AI workload optimization startup Callosum raises $100M

London-based artificial intelligence startup Callosum Ltd. today announced that it has raised $100 million in funding. Atomico led the seed investment with participation from Plural, DCVC, and the U.K. Sovereign AI Fund. The deal follows a $10.25 million raise in February. Callosum offers a cloud service called Tailored Inference that helps developers increase the efficiency of their AI applications. According to the company, its technology can complete some inference tasks 3.7 times faster than GPT-5.6 Luna with better output quality. Callosum also promises to reduce infrastructure costs in the process. The tasks that AI applications perform often comprise multiple steps. Tailored Inference turns the steps that make up a task into standalone software modules called blocks. From there, it sends each block to the AI model best equipped to run it. Simple tasks are routed to low-cost algorithms while more difficult work is performed by frontier models. The new service also optimizes the underlying infrastructure. After it determines which part of an inference task should be performed by what model, the service deploys each model on the chip that can run it most efficiently. Tailored Inference supports AI accelerators from more than a half-dozen companies on launch. One of the chipmakers on the list, Cerebras Systems Inc., announced its partnership with Callosum today against the backdrop of the funding round. The companies will integrate its WSE series of wafer-size inference accelerators into Tailored Inference. Cerebras debuted its newest WSE chip, the WSE-3 Turbo, on Wednesday. It features the same core count as the company’s previous processor but provides twice the performance thanks to an increased clock rate. Cerebras’ engineers also upgraded the on-chip network that links together the cores. The company ships the accelerator as part of a module called the Wafer-Scale Backpack. It features liquid cooling hardware, power management components and other auxiliary parts. Three Wafer-Scale Backpack form the backbone of the CS-4, a rack appliance that Cerebras customers can use to deploy its silicon in their data centers. The company says that a single CS-4 can run models with 10 trillion parameters. The process through which AI models generate prompt responses comprises two steps known as the prefill and decode phases. According to Cerebras, the CS-4 is particularly well-suited for the latter step. The company says that the appliance can be used together with prefill-optimized chips from the likes of Advanced Micro Devices Inc. and Amazon Web Services Inc. Callosum’s platform can run customer workloads on both companies’ silicon. “The next-generation of AI will be defined by how intelligently compute is orchestrated, not simply how much compute is available,” said Callosum co-founder and Chief Executive Officer Danyal Akarca (pictured, right, with co-founder Jascha Achterberg). Photo: Callosum A message from John Furrier, co-founder of SiliconANGLE: Support our mission to keep content open and free by engaging with theCUBE community. Join theCUBE’s Alumni Trust Network, where technology leaders connect, share intelligence and create opportunities. 15M+ viewers of theCUBE videos, powering conversations across AI, cloud, cybersecurity and more 11.4k+ theCUBE alumni — Connect with more than 11,400 tech and business leaders shaping the future through a unique trusted-based network Are you an AWS customer? Support SiliconANGLE financially by buying your AWS services from our Marketplace portal page and links: https://siliconangle.com/aws-marketplace/ About SiliconANGLE Media