SB 253 deadline: a 90-day plan to file your first California emissions report

The SB 253 deadline for your first Scope 1 and Scope 2 report is November 10, 2026. CARB announced on June 24, 2026 that it would defer the original August 10 date by three months, which means that if you are a US company with more than $1 billion in annual revenue doing business in California and you have not started, you have roughly one quarter to produce a greenhouse gas inventory you are willing to sign. That is enough time. It is not enough time to do it badly twice.

This is the plan we would run. It assumes you have no sustainability team, no consultant on retainer, and no emissions data beyond what is already sitting in your finance and facilities systems. That describes most first-year filers, and it is a better starting position than it sounds, because the data you need has been accumulating in your accounts payable system for years.

First, confirm you are actually in scope

Do this before anything else, because it takes an afternoon and it occasionally saves a quarter. SB 253 applies to entities formed in the United States with more than $1 billion in total annual revenue that do business in California. Two details trip people up. The threshold is measured on total revenue, not California revenue, using the lesser of your previous two fiscal years, so a company with $1.3 billion globally and a single California sales office is in. And "doing business in California" is a broad test that catches many companies who do not think of themselves as Californian at all.

While you are checking, check SB 261 too, because its threshold is lower at $500 million and most companies over $1 billion are in scope for both. The two laws are frequently confused, and they ask for completely different things: we pulled them apart in SB 253 vs SB 261.

Days 1 to 20: draw the boundary and find the data

The boundary decision comes first because everything downstream inherits it. Decide which legal entities are inside the reporting boundary and write down why, using either operational control or financial control as your consolidation approach. Pick one, apply it consistently, and record the choice. Auditors do not mind which approach you chose. They mind that you cannot say.

Then go find the data, which in practice means three pulls:

• Scope 1: natural gas bills, fuel card statements, fleet fuel purchases, propane and diesel deliveries, and refrigerant service records. Refrigerants are the one people forget, and in a company with a lot of cooling they are not a rounding error.
• Scope 2: electricity bills for every facility, plus any renewable energy contracts, power purchase agreements or renewable energy certificates you hold. You will report both a location-based and a market-based number, and the contracts are what make the market-based number different from the location-based one.
• Everything else: a full accounts payable export for the reporting year, at line level, with vendor, amount, date and general ledger account. You do not need this for the 2026 filing. You need it to know how bad 2027 is going to be, and to have a screening Scope 3 number in hand before CARB finalizes the rules.

That last pull is the one that stalls, because AP exports tend to arrive as PDFs from a system nobody wants to touch. If your invoice intake is still manual, this is the moment it starts costing you real money, and automating the invoice capture pays for itself well beyond the emissions project. Whatever you do, get the data at line level. A summarized trial balance cannot be classified, and a number you cannot trace back to a line is a number you cannot defend.

Days 21 to 50: classify, and record why

Now you turn records into an inventory. Each Scope 1 and Scope 2 source gets an activity quantity (therms, gallons, kilowatt hours), an emission factor, and a documented reason for the factor you chose. Use published factors: EPA's emission factors for greenhouse gas inventories for most US sources, and the EPA eGRID subregion factors for location-based electricity. Cite the version. Factors get revised, and "we used the 2025 eGRID release" is the kind of sentence that ends an audit question instead of starting one.

The part that decides whether this survives contact with an assurance provider is not the arithmetic. It is whether every reported tonne can be traced back to the invoice, meter read or contract that produced it. Spreadsheets fail here, and they fail quietly. A spreadsheet can hold the number. It cannot tell you, eleven months later, why row 4,812 was classified the way it was, or who decided, or what it looked like before someone edited it. That traceability is the entire job of carbon accounting software, and it is why we built ours to classify a ledger line by line and keep the evidence attached rather than to accept a total and multiply it.

Expect to find gaps. Every first inventory has them: a facility with no separate meter, a fuel card with missing months, an acquired subsidiary whose utilities are still on the seller's account. Estimate the gap using a documented method, label it as an estimate, and keep the label. An inventory with three clearly flagged estimates is credible. An inventory with three quietly buried ones is a problem waiting for an assurance provider to find.

Days 51 to 75: review, and do not skip the market-based number

Somebody other than the person who built the inventory needs to review it. Have them re-derive two or three material lines from the source documents and check the totals against something independent, like the prior year's utility spend. A 40% swing in reported electricity emissions with flat operations means a unit error somewhere, and you would much rather find it now.

Do both Scope 2 methods. Location-based uses the average emissions intensity of the grid you sit on. Market-based reflects the electricity you contractually bought, so it is where your renewable contracts and certificates actually show up. Companies that have paid for clean power and only report the location-based number are, in effect, throwing away the thing they bought. We walk through the boundary between the two in detail in Scope 1 and Scope 2 emissions.

Days 76 to 90: assemble, sign, and set up next year

Assemble the filing package: the inventory by scope, the boundary and consolidation approach, the methods and factor sources, the estimates and how you made them, and the evidence trail. Get it in front of whoever is going to put their name on it early enough that their questions can still change something.

One piece of good news for the first cycle: CARB has said it will exercise enforcement discretion on assurance, and will accept Scope 1 and Scope 2 data companies already hold or were already collecting, whether or not it received limited assurance. That relief is specific to 2026. Limited assurance is expected to arrive later, which is a strong argument for building an inventory now that could pass assurance, rather than one that gets you through November and collapses the first time somebody independent looks at it.

Then spend the last week on 2027, because Scope 3 disclosure begins that year and CARB is still running pre-rulemaking on the framework, boundaries, methods and assurance standards. You will not get final rules in time to start from scratch. What you can do is run a spend-based screening estimate over the AP export you already pulled, see which of the 15 Scope 3 categories actually carry your footprint (for most companies it is category 1 purchased goods and services, and for anything with a plug or an engine it is category 11 use of sold products), and start collecting real supplier data for those categories only. Screening first, precision second, in that order, is how you avoid spending a year chasing supplier data for categories that turn out to be 2% of your total.

What this costs you if you wait

The failure mode is not a missed deadline. It is filing something thin, and then discovering in 2027 that your baseline year cannot be recomputed, your Scope 3 estimate has no evidence behind it, and limited assurance has arrived. Restating a baseline is normal and expected when methods improve. Being unable to restate it, because nobody kept the ledger and the factors that produced it, is the thing you cannot fix later.

The inventory is a bookkeeping deliverable built from data you already own. Treat it that way, keep the evidence attached to every line, and the next four years get easier instead of harder. If you want to see what line-level classification looks like on your own data, the classifier on our SB 253 reporting software page runs on pasted lines from your ledger, no email required, and if you are still choosing a vendor, the honest comparison is on best carbon accounting software.

SB 253 facts here describe California law and CARB rulemaking as published in July 2026. The rulemaking is still open and dates can move, so confirm the current deadline with CARB before you file. This is not legal advice.