Callosum, an AI startup, has raised $100 million to refine software that automatically matches AI tasks to the most suitable chips. The funding round, led by existing investors, will support product development and expand its team.
The company’s technology addresses a growing bottleneck in AI deployments: efficiently routing complex workloads to hardware that can handle them without waste. By reducing latency and improving throughput, Callosum aims to cut costs for data centers running large-scale AI models.
With the new capital, Callosum plans to scale its platform and onboard more enterprise customers. The startup competes in a niche where specialized chips like GPUs and TPUs often sit underutilized due to poor task allocation.
The round signals continued investor confidence in tools that optimize AI infrastructure, a space drawing fresh attention as companies race to deploy generative AI efficiently.
