spot_imgspot_img

Top 5 This Week

spot_img

Related Posts

Sovereign AI startup JohnAIC raises $1.9 million to build up to 100 on-premise AI servers a month


Bengaluru-based JohnAIC, which builds on-premise AI servers for enterprises and government bodies, has raised $1.9 million in a pre-seed funding round led by Crane Venture Partners, with participation from Neon Fund.

Advertisment

JohnAIC said it will use the capital to expand its manufacturing, configuration and supply capabilities for AI servers, with the aim of reaching a production capacity of up to 100 AI servers a month.

The startup is investing in factory and supply chain infrastructure to serve Indian organisations building private AI environments and neoclouds. Such customers need reliable access to GPU servers, as well as integrated, enterprise-ready systems rather than standalone hardware, it said.

What JohnAIC builds

Founded by Sasank Chilamkurthy, JohnAIC sells integrated systems that combine high-performance GPU servers, inference software, optimised open-source models and local data management.

The pitch is that an enterprise can run AI inside its own environment without having to assemble servers, models, deployment tools and security controls on its own.

According to the startup its product as enterprise-hardened servers running voice AI and predictive models behind the customer’s firewall, pre-configured with open-source tooling. It positions the servers as a way to avoid recurring cloud costs and dependence on a single vendor.

Its customers are public sector organisations and large enterprises, including banks, healthcare companies and public sector undertakings (PSUs). The systems come with high-density computing, model optimisation and role-based access controls.

For banking, financial services and insurance (BFSI) institutions, the infrastructure allows AI to be deployed behind organisational firewalls. JohnAIC said the systems are designed to support Digital Personal Data Protection (DPDP) and Reserve Bank of India (RBI) requirements, while keeping sensitive financial data within the organisation’s own environment.

The startup is also working with Andhra Pradesh Eastern Power Distribution Company Limited (APEPDCL), a state power distribution utility. Its server is now embedded in the utility’s routine operational survey processes, where it converts field data into real-time insights.

“As inference workloads form the core of AI investment, we stand out as one of the few companies focusing specifically on the inference layer,” said Chilamkurthy.

“Indian enterprises require not just AI capabilities, but AI infrastructure that they can fully own, operate, and trust. Our mission is to make on-premise AI adoption as seamless as cloud solutions, while ensuring that compute resources, models, and data remain under the customer’s direct control,” he added.

From a ‘cloud in a box’ to enterprise servers

JohnAIC’s product started out as a much smaller machine. At The Fifth Elephant conference on 13 July 2024, the team, then working under the name Von Neumann AI, showed JOHNAIC as a personal AI server, described as “a cloud in a box, with a small footprint and no subscription fee whatsoever.”

That version ran on an Intel i5 12400 processor with an Intel Arc 770 GPU carrying 16 GB of memory, 64 GB of RAM and a 1 TB SSD, with a peak power draw of 500W. It shipped with open-source software including Llama 3, Ollama, ERPNext and Keycloak.

Chilamkurthy argued that cloud suits startups that need flexibility, but that established companies with steady workloads see costs from storage, networking, monitoring and data transfer pile up, potentially reaching “tens of crores per year” for mid-sized enterprises. He also pointed to data egress fees as a barrier to leaving a cloud provider.

The startup now pitches its servers to PSUs, BFSI institutions and large enterprises. Its leadership team, according to its website, includes Swetha Tanamala as Head of AI, Adithya Bhat as Head of Infrastructure, Pushpendra Rawat as Head of GTM and Sales, and Arun Kant Sharma as Head of Engineering.

Why the investors backed it

Anandamoy Roychowdhary, MD for India and APAC at Crane Venture Partners, said the firm was drawn by JohnAIC’s aim of maximising the number of AI tokens customers can generate for every rupee they spend.

“Too much of the current AI infrastructure is supply squeezed and increasingly unaffordable. Our mission is to make AI intelligence cheap and plentiful for India to use,” he said.

Siddhartha Ahluwalia, co-founder and managing partner at Neon Fund, said the firm backs founders building critical infrastructure that lets companies deploy AI quickly, securely and economically.

“JohnAIC is doing exactly that, and Sasank is the right founder to build it: a co-author of PyTorch and who has led AI deployments at Qure.ai. We’re excited to back JohnAIC as it builds the infrastructure powering enterprise AI,” he said.



Source link

LEAVE A REPLY

Please enter your comment!
Please enter your name here

Popular Articles