CO2 Emissions Visibility for Thai Factories — Scope 1 and 2

“We meter our electricity every minute, so we can produce our CO2 number any time.” We hear this often at Japanese-owned factories in Thailand. Then we open the questionnaire the customer actually sent, and it turns out not a single line can be filled in as things stand. Measuring kWh and reporting tCO2e are two different jobs. In 2026 the EU’s CBAM entered its definitive regime and Thailand’s grid emission factors were updated on the same date. CO2 emissions visibility has shifted from a nice-to-have improvement project into a precondition for the paperwork your customers expect. This article narrows the whole question down to a single decision rule for how granular your measurement really needs to be.

Why factories that already meter electricity still cannot produce a CO2 number

Power monitoring only produces activity data

The emissions formula itself is almost disappointingly simple.

Emissions (tCO2e) = activity data × emission factor

For purchased electricity, the activity data is kWh and the emission factor is the grid factor for that country and that year. One multiplication. There is no difficult mathematics anywhere in it. And yet factories with minute-level electricity metering still cannot produce a report, because they only hold the left-hand side of that multiplication.

What a power monitoring system or an energy monitoring system produces is activity data. Minute-level kWh, demand per line, load factor per machine. That data is extremely effective for reduction activities, but it is not emissions. To turn it into emissions you need three additional things: factors, boundary, and evidence.

The three missing pieces are factors, boundary, and evidence

When we come into a project to move an existing power monitoring setup onto carbon accounting, these three are always the first gaps we fill.

Factors means knowing which institution published the value, as of what date, and covering which scope. In Thailand in 2026, the purchased-electricity factor published by TGO (Thailand Greenhouse Gas Management Organization) was updated with effect from 1 January. The same kWh produces a different reported figure once the factor changes. And data where you cannot explain afterwards which factor was applied is unusable in a verification setting.

Boundary means where you draw the line around what counts as your own emissions. Shared-area electricity in a rented factory billed through the landlord, company-owned vehicles mixed in with contracted logistics, a month where a rental generator was running. These situations are entirely ordinary at Japanese-owned plants, and the number moves every time the line is redrawn.

Evidence means the material that lets you explain to a third party where a number came from. Utility invoices, meter calibration certificates, fuel purchase slips, refrigerant charging records. A dashboard can look beautiful, but if there is no reconciliation against invoices behind it, a verification body cannot accept the figures.

kWh management and tCO2e reporting serve different purposes

Invest while still conflating the two and you will end up rebuilding. Here is the difference laid out.

Read down the right-hand column and the point becomes clear: what tCO2e reporting needs is not fineness. What it needs is to be in a form you can explain to somebody outside your company. That is the starting point of this article. How to build the measurement side itself is covered in our article on energy monitoring systems for factories, so please refer to that for how the activity data gets produced.

Four reasons CO2 emissions visibility became necessary for Japanese factories in Thailand in 2026

People have been saying “you will need this eventually” for several years now, but 2026 is the year the situation changed in a concrete way. Here are the four regulatory movements. Dates and numbers matter in this section, so we only state what has been confirmed.

Reason 1 — EU CBAM entered its definitive regime on 1 January 2026

The EU’s Carbon Border Adjustment Mechanism moved into its definitive regime on 1 January 2026. The shape of the mechanism changed from quarterly reporting during the transitional period to an annual declaration. Sales of CBAM certificates begin on 1 February 2027, and the first surrender of certificates takes place in 2027, covering goods imported during 2026.

Most Japanese-owned factories in Thailand do not export to the EU directly, so the reflex is to conclude that none of this applies. What actually shows up in practice is not direct application but inquiries arriving through customers. A trading partner that ships product into the EU needs intensity data for its own declaration, and asks the Thai suppliers of its parts and intermediate materials to provide it. You cannot answer that inquiry with a plant-wide annual emissions total. What is requested is a figure tied to a product or an item.

Reason 2 — TGO’s grid emission factors were updated with effect from 1 January 2026

The Thailand Greenhouse Gas Management Organization (TGO) updated the emission factors for Thai electricity. The values applicable from 1 January 2026 are as follows.

• Scope 2 (purchased electricity) = 0.4750 kgCO2e/kWh — applied to consumption of electricity purchased from the grid
• Scope 3 (upstream) = 0.0812 kgCO2e/kWh — the stages before generation, such as fuel extraction and transport

As a grace period for system updates, use of the previous factors is permitted until 31 March 2026. That grace period is surprisingly awkward in practice. A factory that used one set of factors from January to March and another from April onward must be able to explain, when the annual figure is produced, which months were calculated on which basis. If the factor is hard-coded inside the system, that explanation becomes impossible to construct.

The other point that is easily missed is that a separate upstream Scope 3 factor exists at all. As long as you consume purchased electricity, the extraction and transport of fuel by the power producer exists as upstream emissions. Report only Scope 2 and say “this is our electricity-related footprint” and you are presenting a number with the upstream portion missing.

Reason 3 — GX-ETS enters full operation at the Japanese parent company

The Japanese side is moving too. The amended GX Promotion Act took effect on 1 April 2026, and the emissions trading scheme (GX-ETS) entered full operation from fiscal 2026. It applies to operators whose direct emissions average at least 100,000 tonnes per year over the preceding three fiscal years, covering roughly 300 to 400 companies and about 60% of Japan’s greenhouse gas emissions.

A Thai subsidiary is not itself a direct subject of that scheme. What does reach you is that the parent company now has to talk about emissions on a consolidated basis. Once head office in Japan builds an internal aggregation platform to comply, that requirement flows down to overseas subsidiaries. The sudden message saying “from this period overseas sites will submit monthly in the same format” comes out of exactly this chain.

Reason 4 — Thailand’s Climate Change Act is not yet enacted, but the preparation clock is already running

This is where misunderstanding is most common, so we will state it precisely. Thailand’s Climate Change Act has not yet been enacted.

On 2 December 2025 the Cabinet gave approval in principle. From there it goes to review by the Council of State and then to parliamentary deliberation. Entry into force is expected in the first half of 2027. Once enacted, the framework is expected to include mandatory GHG reporting, an emissions trading scheme (ETS), and a carbon tax, but what will actually have to be reported, and to what depth, depends on subordinate legislation that has yet to be issued — reportedly as many as 50 separate instruments.

So at this moment, an explanation along the lines of “Thai law now requires this, so we have to comply” is simply not correct. There is still a reason to start early. The figures submitted in the first year a regime takes effect are built from the previous year’s data. For a regime entering force in the first half of 2027, starting measurement in 2027 is already too late. What you can do before the details settle is begin collecting activity data, and establish the pattern for boundary and evidence.

The four movements on one page

Look at the right-hand column. The granularity demanded differs from regime to regime. That table leads directly into the central theme of the second half of this article.

Where to draw the Scope 1, Scope 2 and Scope 3 lines in a Thai factory

Translating the three Scopes into actual factory equipment

The GHG Protocol definitions are abstract, so let us map them onto the equipment actually found in a Thai plant.

Scope 1 is what you burned directly and what you leaked directly. Boiler fuel, diesel for the emergency generator, LPG forklifts, petrol for company-owned vehicles, and refrigerant leakage from chillers and air conditioning. Refrigerants are frequently forgotten, but because their global warming potential (GWP) is high, even a small leaked quantity can amount to a non-trivial volume of emissions.

Scope 2 is emissions from the consumption of purchased energy. At a Thai factory this is mainly electricity bought from PEA or MEA, with purchased steam or purchased chilled water added in some industrial estates. This is the category to which the TGO factor of 0.4750 kgCO2e/kWh is applied.

Scope 3 is emissions occurring outside your own facilities. The category list runs from upstream to downstream, but what a factory faces first is the upstream portion of purchased electricity (0.0812 kgCO2e/kWh), plus purchased raw materials and parts, transport, and waste.

Quick reference for assigning equipment

Where the line actually gets drawn wrong on Thai sites

These are the four points we check first in any project. Calling them common would understate it. They come up on very nearly every site.

First, shared electricity in a rented factory. In leased plants inside industrial estates, air conditioning and lighting for common areas are often included in the landlord’s bill on an allocated basis. Does that allocation go into your Scope 2, or is it treated as Scope 3? You cannot decide without reading the lease and the billing detail.

Second, self-consumed solar. Power generated by rooftop solar and used on site is deducted from purchased electricity and can be treated as effectively zero emissions. However, the situation changes if the environmental attributes (renewable energy certificates and similar) have been sold to a third party. Once the attribute is sold, you can no longer claim that portion as zero.

Third, rental generators and temporary equipment. Equipment run only for a production surge or during a maintenance window drops out of the monthly aggregation entirely. The fuel purchase slips exist in accounting, but they never flow into the environmental data collection. That is a completely ordinary state of affairs.

Fourth, the boundary between company-owned vehicles and contracted logistics. The same truck is Scope 1 if you own it and Scope 3 if it is outsourced. A truck with your company name on the side turning out to belong to the logistics provider is a check you need to repeat every time.

None of these four are solved by adding another sensor. This is the work of reading contracts and slips to draw a line, not a measurement problem. The first thing to do when preparing for emissions visibility is not equipment selection. It is this line-drawing.

Calculating with the TGO emission factors — Scope 2 and upstream Scope 3

Working through a calculation (everything below is a hypothetical model)

The figures shown from here are a hypothetical model used for explanation. They are neither measured values from any specific factory nor an industry average. What we want you to look at is not the emissions or the money but the order in which the calculation is assembled.

Assume a Japanese-owned metal processing plant in a Thai industrial estate. Set monthly purchased electricity at 500,000 kWh. The factors used are the TGO values applicable from 1 January 2026.

Scope 2 (purchased electricity)

500,000 kWh × 0.4750 kgCO2e/kWh = 237,500 kgCO2e = 237.5 tCO2e per month

Scope 3 (upstream of electricity)

500,000 kWh × 0.0812 kgCO2e/kWh = 40,600 kgCO2e = 40.6 tCO2e per month

Converted to an annual basis, this is how it looks.

What we want you to take from this table

First, the upstream portion is about 17% of Scope 2. The ratio is 0.0812 divided by 0.4750, or roughly 0.171. Report Scope 2 alone and state that “this plant’s electricity-related emissions are 2,850 tonnes a year” and you are presenting something that is actually a 3,337-tonne story at about 85% of its size. Customer Scope 3 questionnaires frequently ask for a figure that includes upstream, so dropping it here means a resubmission later.

Second, the only input this calculation needs is a single number, monthly kWh. If all you want is a plant-wide annual total, twelve utility invoices are enough. No sensors, no gateways, no dashboard. This matters, because if hearing the phrase “CO2 visibility” sends you straight out for sensor quotations, you are starting from an investment you do not need.

The extra effort that appears when you add Scope 1

What happens when Scope 1 is added to the calculation above? Factors differ by fuel type, so we will not quote specific fuel factors in this article, but the practical workload splits sharply according to where the data comes from.

Purchased electricity (Scope 2) and its upstream portion (Scope 3) both use the same input, monthly kWh, which can be pulled automatically from the utility invoice or the incoming meter. This area is easy to automate. Boiler fuel, by contrast, has to be reconstructed from purchase slips and changes in tank stock, LPG forklifts come from the cylinder issue-and-return log on the floor, and refrigerant leakage comes from maintenance work records. Those last three do not emerge automatically from any system.

The point worth holding on to here is that the items hardest to automate are in fact settled by human operating rules. If maintenance keeps refrigerant charging records on paper, all you need is a named person to collect them monthly and an agreed cut-off date. There is no need to install a system for it. Conversely, installing a system without establishing that operating rule only adds one more master table that nobody populates.

The accident that happens when factors change over

There is one more consideration specific to 2026. TGO’s new factors apply from 1 January 2026, but the previous factors remain permissible until 31 March 2026. Fail to decide how to handle that grace period and the following sequence occurs.

• January to March was calculated on the old factors, April onward on the new ones
• But the system has only one field called “emission factor”, and it was overwritten partway through
• At year end, past months were recalculated and the January to March figures were overwritten with the new factor
• Nobody can now explain which month was reported on which factor

An emission factor should not be a setting. It should be a master record with a validity period. Keep the history in the form “valid from 1 January 2026, value 0.4750” so that past reported figures can be reproduced. This is technically straightforward, but if it is not decided in the initial design, retrofitting it is a nuisance.

Granularity is decided by the recipient of the report, not by the reduction target

This is the heart of the article.

When considering CO2 emissions visibility, most factories start from “how finely do we need to measure in order to reduce”. Measure per machine and you find waste; measure per line and you can compare. That thinking is not wrong, but it does not function as a basis for an investment decision. It only ever points in one direction, namely that finer is better.

The question that does function as a decision basis is a different one.

Who are you giving this number to?

Once the recipient is fixed, the required granularity, the required level of verification and the required update frequency are all determined uniquely. Start from granularity without fixing the recipient and you will always end up with either too much or too little.

The demands of each recipient on one page

The table below organises the reporting destinations we have actually handled at Japanese-owned factories in Thailand, by granularity, verification and frequency.

How to read this table

Compare the third and fourth columns. The more third-party verification a recipient involves, the lower the update frequency tends to be. (The clear exception is the CBAM-related inquiry demanding per-product figures, which we deal with in the next section.)

Internal reduction activity wants real-time data but is verified by nobody. TGO CFO registration undergoes third-party verification but happens once a year. There is a structural fact sitting underneath that.

Real-time capability and verification robustness are two separate requirements.

Assemble fine-grained real-time data and verification will still fail if reconciliation to invoices is missing. Conversely, an annual invoice-based aggregation will pass verification but is useless for reduction activity. Trying to satisfy both with one system is what makes the cost jump. And in most projects, only one of the two is actually needed.

Who actually demands a per-unit figure?

Within that table, the inquiry whose character is clearly different is the CBAM-related one. This is the only one that demands per-product figures.

Bringing a plant-wide emissions total down to a per-unit figure means dividing emissions by production output. In other words, emissions for a period and the quantity of what was produced in that period must close on the same period. And in a plant making multiple product types, you also need a rule for how much is allocated to which type.

There are four allocation methods, in ascending order of accuracy.

1. Allocation by unit count — divide total emissions by the number of units produced. Suits plants with a single product type or only similar products, but weight and process differences are not reflected.
2. Allocation by weight — distribute in proportion to product weight. Suits material processing, casting and extrusion, but differences in processing effort do not show.
3. Allocation by machine running hours — distribute in proportion to equipment run time. Suits high-mix machining and assembly, but requires per-machine measurement.
4. Allocation by measured consumption — distribute using actual measured kWh per line or machine. Effective where an energy-intensive process exists, but requires additional investment in measurement points.

Accuracy rises as you go down the list, and so does the number of measurement points and the linkage effort. Crucially, the top two require no additional measurement at all. Production quantities out of the production management system plus utility invoices are enough to calculate them.

If you are at the stage of simply needing to answer a customer inquiry with something, the realistic move is to produce a figure using unit-count or weight allocation and to state the allocation method explicitly. Measured allocation becomes necessary when the customer specifies the calculation method itself, or when the emissions difference between products starts to bear on price negotiations. Jump into measured allocation without making that judgement and you will invest in measurement points you did not need.

Should granularity be set to the most demanding recipient?

We are often asked whether it is better to build to the finest level from the start, thinking ahead. Our answer is conditional.

What should be fine-grained is the way the data is structured. Not granularity itself, by which we mean the number of measurement points.

Specifically, put the following three in place from the outset and raising granularity later becomes a matter of incremental investment.

1. Hold emission factors as master records with validity periods — as described in the previous section, so that past reported figures can be reproduced
2. Always attach where, when and which category to every piece of activity data — site, measurement point, period and Scope classification, four attributes. Hold the value alone and you cannot allocate it later
3. Decide the reconciliation key against production output — even just aligning the cut-off date between electricity and production dramatically changes whether you can later descend to a per-product figure

Conversely, we do not recommend installing measurement points throughout the plant from the start. Measurement points can be added later; structure cannot be fixed later. Decide the data structure first and add measurement points at the moment a recipient demands them. In that order, rebuilds almost never happen.

How to select measurement points on the electrical side, and how much can be taken from an existing incoming panel, is set out concretely in our article on factory power monitoring systems. If you are approaching this from measurement point design, read that one first.

A 90-day roadmap for moving existing power monitoring onto carbon accounting

If power monitoring or energy monitoring is already installed, moving to carbon accounting is far lighter than building from scratch. Here is the approach we use in projects, laid out in 90 days. The first 30 days are completed entirely on your side. The vendor, us included, only comes in from day 31.

Day 0 to 30 — decide before you measure

Measurement equipment does not come up at all in the first month. There are four things to do.

The inventory of reporting destinations means writing on paper the inquiries already arriving and those expected in one year and in three years. The parent company’s format, customer questionnaires, plans for TGO registration, submission requirements in tenders. This is the work of applying the table from the previous section to your own situation.

Fixing the boundary means drawing lines while reading the lease agreements and the billing detail. Tenant common areas, environmental attributes of solar, rental equipment, own vehicles versus contracted transport. This cannot be settled by the environmental function alone. Finance and purchasing have to be brought in.

Mapping existing measurement points means listing what is currently measured and where. At many factories this list simply does not exist. The figures appear on the BEMS screen, but nobody knows where the raw data sits or how many days of it are retained. That is a common state.

Building the factor register means putting the factors you use, their sources and their periods of applicability onto a single sheet. TGO’s electricity factor, factors by fuel type, refrigerant GWPs. That register can be used as-is later when responding to verification.

Day 31 to 60 — fill the gaps that systems cannot fill with human process

What becomes visible in month two are the items no system will fill. Refrigerant charge quantities, LPG cylinder counts, fuel for rental generators. Do not try to install a new system here. What is needed is a monthly close process specifying who submits what and when.

In our projects we start by defining a one-page input format and a cut-off date. Maintenance submits refrigerant charging records by the fifth of each month, purchasing submits fuel quantities, general affairs submits fuelling volumes for company vehicles. That alone fills most of Scope 1.

In parallel on the system side, implement the factor master with validity periods and attach site, measurement point, period and Scope classification as attributes on the activity data. The part that pulls values from the existing power monitoring can usually reuse a database or CSV export that already exists, so it does not turn into a large-scale modification.

Day 61 to 90 — run the close twice

Month three is the operational trial. Run the monthly close for real, twice, and confirm the following.

• Whether the aggregated electricity quantity matches the value on the utility invoice (where it does not, the cause is usually a gap between the meter reading period and your cut-off date)
• Whether all Scope 1 inputs arrive by the deadline
• Whether an estimation method is defined for when data is missing
• Whether the report format matches what the recipient requires

The reason for running it twice is that something always falls out on the first pass. Fold what was missed in the first close into the process, and pass it through on the second. Only once you have done this do you have a number fit to go outside the company.

Where the money actually goes — the cost drivers in CO2 emissions visibility

Now the budget conversation. When we lay out costs for a project, the item that moves the figure is not the number of measurement points. What moves it is the following six items.

Reading the right-hand column, the item proportional to measurement point count is mainly number 4, with number 3 partly affected depending on how the existing equipment was built. Yet the quotation a factory goes out to get first is almost always for number 4. And the items where the figure becomes unpredictable are numbers 3 and 5.

Number 3 varies enormously with the existing equipment. A panel where values come out cleanly over Modbus or OPC UA and a panel walled off behind a proprietary vendor protocol differ by an order of magnitude in effort. This is the part that cannot be quoted without a site survey, and the part most likely to be left marked “to be advised” as the project moves forward.

Number 5 arises the moment per-product reporting becomes necessary. As covered in the previous section, if unit-count or weight allocation is sufficient, all you need is to receive existing production output data monthly. Once measured allocation is demanded, it comes bundled with the measurement point additions in number 4 and the figure jumps. So whether you check your reporting destinations first translates directly into cost.

On investment incentives

The Thailand Board of Investment (BOI) offers investment incentives covering equipment upgrades that contribute to energy saving and GHG reduction. However, the eligibility conditions are updated each year, so we will not state exemption periods or ratios in this article. Building an internal approval case on the assumption of an incentive at the study stage can leave the premise broken when conditions have changed. Check the current conditions on the official BOI site, and consult a specialist before applying if needed.

One practical point worth remembering when preparing an approval case: what tends to qualify is the measurement equipment and energy-saving equipment in item 4, while the software and labour portions in items 1, 2, 5 and 6 are frequently out of scope.

How to connect a specific reduction theme such as air leakage to measurement is covered in our article on compressed air leak detection in factories. The thinking behind turning a reduction effect into a number is the same.

Three failures to eliminate before aiming for a carbon neutral factory

Finally, three causes of rebuilds that we have actually seen. None of them are technical problems.

Failure 1 — starting from the reduction target and leaving the recipient until later

Set a target of “X percent reduction by year Y” first, then work backwards to a measurement plan. It looks entirely reasonable, but in that order granularity never gets decided. For reduction purposes finer is always better, so no ceiling on the investment amount ever emerges.

Then six months later a CBAM-related inquiry arrives from a customer asking for a per-unit figure, and the situation is that detailed per-machine kWh data exists but cannot be brought down to a per-product level because production output and timestamps were never connected. Fine measurement was in place, and yet the requested form could not be produced. This is the most common rebuild pattern of all.

The countermeasure is simple. Do the inventory of reporting destinations first. Filling in one page of the table from the previous section is enough to determine the granularity you need.

Failure 2 — embedding the emission factor inside the system

2026 brought a TGO factor revision and a grace period running to 31 March. At factories that held the factor as a single item on a settings screen, that changeover overwrote historical data and produced incidents where which month was reported on which factor could no longer be traced.

Factors will continue to be updated. Hold them as master records with validity periods so that past reported figures can be reproduced at any time. This is a design decision on the build side, so a single line in the specification prevents it: “Emission factors shall be managed as master records carrying an effective start date, and recalculation of past periods shall use the factor in force at that time.” That one sentence is sufficient.

Failure 3 — building only a dashboard, leaving no evidence

A well-made dashboard is effective for internal explanation. But what a verification body looks at is not the screen. It is the record of where the number came from.

Concretely: that the aggregated electricity total matches the utility invoice, that where it does not the reason (a meter reading period offset, for instance) can be explained, and that the estimation method used in a month with missing data is recorded. None of this can be created after the fact. It is information that can only be preserved alongside the monthly close.

The same problem occurs in traceability work. The data exists, but because nobody defined what it is evidence of, it turns out to be unusable when submission is actually requested. Our article on the cost and approach for building a traceability system covers how to run such a build, and the thinking on evidence design carries across.

Frequently asked questions

What does CO2 emissions visibility actually mean?

It means a mechanism that takes a factory’s activity data (electricity in kWh, fuel consumption, refrigerant charge quantities and so on), multiplies it by emission factors to produce figures in tCO2e, and aggregates them continuously in a form that can be explained externally. It is frequently conflated with electricity or energy monitoring, but monitoring produces only activity data, not emissions themselves. Turning that into emissions requires three additional things: factors (whose published value, and as of what date), boundary (how far your own emissions extend), and evidence (where the number came from).

What is the difference between Scope 1 and Scope 2?

Scope 1 is emissions from what you burned directly and what you leaked directly. Boiler fuel, diesel for the emergency generator, LPG forklifts, petrol for company vehicles, and refrigerant leakage from air conditioning and chillers all fall here. Scope 2 is emissions from consuming purchased energy, which at a Thai factory means mainly electricity bought from PEA or MEA. Beyond that, Scope 3 covers emissions outside your own facilities, including the upstream of purchased electricity (the extraction and transport of fuel by the power producer), purchased materials, transport and waste. As factors applicable from 1 January 2026, TGO gives 0.4750 kgCO2e/kWh for Scope 2 purchased electricity and 0.0812 kgCO2e/kWh for upstream Scope 3.

How do you calculate a factory’s CO2 emissions?

For purchased electricity alone, you simply multiply monthly kWh by the emission factor. Taking a hypothetical model of a plant using 500,000 kWh a month, Scope 2 is 500,000 × 0.4750 = 237,500 kgCO2e (237.5 tonnes) and upstream Scope 3 is 500,000 × 0.0812 = 40,600 kgCO2e (40.6 tonnes), which annualise to 2,850 tonnes and 487.2 tonnes respectively. These are illustrative assumed values, not figures from a real factory. Because the calculation needs just one input, monthly kWh, twelve utility invoices are enough if all you want is a plant-wide annual figure. Adding Scope 1 requires consumption by fuel type plus the corresponding factors separately.

Is carbon footprint certification required in Thailand?

As a legal obligation, it is not currently mandatory. Thailand’s Climate Change Act has not yet been enacted. The Cabinet gave approval in principle on 2 December 2025, and it is due to go through Council of State review and then parliamentary deliberation. Entry into force is expected in the first half of 2027, and the detail of any reporting obligation depends on subordinate legislation still to be issued. Meanwhile TGO’s CFO (Carbon Footprint for Organization) already exists as a voluntary programme. It conforms to ISO 14064-1:2018 and the GHG Protocol, and involves verification by a TGO-approved verifier followed by registration and publication with TGO. Since third-party verification is increasingly demanded by trading partners and in tenders, the closest description of reality is that it is not a legal obligation but is becoming a practical requirement.

If we already have a power monitoring system, can we produce CO2 figures straight away?

For a plant-wide annual figure, you can produce it from utility invoices alone, with no power monitoring at all. Conversely, having power monitoring installed will not give you a figure fit to go outside the company unless factor management, boundary definition and evidence organisation are in place. Where power monitoring starts to earn its place is at the stage where allocation to product or process level becomes necessary. At that stage, electricity quantity and production output for the same period must close on the same cut-off date, which makes the reconciliation design against the production management system more important than the measurement data itself.

How finely should CO2 emissions be measured?

As finely as the recipient of the report demands. Internal reduction activity gets real value from per-machine real-time visibility but requires no verification. Consolidated disclosure by the parent means per site, monthly. A customer Scope 3 questionnaire means plant-wide, once a year. TGO CFO registration means the whole organisation with third-party verification. Only CBAM-related inquiries demand a per-product figure. What should be made fine-grained is not the number of measurement points but the way the data is held: emission factors as master records with validity periods, activity data carrying site, measurement point, period and Scope classification, and a defined reconciliation key against production output. Settle those three first and granularity can be raised later.

Summary

Measuring kWh and reporting tCO2e are two different jobs. What power monitoring and energy monitoring produce is activity data, not emissions. Turning that into emissions requires three additional things: factors, boundary and evidence.

2026 is a year in which several regulatory milestones coincide. EU CBAM entered its definitive regime on 1 January, and TGO’s electricity factors were updated on the same date, giving 0.4750 kgCO2e/kWh for Scope 2 and 0.0812 kgCO2e/kWh for upstream Scope 3 (with the previous factors permitted until 31 March). In Japan, the amended GX Promotion Act took effect on 1 April 2026, and GX-ETS has been in full operation since fiscal 2026. Thailand’s Climate Change Act, meanwhile, has not yet been enacted. It stands at Cabinet approval in principle granted on 2 December 2025, with entry into force expected in the first half of 2027.

And here is the point we most wanted to convey. Granularity is decided by the recipient of the report, not by the reduction target. Internal reduction activity, consolidated disclosure by the parent, customer Scope 3 questionnaires, TGO CFO registration, and CBAM inquiries arriving via customers. Those five do not share the same required granularity, verification level or update frequency. Go out for sensor quotations before fixing the recipient and you get the rebuild where measurement was fine-grained yet could not be shaped into what was asked for.

The first thing to do is not equipment selection. Write the reporting destinations on paper, read the contracts and billing detail to draw the boundary, and build a one-page factor register. All of that can be done at zero investment. Measurement points can be added at the point a recipient demands them, and that will still be in time.

TOMAS TECH builds production management systems and energy and environmental data foundations for Japanese manufacturers in Thailand. Early-stage inquiries are welcome, including questions such as “a questionnaire arrived from a customer and we do not know where to begin” or “we would like to know whether our existing power monitoring can be used for carbon accounting”. If you bring your current measurement situation and the actual inquiry you received, we can start by working backwards from the recipient to organise the granularity you actually need, together. You can reach us through our contact page.

References

• European Commission, Directorate-General for Taxation and Customs Union — Carbon Border Adjustment Mechanism
• Norton Rose Fulbright — Thailand’s Climate Change Bill, Latest Updates
• The Nation Thailand — TGO sets new emission factors for electricity generation and transmission
• Thailand Greenhouse Gas Management Organization (TGO) official site
• Ministry of Economy, Trade and Industry (METI), Japan — Emissions Trading System
• Thailand Board of Investment (BOI) official site