Gigaton secures $26 million to eliminate the control software utilized in heavy industry.

      A cement kiln is one of the most demanding machines in the industrial sector. Operating at fourteen hundred degrees, it cannot be easily halted, and the software managing its fuel mixture and oxygen levels is often outdated, older than the engineers who oversee it. Gigaton intends to replace that software with AI to manage the kiln instead. On June 3, it secured $26 million in funding to scale this initiative.

      The Series A funding round is led by Plural, with contributions from 2150, Semapa Next, and existing investors including Planet A Ventures, Cambridge Enterprise Ventures, the UCL Technology Fund managed by AlbionVC, and the Clean Growth Fund. This round brings the company’s total funding to over $35 million and will allow for a fivefold increase in staff and expansion into sectors beyond cement, such as steel, glass, and chemicals.

      Previously known as Carbon Re, the company rebranded in late May. It initially emerged in 2020 as the first joint venture between University College London and the University of Cambridge, founded by Daniel Summerbell, Buffy Price, Sherif Elsayed-Ali, and Aidan O’Sullivan. Josh Vernon, a co-founder of the Australian fintech Earnd, will step in as CEO in early 2024. The rebranding reflects a broader goal: to not just focus on carbon reduction but to take control of the entire plant.

      This distinction forms the crux of the company's proposal. Most AI solutions in heavy industry operate alongside the existing control systems, providing recommendations for human operators to accept or disregard. Gigaton claims to have spent five years within control rooms analyzing the failures of these systems and has developed its technology to replace the control stack rather than simply advise it.

      The company’s software simulates process behavior, anticipates the impacts of actions prior to execution, and autonomously adjusts parameters such as fuel mix, kiln speed, and oxygen levels while continuously retraining using real-time data as conditions change.

      Gigaton supports its argument for AI-driven control with data showing that implementations with Mannok, Adani Cement, Heidelberg Materials, and Holcim can yield annual operational savings of $1 million to $3 million, alongside the avoidance of approximately 30,000 tonnes of CO2 per plant, aiming to scale towards $100 million or more for larger multi-site customers.

      These figures are provided by the company rather than independently verified, and while the comparison to emissions from 11,000 UK households is a rhetorical device, the named customers are credible and significant, with a venture investor backing these assessments.

      The company is addressing competitive concerns linked to geographic advancements. China is already developing "dark factories" that operate without on-site personnel, and Gigaton positions the rest of the world as lagging behind.

      There is genuine pressure behind this pitch. Rising energy costs, increasing market volatility, and the transition to alternative fuels have made plant operations more complex. Kevin Lunney, operations director at Mannok, noted that switching to solid recovered fuel instead of coal is "genuinely harder to operate with," as it varies in calorific value and moisture more than coal does, and the real challenge lies in getting control room operators comfortable with such a significant change.

      This highlights the less glamorous aspect of decarbonizing heavy industry: while the benefits in carbon reduction and cost savings can be substantial, the operational transition is where projects can either succeed or falter.

      A more daunting question arises with any autonomous-control proposal. Entrusting a fourteen-hundred-degree kiln to software that continuously retrains on live data requires a level of operator trust that mere recommendation tools do not demand, and the potential failure consequences are physical, not just financial.

      Gigaton assures that operators will understand the rationale behind every action taken. Whether this level of transparency can persuade plant managers to relinquish control is something that the next phase, involving multiple sites, will ultimately test.