Forget Downloads: India's AI Success Demands Population Reach, Deep Engagement

The Indian market requires a nuanced approach to measuring the success of artificial intelligence ventures, moving beyond traditional metrics like download numbers, according to Hemant Mohapatra, a partner at Lightspeed Venture Partners. He highlighted the rapid growth of Sarvam AI, an Indian generative AI company, which reportedly grew 10x in the last month from an already significant base, as an example of how engagement and population-scale reach can be more telling indicators in India.[1] This perspective challenges conventional wisdom and underscores the unique characteristics of the Indian landscape for AI adoption.

Mohapatra's assertion that download numbers are not always "rooted in what works in India" points to a deeper understanding of the Indian consumer and the digital ecosystem.[1] India's vast linguistic diversity, varying levels of digital literacy, and a mobile-first internet population that doesn't always translate to high app engagement for every type of service necessitate a different yardstick for success.[2][3][4] For many AI applications, particularly those focused on voice-first interactions or catering to vernacular language users, metrics such as daily or monthly active users, the depth of engagement, and the ability to reach a wide and diverse population become far more critical than the sheer volume of application downloads.[2][5][6][4] The emphasis is shifting towards solutions that integrate seamlessly into existing user behaviors and platforms, often bypassing the need for a standalone application download altogether. This is particularly relevant in a country with over 700 million smartphone users and where voice search is seeing explosive growth, especially in regional languages and Tier-2/3 cities.[2][4] Google, for instance, reported a 270% year-over-year growth in voice searches in India, largely driven by non-English internet users.[4]

Sarvam AI's recent trajectory serves as a case in point for this evolving understanding. The company, backed by Lightspeed, Peak XV Partners, and Khosla Ventures, has been focusing on building foundational AI models specifically for India, emphasizing support for diverse Indian languages and voice-first interfaces.[7][5][8] Its reported 10x growth in a single month, coupled with running one of India's largest population-scale AI campaigns that reached over 4% of the adult population, signifies a significant penetration beyond what download numbers alone could convey.[1] This achievement is particularly noteworthy given Sarvam AI's mission to develop a "full stack" for generative AI, from custom model training to enterprise-grade deployment platforms.[7][8] The company has also been selected by the Government of India under the IndiaAI Mission to build the nation's sovereign Large Language Model (LLM), a project that will involve dedicated compute resources to create an indigenous foundational model fluent in Indian languages and designed for secure, population-scale deployment.[9][10] This LLM is envisioned to be capable of reasoning and voice-first interactions, addressing the unique needs of the Indian populace.[9][11][10] Sarvam AI's models aim to be competitive with global counterparts and are designed for cost-effectiveness and scalability within India.[9] The company's approach involves combining model innovation with application development to build solutions tailored for India's vast population.[7][5][8]

The implications of this perspective for the broader AI industry in India are profound. It signals a maturation of the market, where investors and companies are looking beyond superficial metrics to gauge genuine traction and impact. For AI companies targeting India, this means a strategic pivot towards developing solutions that are deeply integrated with local languages, cultural contexts, and user behaviors.[12][5] The success of platforms like UPI (Unified Payments Interface) and Aadhaar, which serve over a billion people, has demonstrated India's capacity to build and adopt population-scale digital public goods, setting a precedent for AI systems.[13] There's a growing recognition that India needs a unique approach to AI, tailored to its specific needs and scale.[12] This includes developing AI models that support a diverse set of Indian languages and voice-first interfaces, and co-building domain-specific AI models with Indian enterprises using their data.[5] The focus is increasingly on creating AI that can add tangible value to people's lives by being accessible and relevant to a wide spectrum of the population, including those in rural areas or those with literacy barriers.[7][4] The Indian government's push for AI, including initiatives like the IndiaAI Mission with a budget of ₹10,372 crore (about $1.25 billion), further aims to democratize AI innovation and access.[11]

Lightspeed Venture Partners, where Mohapatra is a partner, has been actively investing in the Indian AI ecosystem, recognizing its potential.[14][15][16] The firm's strategy often involves backing companies that are building for India or building in India for the world, with a focus on technology-led businesses.[15] Globally, Lightspeed has invested significantly in AI, with nearly $1 billion deployed in the 18 months leading up to late 2024 and over $2 billion in more than 100 companies in the past year.[17][15][16] In India, their portfolio includes companies like Sarvam AI, Stimuler (a voice-first AI tutor), Thena, Portkey, Marqo, and Gushwork, reflecting a belief in the application layer of AI and founders who understand the specific problem statements they are solving.[14][15] Mohapatra himself has emphasized the need for India to move from being an importer of AI breakthroughs to creating them, advocating for research-based engineering to build genuine, long-term innovation.[18][19] This involves not just technological prowess but also strong go-to-market strategies, particularly for companies with global aspirations.[18][20] The firm is increasingly seeing founders with clarity on their role in the global landscape and their unique advantages in the application layer.[14]

In conclusion, the narrative championed by investors like Hemant Mohapatra, and exemplified by the progress of companies like Sarvam AI, signals a critical shift in how AI success is defined and pursued in India. The emphasis is moving away from vanity metrics like downloads towards more substantive indicators such as deep user engagement, broad population reach, and the ability to deliver solutions tailored to India's unique linguistic and socio-economic fabric. This approach recognizes that for AI to truly take root and flourish in India, it must be inclusive, accessible, and capable of addressing the specific needs of its diverse billion-plus population, ultimately fostering an ecosystem where India is not just an adopter but an active creator and shaper of AI technology.[12][9][13] The substantial growth of the conversational AI market in India, projected to expand significantly, and the increasing adoption of voice recognition technologies further underscore the importance of these tailored approaches.[2][21][22]