HMD Global, the Finnish phone maker trying to reclaim relevance after Nokia’s fall, just launched its first smartphone: the Vibe 2 5G. On paper, it’s unremarkable. Mid-range specs, a 6,000mAh battery, priced at ₹10,999 ($114). But what matters here isn’t the hardware. It’s what comes preloaded: Indus, an AI chatbot built by Indian startup Sarvam that speaks 22 Indian languages and understands mid-sentence code-switching between Hindi and English.
The partnership, first announced at India’s AI summit in February, feels like a quiet experiment that could say something interesting about how AI actually gets adopted in emerging markets. And right now, the signals are mixed.
When Distribution Matters More Than Downloads
Indus hasn’t exactly set the world on fire. Nearly three months after launch, the app crossed 293,000 downloads across all platforms in India. ChatGPT, by comparison, hit 43.9 million downloads in the same country. That gap is embarrassing, and it’s the reason why what HMD and Sarvam are doing matters.
The companies aren’t chasing the same users. They’re not trying to beat ChatGPT in the smartphone market where everyone already has access to every English-language AI tool available. Instead, they’re targeting what HMD’s CEO Ravi Kunwar called phase two in a TechCrunch interview: driving stickiness through distribution. By preloading Indus on affordable hardware, they’re placing the app directly in front of people who might never search for it themselves.
HMD’s smartphone presence in India is negligible. It doesn’t crack the top 15 according to IDC data. But the company holds 4% of India’s feature phone market, and that’s where the real play might unfold. Kunwar confirmed that a feature phone with Sarvam AI integration is coming in the coming months.
Feature phones. Not smartphones. That detail matters more than it sounds.
The Unsexy Distribution Channel That Actually Works
India still has hundreds of millions of feature phone users. These aren’t people choosing to stay offline. They’re people for whom affordable hardware is the only hardware available. English-language AI tools have no meaningful reach in this segment because the interface itself is a barrier. A chatbot that speaks your language, understands your dialect, and works on your device changes the equation entirely.
Bundling regional AI with feature phones is one of the most direct distribution strategies available in a market as linguistically diverse as India. It bypasses app stores, download decisions, and user awareness. It just shows up. Kunwar’s language in that interview is revealing: “Once they start using it, we will move to phase two to focus on driving more traction and stickiness. Right now, by pre-loading the app, we want to be more accessible to users.”
That’s not a product pitch. That’s a distribution play.
For investors and technology observers watching how AI adoption gets seeded in emerging markets, this is worth tracking. The Indus app currently doesn’t support offline usage, and it lacks device-level shortcuts to invoke the assistant. These are limitations that matter more on feature phones than smartphones. But they’re also solvable problems once usage data starts flowing.
Sarvam’s Real Bet Isn’t the Numbers Game
Sarvam AI itself is in an interesting position. The company has positioned itself as an enterprise player first, focusing on voice-based solutions for business use cases. The consumer play through Indus feels like an adjacent move, a way to build consumer awareness while the enterprise engine runs. Reports suggest a $300 million funding round at a $1.5 billion valuation is in the works, which would make Sarvam one of India’s most funded AI startups.
That kind of capital comes with expectations, and those expectations probably aren’t about competing with ChatGPT for downloads. They’re about proving that there’s a real market for localized AI in India, that it can be monetized, and that it can operate at scale. HMD’s distribution channel is a useful testbed for exactly that.
The early numbers aren’t impressive by Silicon Valley metrics. But Silicon Valley metrics don’t apply here. The question isn’t whether Indus can outcompete global AI products in their existing markets. The question is whether it can reach people that global products have completely failed to reach. That’s a different game entirely, and it’s one where feature phones and preloaded apps actually make sense.
