30 Indian Minds Leading the AI Revolution

Pranav Mistry: It's time to build value-driven AI models to transform industries, daily life

The CEO of TWO AI believes the focus has shifted from developing large language models (LLMs) to applying them to real-world problems

By Naandika Tripathi Forbes India Staff

Published: Jun 17, 2025 02:40:37 PM IST

Updated: Jun 17, 2025 02:46:05 PM IST

Full Bio

Pranav Mistry, CEO, TWO AI

In the heart of Silicon Valley, a Palanpuri technologist, entrepreneur and computer scientist from Gujarat has a straightforward message: There are enough foundational models, and it’s time to focus on creating value-driven, second-layer AI models that can transform both industries and everyday life.

The AI era has entered a new phase, says Pranav Mistry, CEO of TWO AI. The focus, he believes, has shifted from developing large language models (LLMs) to applying them to real-world problems. “We’ve moved beyond building foundational models. They’re now widely available and increasingly commoditised,” says Mistry. “Our focus is on what comes next: Compact, high-performance second-layer models distilled with deep domain expertise, models that understand the nuances of specific industries and context.”

Two AI’s approach reflects a broader industry shift toward second, third, and even higher-layer models—smaller, specialised systems trained with teacher-student distillation for enterprise and consumer use. The company initially trained its own foundational LLM called SUTRA, which supports over 50 languages, including Hindi, Gujarati, Korean, Arabic, Japanese and others, and which it claims outperforms some of the more prominent models in the industry, in multilingual capabilities.

Their goal is to grow into a platform that connects one billion AI users. If the first billion users were English-based, the next billion will come from non-English-speaking countries such as Asia and the Middle East, he says, adding, “We see tremendous potential in building AI that’s deeply contextual to local languages, workflows, and cultures.”

Four years since its launch, TWO AI has already reached over 500 million users through its enterprise customers that include leading telecom providers, prominent banks, and global retail companies. While tools like ChatSutra, its multilingual conversational AI, showcase its capabilities, the company’s deeper focus lies in building end-to-end AI platforms for enterprises and delivering real-world applications for consumers. TWO AI is also preparing to launch a free, consumer-facing, AI search engine—a lightweight, intuitive alternative to products like Perplexity, tailored for Indian users and vernacular discovery.

TWO AI’s key focus is on providing enterprise AI solutions that can be deployed securely and at scale. The company, rather than competing with prominent AI model companies like OpenAI or open-source models like DeepSeek, is leveraging these advances in the global AI race to its advantage in building real-world AI solutions and end-to-end platforms to meet the specific needs of enterprises.

View the full list here

“We’re building second-layer AI systems that are lean, smart, and customisable,” says Mistry. “We’ve trained domain-aligned distilled models that behave like high-performance experts—whether in healthcare diagnostics or retail forecasting, they are ready to deliver.”

Rather than building monolithic models with hundreds of billions of parameters, Two AI’s second-layer approach, distilling large models into smaller, specialised ones, is modular and efficient. Its domain-specific models are small enough to deploy securely on premises or in private cloud environments, while delivering accuracy and responsiveness.

Also read: Niki Parmar: Pushing the boundaries of AI

“The future of AI isn’t just bigger—it’s smarter and more targeted,” says Mistry. “We’ve developed the tools and infrastructure to build and serve these specialist models, and we have access to high-quality, often proprietary datasets that larger companies like OpenAI typically don’t.”

These types of applications are more valuable to enterprises than chatbots, which are only needed for customer support, he says. “That’s why we’re focused on solving real problems. And we’re supporting the developer and startup ecosystem too—we’ve already granted billions of free tokens to over 25 AI startups globally,” he adds.

TWO AI is also expanding its infrastructure to support its growing user base. The company has set up inference clusters in India, South Korea, and the US, with new clusters coming soon in Japan, Singapore, and Doha. These clusters will serve its SUTRA models. While model training currently happens in Silicon Valley, where its core AI and research teams are based, TWO AI is building out compute infrastructure and expanding its R&D team in India to train its next generation of models.

TWO AI had previously secured $20 million (`171 crore) in funding from Reliance Jio and Korea’s internet giant Naver Corp. The company is looking to raise more funds. On track to reach its goal of $20 million ARR by year-end, TWO AI aims to become profitable by 2026–2027.

“We are currently in the process of raising our growth round to expand our customer base and enter newer markets, including Japan, Southeast Asia, and the Middle East,” says Mistry.

He adds that TWO AI is experiencing significant growth in its customer base this year. “We currently have over 10 large enterprise customers globally, with an average annual contract value of $1 million.”

The company recently signed several large clients across Korea, Southeast Asia, and the Middle East, and has received approval for government AI projects in India focused on mission-critical applications.

Mistry was also recently appointed by the Ministry of Home Affairs as an advisor for cybersecurity at the National Forensic Sciences University (NFSU), where he will play a key role in strengthening national security. In this capacity, he will leverage cybersecurity and AI to protect India’s borders and critical digital infrastructure.

His background in AR/VR, AI, and robotics, and his time at IIT-Bombay, MIT, Microsoft, Google, and most recently as executive and president at Samsung has been instrumental in shaping his approach to innovation—whether it was AR glasses, smartwatches, foldable phones and now AI.

It’s now worth watching how Mistry will stand out among the big boys in Silicon Valley and create something of lasting value for India and beyond.

(This story appears in the 13 June, 2025 issue of Forbes India. To visit our Archives, click here.)