Zoho’s Own AI Server “Nathu La”: Redefining Infrastructure Sovereignty

Zoho has unveiled “Nathu La,” its first in-house AI server platform, marking a bold step toward reducing dependence on foreign OEMs and cutting escalating AI infrastructure costs. Built in Nagpur with Intel Xeon 6 processors, Nathu La represents Zoho’s ambition to control its full technology stack—from software to hardware—while ensuring sustainability, sovereignty, and efficiency.

Zoho is one of the very few SaaS firms globally to design its own AI servers (Nathu La), but it’s not alone — other SaaS and cloud-native companies like Salesforce, Oracle, and ByteDance have also invested in proprietary AI hardware stacks to reduce costs, improve sovereignty, and optimize workloads.

Zoho’s Nathu La server is built with security baked into every layer of its hardware and software. Because the entire design is created in‑house, Zoho doesn’t rely on outside companies for security checks, firmware updates, or licensing — giving it complete independence.

The platform also follows open‑source software principles and meets India’s highest local content standards, meaning most of its value is created within the country. This makes it especially relevant for government use.

Why Nathu La Matters

• Cost efficiency: Designed to lower total cost of ownership (TCO) by 20–30%, a critical advantage as global server prices have surged 3–4x in recent months.
• Energy savings: Optimized to reduce power consumption by 12–18%, aligning with Zoho’s sustainability goals.
• AI workload optimization: Tailored for inference-heavy tasks and large language model (LLM) deployments.
• Technology sovereignty: Eliminates reliance on Dell, HPE, Lenovo, and other OEMs, giving Zoho full-stack control.
• Talent pipeline: Developed by engineers trained under Zoho’s SETU program in Nagpur, reinforcing local skill development.

Nathu La vs Traditional OEM Servers

Feature	Nathu La (Zoho)	Traditional OEM (Dell/HPE/Lenovo)
Cost of Ownership	20–30% lower	Rising sharply due to AI demand
Power Consumption	12–18% lower	Higher baseline usage
Deployment	Internal Zoho data centres	Sold commercially worldwide
Optimization	Tailored for Zoho apps & AI workloads	General-purpose enterprise workloads
Technology Control	Full-stack sovereignty	Vendor-dependent
Commercial Strategy	Not for sale	Global enterprise sales

Strategic Context

Global AI infrastructure costs have skyrocketed due to supply chain bottlenecks and compute demand. By building Nathu La, Zoho ensures:

• Long-term sustainability through reduced costs and energy efficiency.
• Data sovereignty by controlling hardware and software layers.
• Performance per watt optimization, critical for scaling AI workloads.

This move also aligns with India’s sovereign cloud initiatives, where domestic firms are increasingly building indigenous infrastructure to reduce reliance on foreign technology ecosystems.

Risks & Trade-offs

• No commercial availability: Nathu La is strictly for Zoho’s internal use.
• High upfront R&D investment: Five years of development and Intel collaboration required.
• Scalability challenge: Initial rollout limited to ~2,000 servers by end of 2026.

Impact on Businesses

• Lower subscription costs over time.
• Faster AI-driven features.
• Stronger data governance and sovereignty.

Zoho’s Nathu La is more than just a server—it’s a statement of technological independence and a blueprint for how Indian SaaS firms can build resilient, cost-effective AI infrastructure.

By aligning with national programs like Make in India, Atmanirbhar Bharat, and the National Supercomputing Mission, Nathu La strengthens India’s ability to build its own high‑performance computing systems and sovereign digital infrastructure.

Altos India Unveils Next-Gen AI Server to Accelerate Enterprise Transformation

Altos India, a subsidiary of Acer Inc., today announced the launch of its latest AI server, Altos BrainSphere™ R680 F7, designed to power the next wave of enterprise AI deployments across India. The new server will be officially available for shipment from September and is engineered to meet the surging demand from enterprises, government, education, and healthcare sectors for high-performance, scalable, and future-ready AI infrastructure.

The Altos BrainSphere™ R680 F7 supports up to 8 NVIDIA GPUs, including RTX PRO 6000 Blackwell Server Edition, H200 NVL, and L40S, along with 6th-generation Intel Xeon processors and DDR5 memory. Its flexible PCIe architecture allows seamless integration of diverse NVIDIA accelerated computing technologies, making it ideal for AI inference, AI model optimization, data analytics, industrial AI, virtualization, visual compute, as well as specialized applications such as AI medical imaging, smart security, and personalized recommendations. This makes it a powerful platform for rapidly building diversified AI workload environments.

The server can also be paired with Altos aiWorks, an AI computing platform that integrates hardware and software resources to simplify deployment, accelerate time-to-market, and support hybrid workload capabilities.

Altos aiWorks now includes advanced functionalities such as model deployment, job scheduling, resource monitoring, inference process management, and comes preloaded with multiple mainstream AI frameworks and models. It also integrates NVIDIA AI Enterprise software, including NIM microservices, enabling enterprises to build scalable, visualized, open, and highly available AI environments tailored to their business needs.

“The Altos BrainSphere™ R680 F7 represents a significant milestone in our mission to bring future-ready AI solutions to Indian enterprises,” said Harish Kohli, President and Managing Director, Acer India. “As AI adoption accelerates across industries, organizations are seeking infrastructure that combines scalability, performance, and ease of deployment. With this launch, Altos India is well-positioned to empower businesses, government, and academia with computing solutions that will shape the next phase of India’s digital transformation.”

This launch further strengthens Altos India’s positioning as a trusted partner for enterprise computing, delivering solutions that align with the country’s growing focus on AI-driven innovation and digital-first growth.

IBM Unveils Power11 in India: AI-Optimized Server Built for Mission-Critical Workloads and Hybrid Cloud

IBM (NYSE: IBM) has officially launched its most advanced Power server to date, IBM Power11, in India, marking a significant milestone in enterprise AI infrastructure. Purpose-built for the AI-first era. Power11 is engineered to deliver real-time inferencing, seamless hybrid cloud deployment, and unmatched reliability—all while reducing energy consumption and IT costs.

Designed for Mission-Critical Workloads

Developed with major contributions from the IBM India Systems Development Lab (ISDL), Power11 is tailored for enterprises operating complex, data-intensive workloads across sectors like banking, telecom, healthcare, retail, and public services. According to IDC, over one billion new logical applications are expected by 2028, and Power11 is positioned to manage this scale with AI-driven performance, security, and control.

Power11 is built for enterprises that demand high performance, security, and always-on operations while preparing for an AI-first future,” said Subhathra Srinivasaraghavan, Vice President, ISDL, IBM.

Key Features and Innovations

AI Where Data Lives

• On-chip inferencing enables AI models to run directly where data resides—whether in private data centers or hybrid cloud environments.
• Supports data sovereignty and compliance mandates critical to India’s evolving digital landscape.

Security and Resilience

• Offers 99.9999% uptime, the highest in IBM Power history.
• Features <1-minute ransomware threat detection via IBM Power Cyber Vault.
• Built-in quantum-safe cryptography protects against future threats like harvest-now, decrypt-later attacks.

Energy Efficiency

• Delivers 2x performance per watt compared to x86 servers.
• Introduces an Energy Efficient Mode with up to 28% better server efficiency than Maximum Performance Mode.

Future-Ready Architecture

• First IBM Power server to support the upcoming IBM Spyre™ Accelerator, arriving in Q4 2025, designed for AI-intensive inference workloads.
• Fully compatible with IBM’s AI stack including watsonx and Red Hat OpenShift AI.

Availability and Strategic Impact

IBM Power11 will be generally available from July 25, 2025, with the Spyre Accelerator expected later this year. The launch underscores India’s growing role not just as a consumer of cutting-edge infrastructure, but as a co-developer of global enterprise technology.

With Power11, IBM is not just delivering a server—it’s offering a platform for transformation. From UPI and 5G to smart manufacturing and e-governance, Indian enterprises now have a resilient, AI-optimized infrastructure to scale innovation responsibly and securely.

IBM Power11 will be generally available July 25, 2025. The IBM Spyre™ Accelerator is expected to be available in Q4 2025. To learn more about Power11, visit here.

IBM Power11 Debuts as the Most Resilient Server in IBM History, Targeting AI-Era Demands

IBM (NYSE: IBM) has officially launched IBM Power11, the latest generation of its enterprise-grade server platform.

Power11 is purpose-built to meet the demands of AI-driven workloads and hybrid cloud environments. It features a complete redesign across processor, hardware architecture, and virtualization software stack.

The new server is engineered to deliver resiliency, scalability, and performance for mission-critical operations.

• Built for the AI Era:
• Next-gen processor with up to 55% performance gains over Power9
• Zero planned downtime with 99.9999% uptime
• Quantum-safe cryptography and ransomware detection under 1 minute
• Support for IBM Spyre AI Accelerator, optimized for inferencing and LLMs
• Hybrid Cloud Flexibility:
• Available on-premises and on IBM Cloud via IBM Power Virtual Server
• Certified for RISE with SAP
• Integrates with watsonx.data and watsonx Code Assistant for i
• Trusted Across Industries:
• Ideal for banking, healthcare, retail, and government workloads
• Continues IBM Power’s legacy of supporting mission-critical, data-intensive operations

Tom McPherson, GM of IBM Power Systems, said, —

Power11 is the most resilient server in IBM Power platform history. It’s built to simplify operations, accelerate AI adoption, and ensure business continuity in a hybrid cloud world.