New Delhi [India], April 13 : AvenuesAI through its subsidiary PhroneticAI has started working on a fully on-premise, full-stack AI models that never leave the client's infrastructure, positioning itself for a growing market where data sovereignty and geopolitical risk are reshaping enterprise AI adoption.
BSE and NSE listed AI Fintech Company, AvenuesAI Limited is making a bold strategic foray into the turnkey AI technology solutions space targeting an emerging and largely untapped market for enterprise-grade, fully private generative AI infrastructure with small language models (SLMs).
The company through its subsidiary PhroneticAI, an artificial intelligence research and development is developing an AI product for corporates, organisations and large institutions offering a PrivateGPT, an on-premise SLM platform built and deployed entirely within the client's own walls, with no data ever leaving the premises.
At a time when data privacy concerns are reshaping the way organisations engage with artificial intelligence, AvenuesAI is betting that the next competitive frontier in enterprise AI will not be who has the most powerful model but who can guarantee that sensitive data never touches a third-party server.
"There is growing, serious concern over data security and privacy not just for the general public, but for corporates who are already processing sensitive data through open-source AI tools," said Vishal Mehta, Chairman and Managing Director, AvenuesAI Limited.
He further pointed to two seismic developments that have accelerated client interest. Separate revelations about the dealings of OpenAI developer of ChatGPT and Anthropic developer of Claude with the US government's Pentagon have publicly raised data privacy and data independence concerns across the world in a way that no internal compliance memo ever could.
The calculus is especially pointing out and drawing a clear line from geopolitical risk to boardroom anxiety amongst the household brands to mid-size to large conglomerates and organisations that have accumulated decades of proprietary business intelligence, trade secrets, and strategic data. "These are not just data security questions. They are questions of strategic autonomy and sovereign decision-making," points out Mehta.
The point was underscored by a precedent from mid-2025 event, when Microsoft suspended cloud and email services for India's Nayara Energy a major petroleum refiner with Russian ownership links citing compliance with European Union sanctions. Nayara challenged the abrupt shutdown in the Delhi High Court, calling it unjustified. Services were restored within days and the case was resolved, but the episode left a lasting impression on risk managers across India's corporate sector: even paid, contracted services from global technology giants can be switched off overnight for reasons entirely outside a client's control.
According to the company, AvenuesAI is staking its competitive position not on access to artificial intelligence, but on the depth of what it actually builds. Unlike conventional AI consulting firms that license pre-built models or layer thin wrappers over existing tools, the company is delivering a fully customised, centralised platform that spans the entire lifecycle of a Small Language Model from pre-training on a client's proprietary data corpus, through a purpose-built distillation pipeline, to on-premise deployment complete with inference APIs, monitoring dashboards, and a continuous retraining loop. Crucially, the smaller footprint of SLMs of 1B to 10B parameter and domain specific AI makes genuine on-premise deployment long a theoretical selling point in enterprise AI a practical reality, opening the market beyond hyperscaler-dependent large enterprises to mid-market clients who require data sovereignty without sacrificing capability.
The private on-premise SLM platform, will be trained in-house and calibrated to the client's specific use cases whether that entails analysing confidential documents, PDFs, spreadsheets, audio, video, or image data with full model ownership residing with the client at all times.
"Once the private, on-premise SLM platform is built and deployed, we are completely out of the system and clients have the option to operate independently" Mr Mehta said an unusually firm commitment in an industry where software vendors routinely retain backend access for maintenance, updates, and, critics allege, data observation. His statement encapsulates company's core pitch: not a service subscription, but a permanent handover of sovereign AI capability, unless, it's called for service.
The company believes it is moving into a market with enormous latent demand and vanishingly few credible domestic competitors. "This is an unmet privacy and secrecy need coming from large clients," Mehta said.
Mehta is candid about the broader trust deficit that AvenuesAI's subsidiary PhroneticAI hopes to capitalise on. "Large corporates and organisations will always be wary of allowing major global AI companies' access to their data irrespective of any ethical, moral, or legal guarantees those companies may offer," he observed. In an era where artificial intelligence is simultaneously the most powerful productivity tool and the most intimate data-handling system an organisation can deploy, that wariness may prove to be PhroneticAI's most durable competitive advantage.