What a Tonne Costs the State

Sweden does not negotiate the price of carbon one project at a time. It prints the price in a status report and lets the queue sort itself. Klimatklivet is the country's national investment support programme for local and regional climate measures, run by Naturvardsverket under Klimatklivsforordning (2015:517). Its own accounting for 2015 through 2024: 42,326 applications received, 27,114 projects approved, SEK 19.7 billion of grant awarded against SEK 46.0 billion of approved investment cost. 86 per cent of the cumulative grant went to companies. In 2024 alone it was 95 per cent. Then the same records publish the cost of carbon. SEK 0.38 of public grant per kilogram of CO2e abated, cumulative. That is SEK 380 for a tonne of carbon, paid by a state that told everyone the rate in advance. In 2024 the figure was SEK 0.45 per kilogram, SEK 450 a tonne. No carbon market set those numbers. A programme rule did, and the rule was published.

Applications in a round are not judged on merit in the abstract. They are ranked on one ratio, the klimatnyttokvot: kilograms of CO2e reduced for every krona invested. Read that direction twice, because it is the whole game. The breakpoints are public record, drawn from the Riksrevisionen review. A ratio of 1.00 cleared the 2015 rounds. It fell to 0.75 in the first 2016 round, came back to 1.00 in rounds two and three, and sat between 0.75 and 1.00 across the four rounds of 2017. The current estimate sits lower again, somewhere between 0.5 and 0.75. Kilograms per krona, not kronor per kilogram. A project that removes a great deal of carbon cheaply scores high. A project that is technically elegant and expensive scores low, and there is no column on the form for elegance. The state published the ranking rule before it opened the round, which means the ranking rule is not a judgement. It is arithmetic.

The programme funds only work that would not have happened without the money. That is not a preference. It is the eligibility text. The project has to be a physical investment. It has to reduce emissions permanently against a business-as-usual baseline. It must not have started when the application is submitted. No procurement, no construction, no signed order before the decision. It has to be anchored in the county's own climate and energy strategy, and the applicant has to be able to co-finance it. Start building before the decision and the grant is revoked. Show that the investment would have happened anyway, and it is refused. The programme's own rejection data puts 22 per cent of refusals on insufficient emission reduction per krona invested. Withdrawn applications, projects already under way and incomplete submissions fill most of the rest. Naturvardsverket does not accept supplements after the deadline. I signed a EUR 20 million personal guarantee on a company that folded on Christmas Day 2008. Nobody handed me a document that priced what that signature was worth at the moment I made it. A published eligibility clause is the same document, except it arrives before you sign, and it is free.

Every number in the calculation comes from one published file, the emission values document dated 2026-03-25. Heating oil carries 288 kilograms of CO2e per megawatt-hour. Coal carries 385. Peat 425. Fossil gas 249. Household waste 144. District heating, 55 on the national average. The Swedish grid, 59. Waste heat carries zero. That table decides the answer before an engineer touches the drawing. A machine that displaces Swedish grid electricity abates almost nothing, because the grid is already close to clean. The identical machine displacing oil for industrial heat abates 288 kilograms per megawatt-hour. Same machine, same capital, different counterfactual. The grant is measured against what you replace, not against what you build.

No carbon market set those numbers. A programme rule did, and the rule was published.

Grant intensity is capped by the EU General Block Exemption Regulation, 651/2014: small enterprises up to 70 per cent, medium enterprises 40 to 65 per cent, large enterprises 20 to 45 per cent with a 50 per cent ceiling. De minimis aid runs to EUR 300,000 over three fiscal years. Companies of any size can apply. So can municipalities, regions, associations, foundations and housing cooperatives. Private individuals cannot. The exclusions are the part worth reading twice. Research belongs to other agencies. Information campaigns are out. Feasibility studies without a physical investment are out. Hydrogen refuelling stations sit in their own tender. Farm biogas for electricity and heat sits in its own round. Public charging has two tender windows of its own. A programme that pays for machines is not a programme that pays for thinking. It says so in the eligibility text, and the eligibility text is published before the round opens.

Entrade built 200 wood-gasification machines across the UK and Europe. I know what a unit of that hardware costs to design, prove and place, and I know what it costs to keep it running. The file is harder. A Swedish application needs a project description, a greenhouse-gas calculation against the official factors, an investment calculation showing payback, a cost breakdown and financing plan, the latest annual report, balance sheet and income statement, and a certificate of registration naming an authorised signatory. All of it in Swedish. A foreign company cannot apply directly. It needs a Swedish organisationsnummer, which means a privat aktiebolag with SEK 25,000 of minimum capital, registered at Bolagsverket in five to ten business days, then registered with Skatteverket for F-tax and VAT. Or a Swedish partner takes ownership of the installation and buys the technology in. Processing takes four to six months after the round closes. I have restructured a EUR 75 million industrial group and built companies across twelve countries. The engineering has never been the gate. The paperwork has.

Every number this piece uses is published. The programme's cost per kilogram is published. The breakpoints are published. The emission factors are published in a single dated file. The grant ceilings are published. The rejection reasons are published. The probability that a project below the breakpoint is approved anyway is not zero. The programme allows it where a project contributes to the diffusion of a technology or to other environmental objectives. But the majority clears on the ratio, and the ratio runs against a table anyone can read before writing a line of the application. A state that publishes its price is not making an argument. It is running a market with one buyer, and the buyer has already said what it will spend per tonne. The response to that number is a decision about what to build, taken years before the window opens. There is no negotiating with a published figure, and no negotiating with a breakpoint set after the round closes. Both exist before you arrive. This piece describes the published structure of the Swedish Klimatklivet programme, from a dossier dated 2026-08-12. It is not advice, and it is not an application in any round.