Callosum

Callosum is a London-based AI infrastructure firm building a software platform to manage and orchestrate artificial intelligence workloads across diverse and heterogeneous hardware. The company was founded by Dr. Danyal Akarca and Dr. Jascha Achterberg, who met while completing their PhDs at Cambridge University where their research focused on the intersection of neuroscience, computing, and AI. Their academic work, published in prominent journals, and industry experience, including collaborations with Google DeepMind and Intel, informed the company's core thesis: that true AI advancement will come from specialized systems working together, much like the human brain, rather than by scaling single models on uniform hardware.

Callosum's primary offering is a systems-level software that allows different AI models to run collaboratively across a variety of chip architectures, including hardware from Nvidia, AMD, AWS, Cerebras, and SambaNova. This approach treats the compute environment as a dynamic ecosystem, orchestrating tasks and assigning them to the most suitable hardware component to maximize performance. The platform is designed to break the dependency on a single hardware provider, a direct challenge to the current market dominance of single-chip solutions for scaling AI. Callosum reports that for complex tasks, its system can deliver significantly higher accuracy and speed at a fraction of the cost compared to conventional single-hardware setups.

In February 2026, Callosum announced it had secured $10.25 million in a funding round led by Plural, with significant participation from the UK's Advanced Research and Invention Agency (ARIA) and several angel investors. This capital is being deployed to expand its London-based team, scale the software stack, and secure compute resources. The company's business model targets two primary client types: enterprises developing complex, multi-agent AI systems and emerging chip manufacturers seeking to validate and market their new hardware. The platform already integrates with major cloud providers like AWS, Google Cloud, and Microsoft Azure, offering clients flexibility in their compute environments.

Keywords: AI infrastructure, heterogeneous computing, workload orchestration, systems-level software, multi-chip architecture, machine learning, deep learning, sovereign AI, compute optimization, neural networks, data center, HPC, high-performance computing, AI hardware, accelerator, GPU, CPU, ASIC, cloud computing, Plural, Danyal Akarca, Jascha Achterberg, ARIA, Cambridge, AI systems, multi-model, silicon, algorithms, compute stack