Gigaton raises $26M to rip out the control software running heavy industry

Gigaton, a startup formerly known as Carbon Re, raised $26 million in Series A funding on June 3 to replace legacy control software in heavy industrial plants with an autonomous AI system. The company, which has raised over $35 million total, will use the funds to expand its headcount fivefold and deploy its technology beyond cement into steel, glass, and chemical production. Gigaton claims its AI, which simulates and adjusts plant operations in real time, has already delivered $1 million to $3 million in annual savings and 30,000 tonnes of avoided CO2 per plant at customers including Mannok and Heidelberg Materials.

A cement kiln is one of the least forgiving machines in industry. It runs at fourteen hundred degrees, it cannot easily be stopped, and the software deciding its fuel mix and oxygen levels is often older than the engineers tending it. Gigaton wants to throw that software out and let an AI run the kiln instead. On 3 June, it raised $26M to do it at scale. The Series A is led by Plural, with 2150, Semapa Next and existing backers including Planet A Ventures, Cambridge Enterprise Ventures, the UCL Technology Fund managed by AlbionVC, and the Clean Growth Fund. It brings the company’s total funding past $35M and will fund a fivefold increase in headcount and expansion beyond cement into steel, glass and chemicals. Gigaton was known as Carbon Re, but rebranded in late May. The original spun out in 2020 as the first joint venture from University College London and the University of Cambridge, founded by Daniel Summerbell, Buffy Price, Sherif Elsayed-Ali and Aidan O’Sullivan. Josh Vernon, who previously co-founded the Australian fintech Earnd, joined as chief executive in early 2024. The renaming signals the wider ambition: not carbon reduction as a feature, but control of the plant itself. That distinction is the company’s whole pitch. Most AI sold into heavy industry sits on top of the existing control stack, offering recommendations a human operator can take or ignore. Gigaton says it spent five years inside control rooms learning why those systems fail, and built its technology to replace the control stack rather than advise it. Its software simulates process behaviour, forecasts the effect of each action before taking it, and then autonomously adjusts parameters like fuel mix, kiln speed and oxygen, retraining continuously on live plant data as conditions shift. The case for letting an AI take the controls rests on numbers the company supplies. Deployments with Mannok, Adani Cement, Heidelberg Materials and Holcim deliver $1M to $3M in annual operational savings and around 30,000 tonnes of avoided CO2 per plant, Gigaton says, scaling toward $100M or more across large multi-site customers. Those are company figures rather than independently audited results, and the comparison to 11,000 UK households’ emissions is a framing device, but the named customers are real and substantial, and a venture investor has put money behind the readings. The competitive anxiety the company is selling against is geographic. China is already building “dark factories”, plants that run without on-site operators, and Gigaton frames the rest of the world as falling behind. There is real pressure underneath the pitch. Energy costs have climbed, market volatility has grown, and the shift to alternative fuels has made plants harder to run, not easier. Kevin Lunney, operations director at Mannok, put that last point concretely. Moving to solid recovered fuel instead of coal, he said, is “genuinely harder to operate with,” varying in calorific value and moisture in ways coal does not, and the real challenge is making operators in the control room comfortable with being asked to do something so different. That is the unglamorous reality of decarbonising heavy industry: the carbon and cost benefits are large, but the operational transition is where projects succeed or stall. The harder question is the one any autonomous-control pitch raises. Handing a fourteen-hundred-degree kiln to software that retrains itself on live data demands a level of operator trust that recommendation tools never required, and the failure modes are physical, not just financial. Gigaton’s answer is that operators see precisely why each action is taken. Whether that transparency is enough to make plant managers cede the controls is the thing the next phase, dozens of sites, will actually test. Get the TNW newsletter Get the most important tech news in your inbox each week.