Safeguard Mechanism in 2026: What's Changing for Covered Facilities

In the first year of the reformed Safeguard Mechanism, 147 out of 219 covered facilities exceeded their baseline. That's 67%. The year before, under the old settings, it was 18%.

That single stat tells you everything about where the safeguard mechanism changes in 2026 are heading. The reform did exactly what it was supposed to do: it eliminated the headroom that let Australia's biggest emitters coast through compliance year after year. In FY2023, covered facilities collectively sat 31.7 million tonnes below their baselines. In FY2024, that headroom vanished — the aggregate position flipped to a net exceedance of 300,000 tonnes.

And baselines keep falling. Every year. 4.9% until 2030. No pause. No grace period. If your facility emits more than 100,000 tonnes CO2-e per year and you haven't built a forward compliance strategy yet, you're already behind.

We build carbon accounting software, so we see this problem from the data side. The environmental managers we talk to aren't confused about what the Safeguard Mechanism requires. They're drowning in the operational detail of actually proving their numbers are right — facility by facility, fuel source by fuel source, reporting period by reporting period. And when your baseline drops another 4.9% this July, the margin for measurement error shrinks with it.

How Baselines Actually Work (And Why They Keep Dropping)

The baseline is the emissions ceiling your facility can't exceed without triggering a compliance obligation. But it's not a fixed number. For most facilities, the Clean Energy Regulator calculates it using a production-adjusted formula: your actual production volume multiplied by an emissions-intensity value for each product, then scaled down by the annual decline rate.

That decline rate is 4.9% per year from FY2023-24 through FY2029-30. So if your facility's baseline was 150,000 tonnes in FY2024, it drops to roughly 142,650 in FY2025, then 135,690 in FY2026, and so on. By FY2030, you're looking at a baseline about 28% lower than where you started. And your actual emissions haven't changed.

That's the squeeze. Production-adjusted baselines mean your ceiling moves with output — if you produce more, the baseline rises somewhat. But the decline rate eats into that adjustment every year. You can't outproduce the ratchet.

There are three baseline types in the scheme. Standard baselines cover most industrial facilities and use the production-adjusted formula. Landfill baselines apply to waste facilities and factor in a default methane capture efficiency of 37.2%. And then there's the sectoral baseline — a collective cap of 198 million tonnes CO2-e per year for all grid-connected electricity generators, which works differently again.

For trade-exposed facilities — the ones competing against international producers who don't face equivalent carbon costs — there's a relief valve. Trade-exposed baseline-adjusted (TEBA) facilities can apply for a reduced decline rate. Manufacturing facilities can get rates as low as 1% per year. Non-manufacturing trade-exposed facilities can get down to 2%. But the application process isn't automatic, and the CER assesses it based on scheme cost as a proportion of your EBIT or revenue. You have to prove the standard rate would cause genuine competitive harm.

The Cost of Exceeding Your Baseline

Miss your baseline, and the penalties are blunt.

The civil penalty for non-compliance is one penalty unit per tonne of excess emissions, plus 100 penalty units per day the excess continues during the two-year period after the compliance deadline of 1 April. One Commonwealth penalty unit is currently $330. So if your facility exceeds its baseline by 5,000 tonnes and you haven't surrendered enough credits by April, that's $1.65 million in per-tonne penalties alone — before the daily accrual even starts. The daily component runs at $33,000 per day.

These numbers aren't theoretical. The Clean Energy Regulator has shown with Beach Energy's enforceable undertaking (for NGER reporting failures, not Safeguard specifically) that it's willing to act. The Regulator published a late reporters list that names entities publicly. Compliance enforcement is real, and it's getting more sophisticated — the CER now uses advanced data analysis tools to identify high-risk reporters for targeted audits.

But here's the thing most environmental managers focus on: the penalties are the stick you want to avoid. The real cost of the Safeguard Mechanism is the cost of managing your exceedance, not the cost of failing to manage it. That's where ACCUs and SMCs come in.

ACCUs, SMCs, and the Maths of Compliance

You've got two main options for covering excess emissions: Australian Carbon Credit Units (ACCUs) and Safeguard Mechanism Credits (SMCs).

ACCUs are carbon credits generated from emissions reduction projects — things like reforestation, landfill gas capture, or avoided deforestation. You buy them on the open market and surrender them to the CER to offset your exceedance. The spot price in early 2026 sits around $37 per tonne. That's well below the government's cost containment measure price of $82.68 for FY2025-26 (which rises at CPI plus 2% each year from a $75 starting point in FY2023-24). The cost containment measure is a backstop — if market prices spike above it, covered facilities can buy ACCUs directly from the government at that fixed price. So far, it hasn't been needed. Market supply is outstripping demand.

In FY2024, covered facilities surrendered 7.1 million ACCUs to manage exceedances. Over 75% of those came from the cheapest project types: human-induced regeneration, landfill gas, and avoided deforestation. That concentration in low-cost ACCU categories is worth watching. As baselines keep declining and demand for ACCUs rises, the cheapest credits will get snapped up first. EY's analysis projects ACCU prices staying flat around $30-35 per tonne until 2028, then climbing toward $70 by 2035 as the squeeze tightens.

SMCs are different. They're not offsets. An SMC is generated when a covered facility reduces its emissions below its baseline. In FY2024, 61 facilities generated a combined 8.2 million SMCs — but 75% of those were created by just 12 facilities. Only 1.4 million SMCs were actually surrendered for compliance. The rest? Banked. Facilities can bank SMCs indefinitely until 2030, and most are holding them to cover future exceedances rather than selling. That banking behaviour makes sense — if your baseline drops another 4.9% next year, today's spare SMCs become tomorrow's compliance currency.

Whether SMCs can be banked beyond 2030 is one of the big open questions the 2026-27 review will settle.

The 2026-27 Review: Why It Matters More Than Any Single Compliance Year

The scheduled review of the Safeguard Mechanism in FY2026-27 isn't a check-up. It's where the government decides the rules for the next decade of industrial decarbonisation in Australia.

Here's what's on the table:

Post-2030 decline rates. The current 4.9% annual decline is legislated through FY2029-30. After that, the indicative rate drops to 3.285%, with rates to be set in five-year blocks. But here's the problem — independent analysis from the Carbon Market Institute and RepuTex suggests that hitting Australia's 2035 NDC target (62-70% below 2005 levels) requires annual decline rates of 4.8% to 6.9%. The indicative 3.285% doesn't come close. The review must set post-2030 rates by July 2027, and those numbers will define whether the Safeguard Mechanism actually delivers on the 2035 commitment.

Coverage thresholds. The current 100,000 tonne CO2-e threshold could be lowered. Analysis suggests broadening coverage would modestly increase compliance demand and bring more facilities into the scheme. If you're currently just under 100 kt, the review could change your world.

TEBA arrangements. Whether the trade-exposed protections are too generous — or not generous enough — will be re-examined. The 1-2% decline rates for TEBA facilities are significantly below the standard 4.9%, and there's tension between competitiveness concerns and the need for the scheme to actually reduce emissions.

Banking and borrowing after 2030. Current borrowing rules let you pull forward up to 10% of next year's baseline, with interest rates increasing from 2% (FY2024-25 and 2025-26) to 10% from FY2026-27 onwards. Those settings expire at 2030. What replaces them will shape how facilities manage multi-year capital investment in abatement.

We don't know exactly where the review will land. And frankly, we're sceptical that anyone can predict post-2030 settings with confidence right now. But the direction is clear: tighter, not looser.

Which Industries Are Feeling It Most

The 219 covered facilities span mining, oil and gas, manufacturing, transport, and waste. Together they account for about 28% of Australia's national emissions — roughly 135.7 million tonnes CO2-e in FY2024, down 2.8 million from the prior year.

Mining and oil and gas facilities dominate the list. These are the operations with the highest absolute emissions and, often, the fewest short-term options for reducing them. You can't electrify a blast furnace overnight. You can't retrofit a natural gas processing plant in a compliance year. The physics of these industries means the abatement curve is long and capital-intensive.

And that's exactly why the flexibility mechanisms exist — MYMPs, borrowing, ACCUs, SMCs. The scheme is designed to give heavy emitters time and options. But "time and options" isn't the same as "indefinite tolerance." Multi-year monitoring periods are capped at five years and can't extend past June 2030. Borrowing gets expensive from FY2026-27 onwards with 10% interest. The longer you wait to invest in actual abatement, the more it costs in credits later. If you're already tracking against your NGER thresholds, you know how quickly these numbers compound.

For the waste sector, the picture is slightly different. Landfill baselines use a separate methodology, and methane capture upgrades can deliver relatively quick wins compared to mining abatement. But the default 37.2% methane capture efficiency baked into the baseline calculation can overstate your actual performance if your capture system isn't operating at that level.

Practical Compliance: What Environmental Managers Should Actually Do

We won't pretend there's a simple playbook for managing Safeguard Mechanism compliance at a facility emitting 100,000+ tonnes per year. The engineering and investment decisions are specific to your site, your process, and your sector. But the data side — the foundation everything else sits on — is where we see facilities consistently dropping the ball.

Model your baseline trajectory out to 2030. This sounds obvious. It isn't being done rigorously enough. Your FY2026-27 baseline is knowable today — you have the decline rate, your production variables, and your emissions-intensity values. Map every year from now to 2030. Then model what happens at the indicative 3.285% post-2030 rate, and again at 4.9% (which independent analysis says is more likely). That gives you a realistic corridor for compliance exposure.

Get your emissions measurement right. When your baseline drops 4.9% and your actual emissions are measured with an uncertainty of plus or minus 5%, you might cross the line without your emissions actually changing. The quality of your NGER reporting data isn't just a compliance requirement — it's the input to every Safeguard Mechanism decision you make. A 2% error in your Scope 1 calculations could be the difference between generating SMCs and buying ACCUs. At $37 per tonne for 5,000 tonnes, that's $185,000 riding on whether your numbers are right.

Build a forward credit strategy. ACCUs are cheap now. They won't stay cheap. If your modelling shows increasing exceedances through FY2028-2030, buying ACCUs at $35-37 today and banking them might look wise in hindsight. But we're honestly not sure the current supply glut — estimated at 14 million surplus units in 2025 — will persist as more facilities tip into exceedance. The market dynamics here are genuinely uncertain, and anyone who tells you otherwise is selling something.

Track the 2026-27 review closely. The post-2030 decline rate decision will be the single biggest determinant of your long-term compliance costs. If the review sets rates at 4.9% or higher, facilities that planned for 3.285% will be badly exposed.

Data Quality Is the Foundation — Not an Afterthought

Here's what ties all of this together. The Safeguard Mechanism isn't just an emissions cap. It's a financial instrument. Your baseline is calculated from production data and emissions intensity. Your compliance position depends on measured emissions. Your credit strategy depends on accurate forecasting. And all of it feeds from the same source: your facility-level data.

When the ANAO found that 72% of NGER reports contained errors, that wasn't just a reporting problem. For Safeguard facilities, every error in your NGER submission directly affects your baseline calculation, your exceedance position, and your credit obligations. A facility that under-reports production might get a lower baseline than it should. A facility that over-reports emissions might buy credits it didn't need.

We built Carbonly to handle the data extraction and calculation layer — pulling consumption figures from utility documents, applying the right NGA emission factors by state and fuel type, and maintaining an audit trail that links every reported number to its source document. For Safeguard facilities, that kind of data infrastructure isn't nice to have. It's the difference between a compliance position you can defend and one you're guessing at.

The 4.9% annual decline won't pause while you sort out your spreadsheets. Your next baseline is already set. The question is whether you know, today, where you stand against it.

Related Reading: