Zoho Corporation has expanded its infrastructure capabilities by launching an in-house-designed server platform named Nathu La. The development marks a strategic shift for the Indian enterprise software provider as it blends proprietary hardware design with its existing cloud and artificial intelligence applications.
The server platform aims to lower data center operating costs and optimize performance for intense computational tasks. According to data released by the company, the platform reduces power consumption by 12% to 18% compared to standard configurations. It also lowers the total cost of ownership by 20% to 30%. These efficiencies target the high operational expenses associated with running artificial intelligence inference workloads.
Developed in collaboration with Intel, the platform uses Intel Xeon 6 processors. Intel supplied technical expertise and enablement support during the design phase. Zoho engineered the modular components, including the datacenter secure control module and network interface cards, through its dedicated hardware engineering team. Local electronics manufacturing service partners handle the physical assembly within India.
The launch follows a shift in the hardware strategy of the company. In mid-2025, the firm suspended a separate $700 million proposal to establish a compound semiconductor fabrication plant under the India Semiconductor Mission. The board paused the fabrication project due to difficulties in securing a technology partner and a lack of certainty regarding the technical roadmap. Instead of commercial chip manufacturing, the company concentrated its resources on direct hardware architecture design.
A dedicated research and development team in Nagpur has spent five years working on the motherboard and chassis platform. The architecture adheres to Open Compute Project standards, prioritizing thermal efficiency and ease of maintenance. The company has filed more than five patents covering its thermal management systems and cost-optimized server design.
By controlling both the underlying server hardware and the software layer, the firm aims to reduce its reliance on third-party technology vendors. This approach provides tighter control over firmware management, data governance, and platform customization. The platform targets workloads such as virtualization, high-performance computing, storage, and AI inference.
The company plans to deploy its own enterprise applications on the platform. This deployment will combine its proprietary graphics processing unit database and contextual AI models with its own physical server infrastructure.