Reclaiming AI Infrastructure
Event Highlights
Kalpa Impact in collaboration with Current AI and Bhashini hosted a workshop and live demo of a prototype at the India AI Impact Summit 2026
The invite-based event, titled “Reclaiming AI Infrastructure: Why AI hardware needs to be open, personal, and multilingual”, held on 17 February at Le Meridien New Delhi, was an exclusive gathering that convened 65 industry leaders across the funding and tech ecosystem.
It offered a sneak peek of Current AI and Bhashini’s first collaborative build: an open source, multilingual, privacy preserving, offline, handheld prototype device. The event provided technology leaders, civil society, engineers, and maker communities with the opportunity to query and experience the device first hand.
Following a technical demonstration by Andrew Tergis – Engineer at Current AI, and Shailendra Pal Singh – Senior General Manager, Bhashini, participants were invited to test the boundaries of their imagination through a workshop to conceptualise unique use cases for the prototype. They were encouraged to think of applications that spark creativity in their own lives and communities, and to consider how and where the demonstrated prototype might benefit those whose voices and needs are rarely centred in big tech product design and deployment.
Remarks from Ayah Bdeir – CEO, Current AI, and Martin Tisné – Founder and Chair of the board, Current AI, anchored the evening, raising critical questions around who gets to design and develop AI products? What does it mean for you, and I, and for local communities to really have a voice in what AI does, and how it is put into our hands?
Sushant Kumar – Founder and CEO of Kalpa Impact, offered a new perspective: smaller, cheaper, privacy by design, open source, purpose built AI models designed for Global South contexts, in local languages and with real world constraints in mind, including AI that can function completely offline.
Across the event, the one recurring message was that the future of AI must be shaped by a wider set of voices, experiences, and contexts. Building AI that is personal, local, and multilingual should be a collective responsibility shared by governments, builders, and communities alike.