Zoho Launches In-House Server ‘Nathu La’ To Lower AI Inference Costs

Enterprise SaaS major Zoho has unveiled its in-house designed server platform, Nathu La, to cut AI inference costs and bolster end-to-end technology sovereignty from hardware to software applications across its global business operations.

The hardware promises 20-30% lower total cost of ownership (TCO) and up to 18% reduction in power consumption, according to the company. Nathu La was developed over five years by Zoho’s internal R&D teams and manufactured with Indian electronic manufacturing partners.

The server platform is built around Intel Xeon 6 processors and features in-house engineered motherboards, DC-SC (Data Centre Secure Control) modules, and network cards. The company claims it delivers substantial operational savings while reducing organisational exposure to foreign licensing dependencies and external security audits.

Zoho, which has filed five new patents for Nathu La for thermal management and modular server architecture, is aiming to deploy the servers across its application and SaaS infrastructure, optimising performance, cost and efficiency for AI inference, virtualisation, high-performance computing, and storage, while improving data governance for its global customers.

The company said this will also help in bringing down inference cost for Zoho’s AI usage. The SaaS firm has already deployed a few hundred Nathu La servers, with 1,000 servers in production and pre-production currently. and a target to deploy 2,000 by the end of the year, it told Moneycontrol.

“With Zoho’s strategy of using contextual, right-sized models, running on our own platform, now on our own servers, accelerated by our own GPU database, we are compounding the benefits accrued from owning and operating our entire technology stack,” Zoho Corporation CEO Shailesh Davey said.

The launch comes at a time when rapid enterprise adoption of AI is driving higher spending on inference, adding to India’s technology import bill. The burden is expected to grow as demand for AI-powered applications increases.

In fact, Zoho’s cofounder Sridhar Vembu recently likened the phenomenon to an ‘oil import bill’ for the AI era, warning that dependence on foreign AI compute infrastructure could become a vulnerability for countries like India, adding extra burden on its current account deficit. Zoho itself spends ‘a few million a year’ on AI model subscriptions, Vembu said then.

The post Zoho Launches In-House Server ‘Nathu La’ To Lower AI Inference Costs appeared first on Inc42 Media.