India AI Impact Summit Insights
All sessions

Designing India’s Digital Future: AI at the Core, 6G at the Edge

Contents

Executive Summary

This keynote and panel discussion from an AI summit focuses on India's strategic opportunity to transition from a consumer of global technology to a shaper of 6G and AI standards. The core argument is that 6G must be fundamentally AI-native from inception (unlike 5G, where AI was retrofitted), and India must build a sovereign, end-to-end AI and 6G ecosystem to democratize intelligence across 1.4 billion citizens while reducing dependency on foreign technology platforms.

Key Takeaways

  1. India is at a historic inflection point to shift from technology consumer to AI+6G innovator: Government support (100 5G labs upgrading to 6G, 6G Accelerated Research Program, reduced 3GPP membership costs) creates real pathways for startups and academia to shape global standards—if execution happens in next 2-4 quarters.

  2. Energy efficiency and affordability are non-negotiable constraints, not optional features: The "democratize intelligence to 1.4 billion citizens" mandate means distributed, edge-first architectures and sovereign models are existential requirements, not luxury choices. Rental models are economically unfeasible at India's scale.

  3. Open APIs + sovereign control is the viable path forward: Success depends on hybrid models—keeping core stacks sovereign while using open, standardized interfaces (like UPI) to enable third-party innovation. Proprietary silos (e.g., Meta glasses locked to Meta apps) are antithetical to the vision.

  4. India-specific data and models matter urgently: Generic LLMs trained on non-Indian data introduce cultural and contextual bias. India's advantage is local language models, agriculture/informal worker-specific models, and use cases tuned to Indian driving patterns, regulations, and economic conditions.

  5. The real value lies in coordinated ecosystem thinking, not point solutions: Individual 6G or AI projects will fail without alignment to standards, data exchange frameworks, and national sandbox policies. The next 12-18 months are critical to move from 100+ isolated pilots to a coordinated, scalable intelligence platform.

Key Topics Covered

  • 6G as AI-native technology: Historical progression from 2G-5G (connectivity-focused) to 6G (intelligence-integrated from design phase)
  • India's policy and ecosystem initiatives: Government programs including 6G Accelerated Research Program, 5G labs, TSDSI support, and Bharati 6G Alliance
  • Network evolution and AI integration: How AI transforms RAN (radio access networks), from device-level to edge to cloud inferencing
  • Uplink traffic transformation: Shift from downlink-heavy (10:1 or 12:1 ratio) to more balanced (4:1) due to AI applications
  • Data as strategic fuel: National data exchanges and frameworks for training India-specific AI models
  • Enterprise value pools: Demand analysis, workflow automation, and security through 6G+AI convergence
  • Sovereignity and open ecosystems: Balancing sovereign AI stacks with open, interoperable APIs and standards participation
  • Use cases across sectors: BFSI, manufacturing, healthcare, mobility, agriculture, informal economy
  • Energy efficiency and cost: Distributed inferencing to reduce power concentration; democratizing AI through edge compute at cell towers
  • Standardization participation: India's shift from passive consumer to active 3GPP standard-setter via startups and research projects

Key Points & Insights

  1. AI integration is now native to 6G standards (ITU framework released 2 years ago), unlike 5G where it was retrofitted—this represents a fundamental mindset shift in how networks are designed from inception.

  2. India's policy support is enabling ecosystem participation: Department of Telecom reduces 3GPP membership costs from ₹5-6 lakhs to ₹10,000 for startups, democratizing access to standard-setting bodies and shifting India from passive consumer to active shaper of 6G standards.

  3. AI-driven traffic will grow from ~5% today to 30% by 2033 (Nokia Bell Labs projection), with fundamental changes to traffic patterns—uplink will shift from 10-12:1 downlink dominance to ~4:1, requiring massive network redesign for persistent, bursty, multimodal agent traffic.

  4. Distributed inferencing is essential for infrastructure viability: Centralizing all AI workloads in data centers is unsustainable due to power consumption; edge inferencing (latency-critical: robotic surgery, autonomous vehicles, industrial robotics) must be tiered across devices, edge, and cloud based on use case requirements and cost.

  5. Token economy and latency are the new network KPIs: AI efficiency is no longer just about speed; reducing latency by 10-20% improves productivity/efficiency drastically. Token consumption rate, latency, and coverage must jointly optimize for different use case classes.

  6. Sovereignity must be end-to-end but selective: India needs sovereign control of the entire value chain (device → cloud → edge → inference) to avoid rental dependency, but open ecosystems and APIs (exemplified by UPI success) are essential for interoperability, scale, and collaboration.

  7. India's competitive advantage is scale + data diversity: With 490 million informal workers and billions of users, India can train models at massive scale and lower costs than others; cultural and linguistic diversity (multiple language models needed) creates unique training opportunities not available elsewhere.

  8. National frameworks and data exchanges are critical infrastructure: Centralized data anonymization platforms and shared training exchanges enable cross-sector AI model development while protecting confidentiality—essential for democratizing AI without siloing data within industries.

  9. 6G enables last-mile enterprise transformation in a wireless-first economy: India lacks fiber penetration; 6G+AI enables enterprises to reach informal workers, rural areas, and unconnected populations through wireless—unlocking productivity gains in agriculture, trades, and informal economy.

  10. Standardization participation + sandbox testing + safety guardrails must advance in parallel: Pilots risk becoming siloed if not aligned to evolving 3GPP standards; national sandboxes, audit frameworks for AI-driven network parameter changes, and explainability requirements are necessary for safe, coordinated deployment.

Notable Quotes or Statements

  • Ashok Kumar (Deputy DG, Department of Telecom): "AI artificial intelligence AI began to be integrated as part of the network functions... but now [in 6G] the artificial intelligence is part of the initial thought itself... ubiquitous intelligence."

  • Ashok Kumar: "It's an opportunity not only to participate in the standard so that our technology our innovations becomes part of the standard but also to build our own end-to-end 6G technology stack."

  • Rajiv Saluja (VP 5G Radio, Reliance Jio): "India cannot afford to rent intelligence. We need to build it. We need to scale it... the complete infrastructure that we are building is from connectivity to the cloud to the edge and then the intelligence ecosystem on top."

  • Surrojit Roy (Nokia India): "With AI you can actually decipher [signals in high-noise environments]... capacity increase maybe 25-30%... this is going to increase the capacity of the network" (on Deep RX/DeepTX).

  • Sundeep Sharma (Tech Mahindra): "The problem is with the pilot and the scale gap is not a technology gap—it's a gap of how we put things together in frameworks which are scalable and referenceable."

  • Sundeep Sharma: "Data is like bread and butter for AI—there's no AI if there's no data... national framework of putting the data together, creating national exchanges where data can come in... this gives a very useful reference."

  • Rajiv Saluja (on sovereignty): "We need to have a sovereign AI ecosystem... this entire value chain has to be made in India... we don't have an option in this."

  • Surrojit Roy (on inclusive scale): "There are approximately 490 million informal users—carpenters, drivers... AI use cases can significantly help out here... but models need to be trained on data coming from India because if trained outside, there will be bias."

Speakers & Organizations Mentioned

Government:

  • Ashok Kumar, Deputy Director General, Department of Telecom (DoT), Government of India
  • Ministry of IT (MIT), India
  • Department of Science & Technology (DST), India
  • ARIB (assumed to be Indian standards body)
  • NITI Aayog (mentioned in context of 30 trillion economy, 2047 vision)

Private Sector / Industry:

  • Reliance Jio (Rajiv Saluja, VP 5G Radio)
  • Nokia India (Surrojit Roy, Senior Telecom Leader, Head of Technology)
  • Tech Mahindra (Sundeep Sharma, VP & Global Head of Emerging Technologies)
  • TCS (Radha, Moderator; Head of Technology Engineering & Innovation)
  • Airtel (mentioned as partner in network API initiatives)
  • AT&T (Siddhu, attendee question)
  • Meta (mentioned in context of proprietary ecosystems)

Standards & Alliances:

  • ITU (International Telecommunication Union) — 5G IMT-2020 framework, 6G framework
  • 3GPP (Third Generation Partnership Project)
  • TSDSI (Telecom Standards Development Society of India)
  • Bharati 6G Alliance / Bharat 6G Alliance (multiple working groups on technology, spectrum, devices)

Research & Testing:

  • Nokia Bell Labs (traffic projections)
  • 5G Labs (100 across India, being upgraded to 6G labs)
  • Terahertz test bed
  • AOC test bed
  • DSIR (Department of Science & Industrial Research, implied via DST)

Technical Concepts & Resources

6G Standards & Frameworks:

  • ITU 6G Framework: Six usage scenarios including "integrated artificial intelligence and communications"; four overarching design principles including "ubiquitous intelligence"
  • 3GPP Release Roadmap: Release 15 (5G start, three-part delivery) → Release 18 (5G Advanced with AI integration) → Release 19, 20, 21 (path to first 6G release)
  • Voice over NR (VoNR): Network slicing evolution from 5G to 6G

AI in Networks:

  • Deep RX / DeepTX: Nokia's deep learning approach to optimize transmitter/receiver signal decipher in high-noise environments; demonstrated ~25-30% capacity gains
  • Token Economy: AI efficiency metric tracking token consumption rate, latency, cost—replacing traditional network KPIs (throughput, latency in isolation)
  • Multimodal Agents: AI agents orchestrating end-to-end workflows, initiating Agent-to-Agent (A2A) traffic; shift from user-initiated to agent-initiated traffic
  • Semantic Communications: Mentioned as key 6G research area

Network Architecture:

  • Network Slicing: Technology for creating logical network partitions; advancing from 5G (limited) to 6G (ubiquitous)
  • Edge Inferencing vs. Cloud Inferencing: Tiered distribution based on latency (robotic surgery, autonomous vehicles, industrial robots require edge; normal consumer use cases can use central cloud)
  • Uplink-Downlink Ratio Shift: Current ~10:1 or 12:1 (downlink-heavy) → Future ~4:1 (balanced) due to AI contextual data transmission needs
  • Spectral Efficiency: 5G Advanced: 400 MHz bandwidth target (vs. current ~100 MHz); 6G: 5× spectral efficiency improvement expected = ~20× capacity gains

Spectrum & Frequency:

  • Terahertz (THz) Technology: High-frequency research for 6G; terahertz test bed operational in India

Data & Models:

  • National Data Exchange: Proposed centralized anonymization + training platform for cross-sector AI model development
  • Language Models: India-specific LLMs for multiple Indian languages to address bias and cultural relevance
  • Model Audit & Explainability: Safety framework for AI-driven network parameter changes; intervention policies needed

Initiatives & Programs:

  • 6G Accelerated Research Program: Indian government scheme; 100+ projects selected across terahertz, AI/ML, semantic communications, sensing
  • 100 5G Labs Initiative: 100 labs across India (announced in budget, inaugurated by PM); operational and being upgraded to support 6G research
  • RDI Scheme (DST): Research-to-Development-to-Industrialization framework; telecom sector newly included to support scale-up
  • Cyber Physical Systems Programs: DST support for 5G/6G projects
  • 5G/6G Sandbox Testing: National frameworks for pilot validation against evolving standards
  • Bharati 6G Alliance: Multiple working groups (technology, spectrum, devices) guiding government policy

Use Case Categories (by Latency/Coverage Sensitivity):

  • Ultra-low latency (1-10ms): Robotic surgery, autonomous vehicles, industrial robots, real-time AR/VR
  • Persistent uplink coverage: Asset tracking, environmental monitoring, field operations
  • High-volume throughput: Standard consumer apps, content streaming (less impacted by 6G shifts)

Sovereignty & Openness Concepts:

  • Token Sovereignty: Control over entire value chain from request initiation through inference delivery
  • End-to-End Sovereign Ecosystem: Device → edge → core → cloud → intelligence layers all designed/built in India
  • Open APIs / Loose Coupling: Necessary for interoperability and third-party innovation (contrasted with proprietary ecosystems like Meta glasses)
  • UPI Model: Cited as successful precedent for open ecosystem driving scale

Transcript Quality Note: The transcript contains some audio artifacts (repeated phrases, incomplete sentences, background chatter) typical of live speech-to-text conversion. Summaries have been reconstructed from context and speaker intent.