepayers could end up shouldering the burden.
On the other hand, if investments are made but a data center is delayed, downsized, or canceled altogether, the built infrastructure risks being underutilized while the fixed capital costs remain. For instance, if a new substation is built to serve a specific data center that later withdraws from the project, it must be clearly established from the outset who bears the costs already incurred, preventing the burden from spilling over onto other consumers' electricity bills.

The US proposes that cost causers bear the costs
The US House of Representatives has passed the Ratepayer Protection Act, which proposes that state regulatory bodies consider standards for data centers with an aggregate peak electricity demand of 100 MW or more within the same area. A core principle of the bill is that the expenses associated with expanding generation capacity and upgrading grid infrastructure to support these facilities should be billed directly to the projects themselves, alongside requirements to provide financial assurance or contribute capital before infrastructure investments begin.
This concept addresses two critical questions: 'who pays for the additional grid infrastructure' and 'who bears the risk if a project fails to materialize as planned.' However, the proposed legislation must still pass through the Senate, and its provisions merely direct state-level authorities to consider such standards. Therefore, it should not yet be interpreted as a generally applicable rule across the US.

Singapore focuses on pairing data center growth with energy efficiency
Facing land, energy, and water constraints, Singapore prioritizes selecting data centers that deliver substantial economic benefits alongside high resource efficiency. Under its Green Data Centre Roadmap, Singapore allocates capacity for new projects to operators that demonstrate a strong commitment to sustainability and economic value creation. To qualify, operators must demonstrate their capabilities in energy efficiency, decarbonization, as well as the economic benefits their projects will bring to the country.
While Singapore's measures focus primarily on guiding the growth, efficiency, and energy sources of data centers, they do not necessarily involve fully separating data center-related electricity costs from those borne by the public. Nevertheless, an applicable lesson is that attracting investment should consider not only the economic benefits of a project but also its electricity requirements and the readiness of the local grid. Data center operators could be required to commit to their projected electricity consumption plans and report actual usage over time. Such measures may help reduce the need for additional grid investments and curb overcapacity risks. Thailand could complement its cost-allocation and financial assurance mechanisms with this approach to effectively manage surging electricity demand while safeguarding a fair and equitable cost distribution.

Thailand has established the principles; the priority now is implementation
            Following the resolution of the NEPC on July 15, 2026, Thailand approved the introduction of a separate electricity tariff for data centers to reflect incremental electricity procurement and grid expansion costs. These costs may include importing additional liquefied natural gas (LNG) to generate more electricity, the construction of substations and transmission lines, and upgrades to ensure a reliable and uninterrupted electricity supply. The initiative aims to recover data center-related costs directly from operators rather than spreading the burden across ordinary citizens and businesses. Furthermore, it requires large-scale data center developers to provide financial assurance before the public sector undertakes grid expansion investments, thereby validating project readiness and mitigating the risks associated with premature capital expenditures.
In addition, following the Data Center Board's decision on October 1, 2026, data centers were classified into five categories based on electricity consumption, ranging from those using no more than 1 MW to those consuming more than 200 MW. The proposed framework is expected to introduce two main electricity tariff rates to data center users. Data centers powered by clean energy would be charged a lower tariff than those reliant on natural gas generation while the Net Zero target would also be incorporated into the regulatory framework governing the sector. However, it is not yet possible to determine how much more data centers will pay, when compared with other electricity users, or how much lower the clean-energy tariff will be on a per-unit basis, as these details remain subject to further consideration by the subcommittee.
In this regard, Thailand's approach aligns in principle with the US Ratepayer Protection Act, which aims to ensure that large-scale electricity consumers bear the specific costs and risks associated with their own projects. For Thailand, establishing transparent, upfront terms will both protect ordinary ratepayers and provide developers with cost certainty. Crucially, the framework should permit phased load ramp-ups aligned with agreed schedules. This would ensure that utility capital expenditures directly match commercial operations, simultaneously shielding general consumers and preserving Thailand's investment attractiveness.

How much more will data centers pay once a separate electricity tariff is introduced?
A simulation of electricity costs for a 25-MW [1] data center in Virginia, United States, found that the average electricity rate increased from approximately 6.5 to 8.0 US cents per unit – an increase of roughly 23%. This was primarily driven by higher costs associated with the transmission and distribution systems.
Additionally, Virginia requires qualifying large-scale electricity users to pay for at least 85% of their allocated transmission and distribution system costs, even if their actual electricity consumption falls below that level. This mechanism is designed to prevent the costs of infrastructure built to serve these users from being shifted to other electricity consumers.
For Thailand, these examples suggest that the design of electricity tariffs for data centers should consider not only the per-unit electricity rate but also the responsibility for the infrastructure built to support such projects. Several scenarios should be evaluated, including cases in which a ​project operates at its planned electricity demand, experiences delays in commencing operations, or consumes less electricity than originally projected. This would help clarify how much of the incremental costs data center operators would be required to bear and the extent to which such measures can protect general electricity consumers from shouldering these costs.

As a next step, Thailand may adapt approaches from the US or Singapore to build upon its initiated measures across three key areas
        Thailand has begun laying the groundwork for data center growth through dedicated electricity tariffs, financial assurance requirements, and project eligibility assessments. Moving forward, the focus should be on detailing and integrating these mechanisms so that investors can forecast costs with certainty while insulating ordinary ratepayers from adverse impacts. Three key areas may warrant further consideration:

1. Ensure transparent and auditable cost allocation by clearly distinguishing between expenditures incurred specifically to serve data centers and system-wide upgrades that deliver shared public benefits, thereby ensuring a fair distribution of costs.
2. Establish upfront accountability before capital deployment by requiring developers to commit to project capacity and energization schedules, which are backed by appropriate financial assurance and clear liability terms in the event of downsizing or project cancellation, thereby building upon the existing financial assurance framework.
3. Promote energy-efficient projects that are well-suited to local conditions by incorporating lessons from Singapore and building on Thailand's Net Zero framework. This would help ensure that data center investment expands in line with the readiness of the electrical grid while supporting the country's sustainability objectives.

 
[1] Energy and Environmental Economics, Inc. (2026, May). Large load tariff whitepaper.

Scan QR Code