Malaysia-based Athenatech AI has announced that it will be spinning off its AxxonAI platform into a new subsidiary that focuses on delivering LLM-driven synthetic data generation for helping enterprises work around two of the biggest problems in AI development today – the lack of usable local data and increasingly strict data privacy regulations.
According to founder and CEO Sonny (Soumitra) Dey, the philosophy behind AxxonAI is about proving that innovation and ethics don’t have to be mutually exclusive. He says the platform is built specifically for Asia and ASEAN, with localisation and personalised synthetic data as key differentiators, while hinting that upcoming versions will integrate agentic functions to further enhance decision-making and automation.
AxxonAI’s goal closely aligns with Malaysia’s broader ambitions to become an AI Nation by 2030 under the National AI Action Plan 2026–2030, and through working in step with priorities set by the National AI Office, the company will be setting up a Global Delivery Centre focused on AI training, synthetic data, and generative AI insights.
Specifically, AxxonAI is a privacy-centred, multi-tier synthetic data cognitive intelligence platform built for enterprise use that doesn’t rely on sensitive, true customer data, and instead, utilizes techinques like Generative Adversarial Networks and Retrieval-Augmented Generation with LLMs to create statistically equivalent, privacy-safe datasets.
These datasets can then be fine-tuned to match local markets, regulatory requirements, and specific business objectives, something that global datasets often fail to do. The platform also introduces data decoupling and fine-tuning as key features, giving organisations more control over data quality and utility, particularly for segmentation and hyper-personalisation in complex BANI consumer environments.
This regional focus also addresses a heavily weighted pain point for ASEAN enterprises, where localised and structured data is often scarce, global datasets are expensive, and privacy laws such as PDPA and GDPR limit how real customer data can be used. AxxonAI’s approach aims to reduce compliance risk, improve accuracy, and remain affordable enough for SMEs, helped by a localised LLM that understands regional languages and business practices, alongside a built-in Quality Score feature.
In terms of industrial applications, healthcare will be one of the first sectors to see this platform in action. Through a partnership with Sancy Berhad, AxxonAI is enabling clinicians and administrators to extract actionable insights from patient data without exposing personal identities. The collaboration will initially focus on Connected Care, a healthcare tourism solution, and Ring, a patient-centric care model that uses synthetic data and predictive intelligence to improve outcomes, reduce costs, and streamline care from hospital to home.
Looking ahead, AxxonAI is already developing multiple versions of the platform, including Small Language Model-based implementations and SME-focused agentic integrations. These domain-specific solutions are expected to roll out across industries such as agriculture, automotive, BFSI, telecommunications, logistics, and media and entertainment by early 2026, signalling that this Malaysian-built platform has ambitions well beyond its home market.
