Startup AI Gigaton Raih $26 Juta untuk Gantikan Software Kontrol Industri Berusia Puluhan Tahun

Heavy industry runs on software that hasn't kept pace with the demands of modern production. Cement kilns turning at 1,400°C, glass furnaces, and chemical reactors are often guided by control systems that are decades old—forcing plant operators to make constant manual adjustments while burning more fuel than necessary. Gigaton, a London-based AI company spun out of the University of Cambridge and UCL, has just raised $26 million in Series A funding to throw that legacy stack out and let a self-learning AI take the controls directly .

On June 3, 2026, the company announced the round was led by Plural, with participation from 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 . The raise brings Gigaton's total funding to more than $35 million .

The capital will fuel a significant expansion: the company plans to grow its team fivefold and move beyond its initial stronghold in cement into steel, glass, and chemicals production . CEO Josh Vernon told Global Cement that the funding enables the company's mission to reduce emissions on a “gigaton” scale, with plans to scale deployment across “dozens of sites” in the next growth phase .

Rip and replace, not bolt on

Where most industrial AI offerings layer optimization on top of an existing control system, Gigaton replaces the underlying control stack entirely. The company describes the approach as “ripping out” the legacy software so its AI can directly run the plant . This is a fundamentally different architecture from conventional Advanced Process Control (APC) tools that sit on top and make suggestions.

In practice, the AI autonomously adjusts several critical parameters in real time: the fuel mix feeding a kiln or furnace, the rotational speed of the kiln itself, and the oxygen levels required for efficient combustion . These variables are interdependent and change constantly based on raw material quality, ambient conditions, and production targets. Gigaton’s system learns the plant’s behavior continuously and makes closed-loop decisions without waiting for operator input.

Proven results in cement, with deployment times measured in weeks

The company’s initial focus has been cement manufacturing, one of the hardest-to-abate industrial sectors. A case study with Heidelberg Materials documented concrete operational improvements: a 4% reduction in fuel cost index, driven by a 2.2% reduction in specific heat consumption, alongside a 33% decrease in C3S variability and a 2% reduction in fuel-derived carbon emissions . The system went from integration to live operation in eight weeks .

In its white paper, Gigaton reports that its AI can reduce fuel-derived carbon emissions by up to 5% at the pyroprocess stage—the most energy-intensive part of cement production . The software integrates with existing APC systems like ABB Ability and FLSmidth ECS/ProcessExpert, but takes over dynamic target-setting rather than just recommending adjustments .

From Carbon Re to Gigaton: five years of plant-floor research

The company was founded in 2020 as Carbon Re, a deep-tech spinout from the University of Cambridge and UCL . Early development involved more than five years of work alongside industrial plant operators, giving the team direct exposure to the constraints and failure modes of real production environments . The recent rebrand to Gigaton reflects a broader ambition: the name signals a commitment to removing billions of tons of CO2 across multiple heavy-industry verticals, not just cement .

Gigaton is part of a wave of companies applying AI to the physical world rather than to office-workflows or consumer software. As one analysis noted, this is “a different AI story from chat, search, or office workflow”—it sits inside physical production where timing, energy use, process stability, and equipment reliability matter in ways that a hallucination can't be tolerated .

What the $26 million will unlock

The Series A will fund two parallel tracks: continued development of the next-generation platform and broader deployment across the four target sectors . The fivefold team expansion signals that Gigaton is moving from a research-heavy phase into commercial scaling. Expansion beyond cement into steel, glass, and chemicals suggests the core technology is sector-agnostic—if an AI can learn to control one type of thermal process, it can likely learn another.

For heavy industry, the timing is pressing. Energy costs remain volatile, carbon pricing is expanding across jurisdictions, and plants face increasing pressure to reduce emissions without sacrificing output. A self-learning control system that can cut fuel consumption and emissions simultaneously, and go live in under two months, offers a tangible path forward for an industry that has been slow to digitize.