Delos Data, which develops network chips and software to connect AI chips in a data center, raised $100M from Matrix, Playground, Socratic Partners, and others

The cost to connect AI hardware falls substantially, potentially lowering the barrier to entry for advanced AI deployments.

Delos Data, a company specializing in network chips and software for connecting AI chips in data centers, has successfully raised over $100 million in a funding round. The investment was led by Matrix, Playground, and Socratic Partners, with participation from Capricorn's Technology Impact Fund, Matter Venture Partners, IAG, and several industry experts. This funding will support the expansion of Delos Data's software and hardware engineering teams, accelerate product development, and bolster sales efforts. The company aims to address what it identifies as the next bottleneck in AI infrastructure: the network, particularly for agentic inference workloads. Delos Nonstop AI is their proposed solution, designed to handle persistent, large-scale AI tasks across diverse hardware and ensure continuous operation through failures, anticipating a significant increase in token demand by 2030.

Delos Data's significant funding round highlights a growing industry consensus that network infrastructure, not just compute or memory, is becoming the critical bottleneck for advanced AI applications. The company's focus on "Delos Nonstop AI" suggests a strategic bet on the future of persistent, agentic workloads that demand highly reliable and efficient interconnectivity across a heterogeneous mix of processing units. This positions Delos Data as a key player in enabling the next wave of AI development, where network performance and resilience are paramount to achieving scalability and cost-effectiveness.

The substantial investment in Delos Data indicates a shift in venture capital focus towards the foundational infrastructure required for next-generation AI. As AI workloads become more complex and data-intensive, the limitations of current network architectures are becoming increasingly apparent. Delos Data's approach, aiming to provide specialized chips and software for AI interconnectivity, suggests that dedicated solutions will be crucial for overcoming these challenges and unlocking further advancements in AI capabilities, potentially disrupting established networking hardware providers.