CoreWeave has agreed to acquire London-based Monolith AI, a company that develops machine-learning software for engineering and industrial applications. The deal extends CoreWeave’s cloud business beyond AI training and inference workloads and into industrial research and development. Terms of the transaction were not disclosed.
CoreWeave said the acquisition will combine its GPU-based cloud infrastructure with Monolith’s simulation and test-driven machine-learning tools. The company described the move as a way to support customers in manufacturing, automotive, aerospace, and other sectors that depend on complex physics and engineering design.
“Every leader we meet across the industrial and manufacturing sectors knows AI can transform their business. What they need are the right tools to use the technology to solve intractable physics and engineering problems,” said Brian Venturo, CoreWeave’s co-founder and chief strategy officer, in the announcement. “Those challenges have historically slowed industrial innovation, and Monolith has closed that gap”.
Monolith, founded by Dr. Richard Ahlfeld, applies AI to engineering workflows such as anomaly detection, test-plan optimization, and next-test recommendation. Its software is used by companies including BMW, Nissan, and Honeywell to reduce physical testing and shorten product-development cycles. “Monolith was founded to put AI directly into the hands of engineers, enabling them to create breakthrough technologies,” Ahlfeld said. “Joining CoreWeave will allow us to scale that mission dramatically”.
According to McKinsey & Company research cited in the release, applying AI in complex manufacturing environments can raise R&D efficiency by 20 to 80 percent. CoreWeave said the combined offering will help enterprises apply those gains by pairing Monolith’s domain-specific models with large-scale compute resources.
The transaction follows several other CoreWeave acquisitions this year. In March 2025 the company purchased Weights & Biases, a U.S. firm that provides experiment-tracking software used in machine-learning development. In September it acquired OpenPipe, a startup focused on reinforcement-learning tools for building AI agents. Together, those deals extend CoreWeave’s platform from infrastructure into model management, training, and now domain-specific applications.
CoreWeave has grown rapidly over the past two years as demand for high-performance GPUs surged. Once a cryptocurrency-mining operation, the company pivoted to AI infrastructure in 2019 and now operates 32 data centers in the United States and Europe. It supplies computing power to large AI developers including OpenAI and Meta Platforms. In March 2025 it raised $1.5 billion in an initial public offering on Nasdaq.
In the months leading up to the Monolith announcement, CoreWeave reported several large commercial agreements. It signed a five-year, $11.9 billion cloud-computing contract with OpenAI in March and a $14.2 billion infrastructure deal with Meta in September. Those contracts helped diversify revenue after Microsoft Corp. accounted for more than 70 percent of sales in 2024. Chief executive Michael Intrator told Bloomberg that expanding beyond a single major customer was a priority following investor concern about concentration.
The company has also been consolidating its physical footprint. In July, CoreWeave said it would acquire Core Scientific Inc. for $9 billion in stock, bringing 1.3 gigawatts of data-center capacity directly under its control and eliminating about $10 billion in future lease obligations. “Verticalizing the ownership of Core Scientific’s high-performance data center infrastructure enables CoreWeave to significantly enhance operating efficiency and de-risk our future expansion,” Intrator said at the time.
Industry analysts view the Monolith acquisition as consistent with that pattern of building control across the technology stack. Holger Mueller, vice president and principal analyst at Constellation Research, said the deal represents “CoreWeave’s first entry into the AI applications space” and could help balance its data-center utilization between large AI-lab contracts and steady enterprise workloads.
CoreWeave’s financial filings show the company reported nearly $2 billion in revenue in 2024, an increase of more than 700 percent from the previous year, and a net loss of $863 million. The business remains capital-intensive, with about $8 billion in debt as of 2024, largely tied to GPU purchases and data-center build-outs.
Expanding into industrial AI gives CoreWeave exposure to sectors that invest heavily in simulation and modeling. In the automotive industry, where design validation requires extensive computational testing, customers such as the Aston Martin Aramco Formula One Team are already using CoreWeave’s infrastructure. CoreWeave serves as the team’s official AI cloud-computing partner, supporting its first large-scale cloud facility for race-car development.
Nvidia, which holds a minority stake in CoreWeave, has promoted similar industrial-AI capabilities through its Modulus and Omniverse platforms. CoreWeave’s move to integrate Monolith’s software follows that trend among infrastructure providers to pair GPU capacity with domain-specific tools. Other firms in the same segment, including Lambda Labs and Rescale, have also expanded into higher-level software to strengthen customer engagement.
CoreWeave said the acquisition remains subject to customary closing conditions. Once complete, Monolith will operate within CoreWeave’s platform division. Both companies stated that their immediate focus will be on providing engineering and manufacturing customers with access to scalable GPU resources combined with Monolith’s AI-driven test-and-simulation tools.
The deal continues a period of rapid consolidation in the AI-infrastructure sector, where companies are seeking to link hardware capacity with specialized software in order to address specific enterprise workloads. For CoreWeave, the addition of Monolith places its cloud infrastructure in closer alignment with industrial and engineering use cases that depend on large-scale computation and data-driven modeling.
