• Artham, a 7-billion-parameter Small Language Model (SLM) trained exclusively on Indian Financial & Capital Markets, was unveiled first time at AWS re:Invent 2025
  • Built by Raise Financial Services and hosted in India, Artham already powers fuzz (https://askfuzz.ai/) and ScanX (https://scanx.trade/) with source-backed insights, real-time news and contextual understanding for Long-Term Investors, Finance and Capital Market Professionals
  • Developed over 9 months by Raise AI, Artham uses a curated 70:30 blend of public and proprietary Indian financial data and supports native tool-calling for secure access to real-time data and analytics

MUMBAI : At its flagship event – AWS re:Invent 2025 conference currently underway in Las Vegas, Amazon Web Services (AWS) unveiled Artham – India’s first Small Language Model (SLM) trained exclusively on Indian Financial & Capital Markets and built by Raise Financial Services. Raise is the company owning and operating India’s fastest-growing stock trading platform Dhan, and AI-powered financial markets research assistant fuzz.

The announcement was made in a keynote address by Dr. Swami Sivasubramanian, Vice President, AWS Agentic AI, with Raise being showcased as a standout example of how forward-thinking customers are building domain-specific AI products on AWS. This event marks a rare recognition and milestone for Raise Financial and the broader Indian fintech ecosystem for an Indian startup to feature on re:Invest keynote stage on the global platform. Earlier, AWS CEO Matt Garman in his keynote address introduced DevOps Agent (https://aws.amazon.com/devops-agent/) recognising Dhan as an early customer and adopter of this platform globally.

Positioned as “The Financial Brain, Built for India,” Artham is India’s first dedicated Small Language Model (SLM) for capital markets. It is designed to understand the structure, vocabulary and regulations of India’s capital markets, and to power intelligent, context-aware financial experiences for serious investors, traders, capital market enthusiasts as well as  financial professionals.

“AI in financial services needs to be precise, explainable and deeply contextual to local markets. With Artham, we are building an India-first financial AI layer that strengthens fuzz and ScanX for serious investors and market participants, while staying fully aligned with the responsibilities of operating regulated platforms like Dhan as part of Raise Ecosystem.” said Pravin Jadhav, Founder & CEO, Raise Financial Services.

Built entirely in India and hosted on cloud infrastructure located within the country, Artham has been engineered to build causal relationships between events spanning Indian companies, macro developments and their impact on the financial landscape, specifically to help connect the dots between India’s vast financial data points. It is tuned on Indian market data, regulatory content, company disclosures and financial knowledge focused on India’s listed universe and its ecosystem of participants, rather than generic global internet text.

Artham is a 7-billion-parameter model that has been in development for around nine months, led by Raise AI. The training corpus combines both public and proprietary data from the Indian financial ecosystem; roughly 70% is public data and 30% is proprietary. Over time, the share of proprietary data is expected to increase as Raise deepens Artham’s coverage. Artham is also built with native tool-calling capabilities, allowing it to securely invoke internal data services and workflows, such as real-time market data, analytics and retrieval systems, when responding to user queries.

“We wanted a small, efficient model that we could tune deeply on Indian financial data, evaluate rigorously and run with strong controls around data sovereignty,” said Alok Pandey, Co-founder and CTO, Raise Financial Services. “Working with a lean internal team at Raise AI has helped us move from experimentation to production-grade financial AI in a very short time, with Artham now quietly powering experiences across fuzz, ScanX and Dhan.”

The model is already live in production across the Raise ecosystem. A share of real-time financial news on ScanX and Dhan is generated using Artham, and some conversations on fuzz are also powered by the model, helping users uncover insights on Indian markets faster, with answers grounded in facts and backed by verifiable insights in the form of source links, official announcements and references.

Every experience powered by Artham is built with clear guardrails: it offers information and educational insights, not investment advice or recommendations. Raise and its products do not provide investment advice, and users remain fully responsible for their own investment decisions.

Artham is a foundational capability for Raise AI’s future roadmap, enabling new India-first tools across research, education and analytics for market participants and finance professionals.