Zoho Introduces ‘Nathu La’ to Power Efficient AI Inference

Chennai, June 12: Zoho Corporation today announced the launch of “Nathu La”, its indigenously developed server platform designed to power AI inference workloads while significantly reducing data centre operating costs.

Built over five years by Zoho’s engineering teams in India, Nathu La represents a major step in the company’s expansion into hardware infrastructure, complementing its existing cloud software ecosystem. The platform is designed for modern compute demands including AI inference, cloud services, virtualization, and high-performance computing (HPC).

Focus on efficiency and cost reduction

According to Zoho, the Nathu La server architecture delivers improved efficiency with measurable gains in power and cost optimization, including:

  • Reduced power consumption compared to conventional server systems
  • Lower total cost of ownership (TCO) for enterprise workloads
  • Optimized performance for AI inference operations

The company said these improvements are aimed at addressing rising infrastructure costs driven by large-scale AI adoption.

Built for modern data centres

Nathu La has been developed with a focus on modularity, scalability, and secure infrastructure design. Key features include:

  • Custom-designed hardware architecture
  • Data Centre Secure Control Module (DC-SCM) integration
  • Modular server design aligned with Open Compute Project (OCP) principles
  • Enhanced thermal and power efficiency systems

Zoho also stated that the platform incorporates multiple in-house innovations, with several related patents filed in the area of server design and efficiency optimization.

Deployment and roadmap

The servers are already being deployed across Zoho’s own data centres as part of a phased rollout strategy. The company plans to scale deployments to thousands of units over the coming year as part of its infrastructure modernization program.

"DEHYDRATED FOODS The New-Age Snack?"