WHAT IS AI-OPTIMIZED CONCRETE? THE ANSWER IS IN META’S MINNESOTA DATA CENTER.
With digital demand accelerating, Meta is scaling its data centers while targeting net zero by 2030. Photo courtesy of Meta.
The digital economy runs on physical infrastructure. Every AI query, every social media post, every streamed video passes through a data center — a building that demands precision construction at massive scale. Now, as Meta moves to more than double its AI infrastructure investment, the tech leader isn't just building faster. It's building smarter.
That includes Meta’s concrete. For its data center in Rosemount, Minnesota, Amrize, Meta and the University of Illinois Urbana-Champaign partnered to develop a first-of-its-kind, AI-optimized concrete mix. The ingredients? Academic research, material science expertise and open-sourced AI models. The result? A first-of-its-kind advanced mix with early strength and carbon footprint benefits.
OUR AMBITION: A MIX THAT COULD MOVE AT THE SPEED OF AI
Data centers are where Meta houses the technology that powers AI — the computing resources, cooling systems and data storage that millions of people depend on, explains Meta Principal Research Scientist Julius Kusuma. And the concrete matters more than many realize.
Mixing high-performance concrete for these types of slab or surface applications is more difficult than doing so for foundations or formed shapes, Julius says. “It requires a good finish quality for the installation, operation and maintenance of mission-critical servers, cooling and power equipment.”
To meet their practical and strategic needs — including being net zero by 2030 — Meta needed a mix that could deliver high strength, controlled set time, and reduced carbon load, all while supporting its growing AI ambitions. Ironically, AI was the answer.
THE SOLUTION: A FORMULA MIXING ALGORITHMS WITH FIELD EXPERIENCE
Lab-built data and open-source AI laid the groundwork for concrete ready for full-scale construction. Photo courtesy of Meta.
AI is good at processing a lot of data. And a concrete mix design has a lot of variables. So it made sense to use AI to help identify the right combination of ingredients — fast.
Amrize provided all ingredients, from Minnesota to Illinois, along with real-world materials expertise. University of Illinois engineers tested the 100+ unique mixes in the lab, creating a novel dataset that trained Meta’s open-source AI model. With all this input, the model could predict concrete strength and optimize ingredient proportions before a single yard was poured.
“AI-driven mix design lets us optimize concrete for performance, cost and carbon in one step,” said Illinois Grainger Engineering professor Nishant Garg, who led the data-generation effort in his lab. “As mixes become more complex — with innovative raw materials and product formulation — AI can also forecast critical properties like strength gain over time.”
Using its ECOtect™ low-carbon concrete, Amrize translated the AI’s predictions into a mix that could actually perform at full construction scale — not just in a lab — and meet the project’s ambitious performance criteria.
THE RESULTS: A FASTER, STRONGER, LOWER-CARBON STEP TOWARDS THE DIGITAL FUTURE
AI-guided concrete development is expanding what’s possible across the industry. Photo courtesy of Meta.
The mix excelled. It met every strength and performance requirement Meta set and was approved for expanded use across additional areas of the facility. The numbers tell the rest of the story: 43% faster in early strength development. An estimated 35% reduction in total concrete carbon footprint. A win for AI, and for humans.
“Partnering with Meta and using AI to develop an innovative concrete mix that meets the unique needs of data centers is just the beginning,” said Jaime Hill, president of Amrize Building Materials. “Using AI, we can optimize our specialized concrete formulations for data center requirements, from performance needs like strength and durability to thermal regulation and energy-efficiency. We look forward to continuing this exciting work with Meta to scale up the adoption of advanced tailored concrete mixes.”
