Renewable Power Procurement for Thai Plants — UGT and I-REC

Most plant teams that start looking into renewable power procurement in Thailand get stuck at the same point. Switch the contract and the electricity turns green. But head office and customers do not ask you to prove that you bought green power. They ask you to prove that you used it. In 2026 Thailand finally has the schemes to support that claim. This article walks through the three procurement routes now open to factories, and through the metering data that turns a contract into evidence.

Thai factories have more renewable procurement options in 2026

The regulatory picture first

Thailand’s green electricity framework traces back to the resolution of the National Energy Policy Council meeting 7/2565 on 7 November 2022, after which the Energy Regulatory Commission built out the rules. According to the ERC’s own page, the framework has two tiers. UGT1 rates have been published for the 2568 and 2569 service years, which correspond to 2025 and 2026. UGT2 was approved by ERC resolution 13/2569 on 1 April 2026.

Add standalone I-REC purchases and rooftop solar for self-consumption, and a factory realistically has three families of options. They are not mutually exclusive. In practice you combine them and use one to fill the gap left by another.

For the cost-reduction side of factory electricity, we covered the ground separately in why measuring power alone does not lower your electricity bill. This article picks up where that one ends, on the question of explaining where your electricity came from.

UGT1 is the 1-year entry point with no choice of source

UGT1 lets you buy renewable electricity without specifying the generating plant. According to the official UGT site operated by EGAT, the supply comes from 7 EGAT hydropower plants, offers roughly 2,000 GWh per year, and corresponds to about 2 million RECs annually. Contracts run for 1 year, and any class of electricity user can apply.

Pricing works as a premium on top of the normal tariff. For the 2026 service year, the ERC ran a public consultation from 7 to 19 January 2026, and MEA, the Metropolitan Electricity Authority, published a notice dated 12 March 2026 setting the premium at 0.0375 baht per unit and opening registration from 16 March to 10 April 2026. The same notice confirms that the supplied electricity carries renewable energy certificates under the I-REC Standard.

Thai energy media reported that this 0.0375 baht premium represents a 37 percent reduction from the previous year. For a plant buying 12 million kWh a year, the premium adds 450,000 baht. As an entry point, that is remarkably cheap.

Take-up, however, has been thin. The Thai business daily Thansettakij reported on 19 April 2026 that although the 7 hydropower plants can supply 1,329 million units a year, uptake since the middle of 2025 has amounted to 41 corporate users and roughly 142 million units. The capacity exists. It is simply not being used.

UGT2 lets you name the source, on a 10-year contract

UGT2 lets you specify the generating plants. The official UGT site lists 245 solar, wind and solar-plus-battery facilities, with a combined annual supply of roughly 12,000 GWh across portfolios A and B once all of them are in commercial operation. Contracts run for 10 years, and eligibility is restricted to Thai tariff classes Type 4, large general service, and Type 5, specific business service.

On pricing, Thansettakij reported on 19 April 2026 that UGT2 uses a fixed-rate structure with a management fee of 5 satang, or 0.05 baht, per unit, and that reported contract terms range from 10 to 25 years, which makes the 10 years stated on the official UGT site the lower bound. Several outlets reported a unit price of around 4.56 baht for portfolio A, which is made up of plants reaching commercial operation between 2568 and 2570.

The practical significance of UGT2 is not the price level but the contract length. 10 years means 10 years of a fixed unit rate insulated from the fuel adjustment charge embedded in the ordinary tariff. Evaluate it purely as a cost uplift and you will reach the wrong conclusion.

Buying I-RECs on their own

You can also buy certificates separately. In Thailand, EGAT is the issuing body, accredited by the International Tracking Standard Foundation as the local issuer for certificates under the I-REC(E) code. Eligible generation includes solar, wind, hydro, biomass, biogas and geothermal.

The decisive difference between UGT and standalone I-RECs is whether the certificate arrives already tied to a volume of electricity. UGT uses what is called the bundled REC method, in which the physical delivery of electricity and the certificate travel together and appear as line items on the same electricity bill. A standalone certificate is procured through an entirely separate channel and lives in a separate ledger. That difference drives the reconciliation problem discussed below.

Direct PPA is not on the table for factories

The Direct PPA pilot approved by the NEPC on 25 June 2024 is often described as the moment factories gained the ability to choose their power source. Read the conditions and the scope narrows sharply. According to a commentary published by Nagashima Ohno and Tsunematsu on 22 December 2025, the pilot covers 2,000 MW and is limited to data centres holding a BOI promotion certificate. Additional conditions apply, including an IT base load of at least 50 MW per building, a parent-company mandate for 100 percent renewable energy, and the requirement that the facility not already be operating.

An ordinary manufacturing plant cannot enter this quota today. The routes actually available to a factory in 2026 are UGT1, UGT2, rooftop solar for self-consumption and standalone I-RECs, and only UGT and rooftop solar change the electricity itself.

Buying it is not the same as proving it

Location-based and market-based are two different calculations

The effect of renewable procurement shows up in Scope 2 accounting. What matters here is that Scope 2 has two calculation methods and that reporting both is the default expectation.

The location-based method uses the average emission factor for the local grid, in Thailand the factor published by the Thailand Greenhouse Gas Management Organization. Under this method the source of your electricity is irrelevant. Consume the same kWh and you report the same emissions. Signing a green tariff moves this number by exactly zero.

The market-based method uses emission factors derived from contracts and certificates. Hold certificates and you apply a zero emission factor to the corresponding volume. This is the only place where renewable procurement shows up.

Switch contracts without understanding this and you will receive a report saying the plant went renewable but CO2 did not fall. Nothing failed. The number that moves simply was not produced. We cover the underlying calculation work in our article on visualising CO2 emissions and calculating Scope 1 and Scope 2 at Thai plants.

A certificate only means something once it is tied to a volume

To claim a zero emission factor under the market-based method, the number of certificates has to correspond to the electricity volume you apply them to. One I-REC corresponds to one MWh. A site consuming 6 million kWh a year needs 6,000 certificates to cover its full load.

The problem is who maintains that correspondence. On the consumption side, a plant typically has monthly utility invoices and, at best, records from submeters inside the fence. On the certificate side, purchase timing and quantities sit in a different system entirely. In most sites, the two are reconciled once a year, by hand, in a spreadsheet.

With 3 sites you also have an allocation problem

Japanese manufacturers with several plants in Thailand carry an extra layer. If certificates are bought centrally and allocated across sites, the allocation basis has to be actual consumption or it cannot be defended.

The following is a model used throughout this article. It does not describe any real company.

The consumption ratio in that table becomes the allocation ratio for certificates. In practice, one of the 3 sites always breaks the assumption. Output rose and consumption with it. A submeter circuit was down and the actuals are missing. Site C turns out not to fall under Type 4 and cannot join UGT2 at all. Any one of these breaks the allocation.

Three checks that prevent double counting

The question asked most severely in the certificate world is whether two companies are claiming the same MWh. Three checks address it.

First, whether the generation period and the consumption period fall inside the same window. Certificates record the period in which the electricity was generated, and a certificate that straddles reporting years generally cannot be applied.

Second, whether the geographic boundary lines up. A certificate issued on the Thai grid is meant to be applied against consumption in Thailand.

Third, whether the certificate has been retired. Once used, a certificate is retired in the registry and can never be used again. Report against a certificate you bought but never retired and an auditor will raise it.

None of these three can be confirmed without matching the consumption data held by the plant against the attributes of the certificate. The coarser your matching unit, the more formal and less meaningful the check becomes.

The GHG Protocol revision will decide your metering granularity

From annual matching to hourly matching

Everything so far assumes the current Scope 2 Guidance, published in 2015. That assumption is now changing.

According to the GHG Protocol’s own blog, the Scope 2 technical working group has advanced three proposed changes to the market-based method.

The first is hourly matching. Today, certificates and consumption need only be matched on an annual basis. Under the proposal they would have to be matched hour by hour.

The second is a deliverability requirement. Energy attributes would have to be sourced from grid regions that can physically deliver power to the point of consumption.

The third concerns allocation of standard supply service. Each reporter could claim only its pro-rata portion, and unclaimed attributes would not be available for voluntary market-based claims.

Publication of the revised Scope 2 Standard is anticipated in late 2027, with implementation phasing in over several years after that.

What happens to a plant that only has monthly invoices

Once hourly matching bites, the plant-side assumptions change completely.

Annual matching needs one scalar value, total consumption for the year. Add up 12 monthly invoices and you have it. Hourly matching needs a time series of consumption by hour. 24 points a day, 8,760 points a year. You cannot build that from invoices.

Drop to that granularity and one mismatch always surfaces, between solar-derived certificates and night-time consumption. Certificates that looked sufficient on an annual average decompose into a surplus during the day and a shortfall at night.

To make this concrete, split the day into 6 blocks of 4 hours. Solar generation realistically covers 2 of them. A plant running 3 shifts around the clock has 4 remaining blocks whose consumption no solar-derived certificate will match. What balanced out as an annual total appears as 4 holes at block level.

What interval to keep, and for how long

The answer for plant teams is straightforward. Start capturing and retaining hourly consumption data before hourly matching applies.

In Thailand, metering at the point of supply is performed by the utility’s revenue meter. That data comes back to the plant only in the form of an invoice. To hold hourly values yourself, you either install your own meter at the incoming point or pull data from the demand monitoring device already installed.

There is a rule of thumb for granularity. Thai demand charges are set on a 15-minute average demand, so any site already running demand monitoring holds 15-minute values. From those, hourly values can be aggregated. The reverse does not work. Capture at the fine interval and derive the coarse one.

Retention needs a decision too. Certificate reporting is annual, and once you allow for third-party verification and customer audits going back through the record, the minimum is 3 years and 5 is better. 15-minute values for 5 years across a combined 15 points, being 3 incoming points and 12 major loads, comes to 15 points times 35,040 points times 5 years, roughly 2.63 million rows. Trivial for a time-series database. Impossible in a spreadsheet.

We cover metering point design itself in our guide to the cost and implementation steps of an energy monitoring system.

A cost model

Stating the assumptions

What follows is a model, not a quotation. Treat every figure as something that will change at your site.

Assume a Japanese manufacturer with 3 plants in Thailand. Combined annual purchased electricity is 12 million kWh, split as 6 million at Site A, 4 million at Site B and 2 million at Site C. Assume a current average unit price of 4.20 baht per kWh. The UGT1 premium is the 0.0375 baht per kWh published by MEA, and the UGT2 unit price is the reported 4.56 baht per kWh.

Three procurement scenarios

A note on how each line was derived follows.

Scenario 1 is 12 million kWh at 0.0375 baht, or 450,000 baht. Scenario 2 takes the 0.36 baht gap against the current unit price across 12 million kWh, or 4,320,000 baht. Scenario 3 puts 1,000 kWp on the roof at Site A, assumes Thai yield of 1,400 kWh per kWp per year for 1.4 million kWh of self-consumption, saves 1.4 million kWh at 4.20 baht for 5,880,000 baht, and applies UGT1 to the remaining 10.6 million kWh for 397,500 baht. Net benefit is 5,482,500 baht a year, and simple payback on the 25 million baht investment is about 4.6 years. Whether the array is actually producing what the plan assumed is a separate question, covered in the four blind spots that hide falling solar output.

The cost of proving it sits on top

Everything above concerns the electricity itself. The cost of making the claim provable is separate. Building it up in layers keeps things from being missed.

Choose scenario 1 and the electricity itself costs 450,000 baht a year extra, while proving it costs 1,775,000 baht upfront and 680,000 baht a year. The proof costs more than the power. That is not an anomaly. It follows directly from how cheaply the UGT1 premium is set.

Read the other way, layers 1 through 3 are the components of an energy monitoring system and are shared with electricity cost reduction and ISO 50001 operation. Invest in them solely for certificates and the payback case falls apart. At a site already pursuing power visibility, the incremental work is layers 4 and 5 only. Our view on running this inside an ISO 50001 framework is in ISO 50001 energy management and why payback starts with power visibility.

Four things that actually trip up Thai sites

Nobody knows the tariff class

UGT2 covers Type 4 and Type 5. Yet ask a plant contact what tariff class the site is on and few can answer. The class appears on the electricity invoice, but the invoice is processed by finance and never reaches the plant. If you are considering UGT2, the first task is not comparing vendors. It is reading one invoice.

Sites supplied through an industrial estate follow a different path

Many Japanese plants in Thailand sit inside industrial estates and receive electricity through the estate operator rather than directly from the utility. That raises the question of who the UGT counterparty is. Where the estate takes supply in bulk, a tenant plant may not be able to subscribe on its own. Treatment varies by estate, so reading the scheme documentation will not settle it. Ask the operator first.

Head office timing and the registration window do not line up

MEA’s UGT1 registration for 2026 ran from 16 March to 10 April. Once a year, for a limited window. Meanwhile parent-company budgets and sustainability policies in Japan are typically finalised across March and April. Wait for head office and you miss that year’s window. Working backwards, the local team needs to be in a position to decide by the fourth quarter of the preceding year.

Certificates get bought but never retired

Retiring a purchased certificate in the registry tends to fall outside procurement’s scope of work. Once bought and paid for, the task looks finished, and year end arrives with the certificates unretired. Report against unretired certificates and you invite a double-counting finding. Decide who retires them and when, at the moment of purchase.

How to approach implementation

Five steps

Step one is fixing the destination of the report. A parent-company integrated report, a customer supplier survey and an initiative such as RE100 each demand different certificate types and different granularity. Start procuring before the destination is settled and you will end up holding certificates you cannot use.

Step two is taking stock of your current consumption data. How many incoming points are there, what is captured beyond monthly invoices, is a demand monitoring device installed, and are its values retained. Discovering here that 15-minute values already exist shortens everything downstream.

Step three is choosing the procurement route. Whether the tariff class is Type 4 or Type 5 determines UGT2 eligibility, and roof load capacity and area determine the ceiling on self-consumption. Both are physical constraints, so settle them first.

Step four is building the reconciliation. Metered values and the certificate ledger have to be matchable by site and by period. This step is the one most often deferred and the one most often questioned in an audit.

Step five is running both calculations in parallel. Produce location-based and market-based figures every month, so that the difference between them can be explained as the effect of procurement.

Five points to check when selecting

Metering granularity of 15 minutes or finer. A product fixed at 1-hour intervals may survive future hourly matching but is useless for analysing Thai demand charges.

The certificate ledger reconcilable against consumption data on the same screen. Manage them in separate systems and reconciliation reverts to manual spreadsheet work.

Site allocation rules held as configuration. Embedded in code, they require a change request every time the site structure changes.

Both location-based and market-based outputs. A product that produces only one of them fails at the last step of reporting.

Retrospective recalculation of historical data. Emission factors get updated afterwards, and if past years cannot be recalculated, your time series comparison breaks.

Who owns this

This project does not close within plant IT alone, nor within the sustainability function alone. Meter installation belongs to electrical maintenance, tariff class confirmation to administration or finance, certificate purchasing to procurement, and reporting to head office sustainability.

The workable approach is for the plant to complete step two first and enter the discussion with head office holding one sentence, namely whether or not the site already has 15-minute values. That one sentence changes how long the discussion takes.

Frequently asked questions

Should we choose UGT1 or UGT2

It depends on the level of proof being demanded. If a certificate is sufficient, entering through UGT1 and reviewing annually is the low-risk path. If the customer or parent company specifies the generation technology, or if a long fixed rate has value, UGT2 is the answer. Sites not on Type 4 or Type 5 cannot join UGT2 at all, so the question does not arise.

Can we claim 100 percent renewable by buying I-RECs alone

If the certificate volume covers consumption, you can claim it under market-based Scope 2. Check the rules of the reporting destination first. Some initiatives impose additional conditions on the alignment of issuance and consumption periods, on the issuing region, and on the commissioning year of the generation asset. Discover those after purchase and you will be buying again.

Do we have to report market-based Scope 2 only

Both, as a rule. The location-based figure uses grid average factors and does not respond to procurement, while the market-based figure responds to contracts and certificates. Publish only one and the substance of the reduction cannot be read. Even where head office accepts only one format, keep both internally.

Do we still need certificates for rooftop solar we consume ourselves

Self-consumed generation is generally treated as renewable without certificates, since you generate and consume it yourself. The exception is where that output is exported to the grid, or where certificates have been issued against it and sold to a third party. What you sold belongs to someone else. Without measured generation data, you cannot make that separation.

Will Direct PPA become available to factories

The pilot as of 2026 is limited to large data centres holding BOI promotion. Extension to manufacturing belongs to the next stage of scheme design and depends on third-party access rules being settled. It should not be written into current plans.

What interval should we retain to prepare for hourly matching

15 minutes. Thai demand charges are set on 15-minute average demand, so any site with a demand monitoring device already holds those values. From 15-minute values you can aggregate hourly values, but not the reverse. For retention, 5 years is a reasonable target given third-party verification and customer audits.

In summary

By 2026, Thailand’s renewable procurement options are complete at a practical level. UGT1 offers roughly 2,000 GWh a year of hydropower on a 1-year contract at a 0.0375 baht per kWh premium. UGT2 offers 245 solar and wind facilities and roughly 12,000 GWh a year on a 10-year contract. Both come with I-RECs attached. Direct PPA remains data-centre territory for now, leaving UGT, rooftop solar and standalone I-RECs as the realistic mix for a factory.

Switching procurement, however, does not move the reported number on its own. Only market-based Scope 2 responds, and that calculation needs certificates matched to consumed kWh. The GHG Protocol revision then moves the matching unit from the year to the hour. What that will require is not a new contract but the retention of 15-minute values you probably already generate, in a form that can be reconciled against a certificate ledger.

Contracts can be changed in a few months. Historical metering data cannot be created retroactively. Start measuring only after the revised standard lands in late 2027 and the past you need for comparison will simply be blank. Metering comes first, the contract second.

TOMAS TECH has supported Japanese manufacturing sites across Thailand and ASEAN in putting shop-floor data to work, through PEGASUS and other production and energy management systems. Deciding where to place metering points, at what interval, and for how many years to keep the data affects both electricity cost reduction and CO2 reporting. We are happy to start well before any product comparison, at the stage of taking stock of what is and is not being captured today. To get in touch, use our contact page.

References

• MEA notice on UGT1 registration for 2026 and the 0.0375 baht per unit premium https://www.mea.or.th/public-relations/corporate-news-activities/announcement/mea-ugt1-registration-2026-green-energy
• EGAT official Thailand Utility Green Tariff site, covering supply sources, volumes, contract terms and the bundled REC method https://ugt-thai.com/
• ERC official Utility Green Tariff page, covering the founding resolution and the approval of UGT2 https://www.erc.or.th/th/util-green-tariff
• ERC public consultation on the UGT1 rate for the 2569 service year, held from 7 to 19 January 2026 https://www.erc.or.th/th/listen-to-opinions/590
• Thansettakij report of 19 April 2026 on the UGT2 management fee, eligible tariff classes, contract terms and UGT1 take-up https://www.thansettakij.com/sustainable/net-zero/656948
• SolarQuarter, Thailand Launches UGT2 Program To Boost Corporate Access To Renewable Energy, 20 March 2026 https://solarquarter.com/2026/03/20/thailand-launches-ugt2-program-to-boost-corporate-access-to-renewable-energy/
• EGAT I-REC Issuer, the local issuing body for I-REC(E) in Thailand https://irecissuer.egat.co.th/
• GHG Protocol, Upcoming Scope 2 Public Consultation on Hourly Matching and Deliverability https://ghgprotocol.org/blog/upcoming-scope-2-public-consultation-hourly-matching-and-deliverability
• Nagashima Ohno and Tsunematsu, Thailand’s Direct PPA Pilot Project, 22 December 2025 https://www.nagashima.com/en/publications/publication20251222-2/
• Reccessary, Thailand launches Utility Green Tariff, offering 2 billion kWh for SMEs https://www.reccessary.com/en/news/thailand-launches-utility-green-tariff-for-business
• Reccessary, Thailand to pilot 2 GW direct PPA for data centers in 2026 https://www.reccessary.com/en/news/thailand-dppa-data-centers
• Energy News Center report on the proposed 37 percent reduction of the 2569 UGT1 premium https://www.energynewscenter.com/3-%E0%B8%81%E0%B8%B2%E0%B8%A3%E0%B9%84%E0%B8%9F%E0%B8%9F%E0%B9%89%E0%B8%B2%E0%B9%80%E0%B8%AA%E0%B8%99%E0%B8%AD%E0%B8%9B%E0%B8%A3%E0%B8%B1%E0%B8%9A%E0%B8%A5%E0%B8%94%E0%B8%84%E0%B9%88%E0%B8%B2%E0%B8%9E/