A UK startup called Gigaton has built an AI platform to optimise how cement is manufactured, cutting costs and carbon emissions in the process, and now it has raised a Series A of $26 million to expand its business.
The company has now raised $35 million and its not disclosing valuation.
Cement is known for being ‘hard’ as a substance, but it’s also hard in terms of production, and a look at how it is made highlights some of the bigger gaps that exist as companies and countries look for more resilience in the industrial sector.
After water, cement is the world’s most-used and most-produced industrial material, with more than 4 billion tonnes made annually. This material is used directly as-is, but even more importantly, cement is a critical composite of one of the globe’s literal building blocks, concrete.
“Cement is a true foundational material,” said Gigaton’s CEO Josh Vernon in an interview with Resilience Media.
However, the process is far from perfect. Today, the most common method for cement production is carried out using legacy machines that are decades old, developed to be managed by teams of people, set up to make cement from the same materials that have always been used, and estimated to be responsible for some 8% of carbon emissions globally due to the amount of heat that is used in their kilns and the amount of fossil fuels — typically coal — that is used to produce that heat.
The production of modern concrete has other issues. First, there is a rising push to use products beyond coal for heating and cement production — a list of ‘no one wants this’ waste products that ranges from auto parts to sewage — but while these might reduce fossil fuel reliance, they are significantly less stable and need more management when used. Second, modern concrete processes often swap out cement for organic compounds for binding, one more driver for fossil fuel consumption.
Added to all this, cement is locked into a long-distance supply chain: more than half of the world’s cement made in China — a production cycle that many countries in the world are keen to shift and bring closer to home for all kinds of materials and goods. China, it should be noted, is quickly modernising and aiming ultimately to have ‘dark’ plants that work completely automatically, so the goal for any company looking to replace China as a supplier is to essentially improve and modernise its own process, and make it ‘darker’ too, to make it cost-effective. There is still time to develop this, Vernon said.
“In China, we see the most advanced set ups and they are moving to dark plants, but it’s half a decade off still if not more,” he said. “The largest companies in the west are moving towards remote operated control centres managing 10-20 plans. This is step one.”
Formed in academia, hardened in the real world
Gigaton’s not out to save everything, but rather to tackle the particular hard problem of cement as described above.
The company’s technology was initially developed in labs at Cambridge and UC. One of the company’s founders is Daniel Summerbell, who focused his PhD work at Cambridge University on looking for improving the production methods of cement, and then years after that honing that work as a researcher and fellow of the university’s Institute for Manufacturing. Summerbell is the company’s chief solutions architect.
A piece of hardware that Gigaton has developed sits on the equipment that’s used to produce concrete. This monitors the kilns’ performance and sends signals to control rooms to be able to control the heating more accurately, which especially important when the burning fuel is not the same coal that’s been used before and especially important as factories slowly move to increasingly dark.
“This tech is going to be the underlying infrastructure to move to these dark plants,” said Vernon. Right now, he added, the plants still require for someone to come and change the weights and balances of the kiln processes each time the measurements are taken and signal that shift.
The system in the process of that collects a lot of data about the cement manufacturing process that it uses to improve how it works, and to lock in information that will be used down the line at Gigaton for future products.
The company’s customers today include Adani Cement, Heidelberg Materials, and Holcim, and they estimate that using Gigaton’s are saving $1 million+ per year in energy costs and emissions.
While Gigaton is working to change the industry, it’s also undergone some changes of its own.
The company has been around since 2022 and until just a week ago, it was called Carbon Re.
The name change seems indicative of the bigger shift in climate tech: many companies in the space have found funding and interest in carbon emissions to have gone out of style, and so while a climate tech company may still be striving for the same end to improve emissions, the staying power of any single product might lie in whether a company can convincingly show that it is improving something else for customers at the same time. In its case, that improvement is in costs and controls, especially as plants modernise.
Gigaton’s first several years also were about more than cement: the company initially set out to look at steel and glass production alongside it. Today the focus is honed just on cement, although down the line soon it will bring steel and glass into focus, and use its vast data store to look at other functions it can improve in the process.
The startup may still have Summerbell on its solutions team, but others have moved on: Vernon only joined the company in 2024, taking over from another co-founder who had been the CEO, Sherif Elsayed-Ali, who is no longer at the company. Elsayed-Ali is a longtime figure in climate tech, who had previously headed up the climate business at the too-early-for-its-time, now-defunct Canadian startup Element AI. Another co-founder, Buffy Price, also worked with him at Element AI and is also still at the startup. (She is pictured here with Vernon, centre, and Summbell on the right.)
Indeed, while some things have shifted, other aspects of the startup have crystallised. The company is still very focused on large reductions of carbon emissions — a ‘gigaton’ of them — and as resilience has risen as an investment thesis, the startup’s work squarely fits into how a nation’s industrial base might see its own resilience take shape.
“It’s imperative that we work out how to build and manufacture more, and the base of that is the materials we build with,” Vernon said. “If it [forever] costs less to build in China, then we are royally screwed. We need tech to allow us to more sustainably produce materials here.”
This is largely what motivated this latest Series A, which is being led by one of the VCs that building a reputation around resilience tech, Plural, with participation from 2150, Semapa Next, and existing investors Planet A Ventures, Cambridge Enterprise Ventures, UCL Technology Fund managed by AlbionVC with UCL Business, and Clean Growth Fund.
“Cement, glass, and steel are the materials civilisation runs on, but producing them consumes about a quarter of global energy,” said Carina Namih, a partner at Plural, in a statement. “The Gigaton team combines deep AI expertise with years spent inside these plants understanding how they actually operate. By running a single facility using Gigaton’s AI, Adani, Heidelberg, and Holcim are already saving millions a year. The scale from here is enormous.”
And while so much AI focus right now is on Large Language Models and how they are being used (at great expense) to write code, write words, and do other things in place of humans, Gigaton is building AI for processes that humans can’t do, and to effectively replace and augment existing machines.
In effect, while a lot of AI these days is commonly associated with the development and application of Large Language Models into replicating a lot of human processes, Gigaton’s application of AI is around machine learning to help machines.
The startup is using some of the funding to hire, and it’s trying to encourage AI engineers who want to work on those problems and maybe a little less on LLMs to get in touch.
“This company is a place for them to run away from the LLM hype, frustrated that AI research is going into chatbots rather than something important,” Vernon said.
