Nexthop AI builds high‑efficiency networking hardware and software for hyperscale cloud operators running large AI workloads. Its scale‑out and scale‑across switches use integrated optics and an open‑source SONiC OS to deliver higher bandwidth per watt, lowering power costs and improving reliability, while a co‑development model lets providers customize solutions for rapid AI training and inference deployment.
Funding
$110M raised to dateRaised to date based on public sources. This may differ from the amount the company actually raised and is based only on what is publicly available on the internet.
3OFounders
Product
Problem
Cloud operators scaling AI workloads face rapidly increasing power consumption, high infrastructure costs, and networking bottlenecks that limit performance and reliability. Traditional data‑center switches cannot keep up with the bandwidth density and latency requirements of modern large‑scale AI models.
Solution
Nexthop AI designs and delivers high‑efficiency networking hardware and software optimized for hyperscale AI deployments. Their scale‑out and scale‑across switch architectures incorporate integrated optics to provide higher bandwidth per watt, reducing the cost per token processed. By offering a co‑development partnership model, Nexthop works directly with cloud operators to customize networking solutions that accelerate AI training and inference while maintaining rapid deployment cycles. The resulting infrastructure improves power efficiency, lowers total cost of ownership, and enhances reliability for the world’s largest AI workloads.
Target Audience
Primary customers are hyperscale cloud providers and NeoCloud operators that run large‑scale AI training and inference workloads and require ultra‑efficient, high‑performance networking solutions.
Features
- Scale‑out and scale‑across switch designs that enable seamless expansion across large data‑center fabrics
- Integrated optical interconnects delivering high bandwidth density with superior power efficiency
- Open‑source SONiC‑based network operating system for flexible, programmable networking
- Co‑development partnership model allowing hyperscalers to tailor hardware and software to specific AI workloads
- Fast deployment pipelines that reduce time‑to‑market for new AI infrastructure upgrades
- Built‑in reliability features such as redundancy and automated fault detection for mission‑critical AI services