This keynote by Pramod Varma explores the necessary evolution of Digital Public Infrastructure (DPI) beyond its initial foundational layers. It argues that to achieve global scale, inclusion, and unlock exponential economic value, DPI must shift from purpose-specific, often siloed systems towards universal, user-centric, protocol-based infrastructure, akin to the internet itself. Viewers will learn about the limitations of current approaches, the importance of user control over data and assets, and emerging concepts like Finternet, open networks, tokenization, and the role of AI in shaping the future of DPI.
Synthesized Summary
Pramod Varma posits that while foundational DPI blocks like identity and payments have made significant strides in formalization and enabling frictionless transactions over the last decade, the current approach to building digital infrastructure often remains fragmented and system-centric. He argues that true DPI is defined by its reusability and ability to serve as general-purpose building blocks, enabling a diverse ecosystem to innovate on top, much like the internet protocol (TCP/IP) enabled the web. The prevailing model of system-to-system data sharing is inherently limited, costly, creates dependencies, and hinders the scalability required for global and cross-border applications.
The future of DPI, Varma suggests, lies in a fundamental shift towards user-centricity, particularly concerning data and digital assets. Individuals and businesses must gain control over their data (identity information, credentials, transaction history, entitlements, etc.), allowing them to share it securely and with consent via mechanisms like digital wallets and verifiable credentials. This user-mediated flow is crucial for unlocking data as a powerful economic asset. Concepts like “Finternet” aim to generalize financial infrastructure beyond specific use cases (payments, lending) into a universal, protocol-based layer. Similarly, open networks for commerce (e.g., ONDC using Beckn protocol), energy (e.g., Uei), agriculture, and skills are emerging, designed to be decentralized and interoperable, breaking down platform monopolies. AI is positioned to supercharge DPI adoption by overcoming language and interface barriers, requiring democratized access to compute, data, and models (both large and small, specialized). To accelerate progress globally, Varma advocates for thinking in terms of universal protocols and standards and exploring rapid deployment models like DPI as a Service (DaaS), moving away from the slow, costly process of building bespoke systems country by country.
Key Learnings & Recommendations
- DPI Defined by Reusability: The core principle distinguishing DPI from mere digitization is its design as reusable, general-purpose infrastructure components (building blocks) that others can leverage. [04:17], [04:55]
- Three Layers of DPI: Effective DPI comprises Open Tech Standards/Building Blocks, Participatory Governance frameworks, and enables Diverse/Inclusive Solutions built atop. [04:25]
- User-Centric Data Sharing is Essential: The current system-to-system data sharing model is flawed and limits scale. A user-centric model, where individuals control the flow of their verifiable data via wallets/credentials, is necessary for unlocking data’s economic potential and enabling cross-border use. [09:37], [11:13]
- Data as an Economic Asset: Personal data (identity, certificates, affiliations, transactions) is a powerful economic asset. Free, user-controlled flow of trusted, verifiable data is key to unlocking growth. [12:36]
- Digital Assets & Tokenization: Real-world and digital assets (money, data, credentials, property, IP) can be tokenized, restoring user control and enabling programmable transactability. Ledgers are crucial for assets requiring “singleness” (non-duplicability). [11:18], [20:01], [25:02]
- Protocols Over Platforms: To achieve universal, interoperable infrastructure like the internet or GSM, DPI must be built on open protocols and standards, not closed, siloed platforms that tend towards monopolies. [17:31], [33:01]
- Emergence of Open Networks: Decentralized, open networks are being built for various sectors (commerce, energy, agriculture, skills) using common protocols (e.g., Beckn) to foster competition and innovation beyond single platforms. [27:34], [28:35]
- Finternet Vision: A future financial system (“Finternet”) based on universal, protocol-based infrastructure, generalizing specific functions like payments, lending, and asset management. [14:06], [20:41]
- AI for Inclusion: AI can significantly enhance DPI adoption by overcoming language and interface barriers, making digital transactions accessible to the next billion users. This requires democratizing compute, data, and models (including smaller, specialized ones). [29:30], [31:06]
- Need for Accelerated Deployment: The current pace of building DPI country-by-country, system-by-system is too slow. Faster deployment mechanisms, potentially “DPI as a Service” (DaaS), are needed. [32:04], [32:52]
Key Visual Information
- DPI Layers Diagram [04:17]: Illustrates the three interdependent layers: Technology Standards & Building Blocks (base), Participatory Governance (middle), and Diverse & Inclusive Solutions (top). Reinforces the idea of building solutions on top of foundational layers.
- User-Centric Data Sharing Model [11:13, 18:02]: Contrasts the limited system-to-system data sharing model with a user-centric model where the user/owner holds credentials/data (in a wallet) and mediates sharing between issuers and service providers upon request and consent.
- Digital Identity/Assets/Transactions [09:38]: Shows the user at the center, interacting with Digital Identities (“Who I Am”), Digital Assets (“What I Have” - including data, money, credentials), and Digital Transactions (“What I Do” - payments, data/credential exchange). Highlights data as a key asset class. [12:36]
- Finternet Concept [20:41]: Depicts “Finternet” as a universal infrastructure layer for finance, aiming to generalize various purpose-specific financial functions (payments, lending, tokenization, etc.).
- Technology Evolution Graphic [21:27]: Shows how purpose-specific technologies (like separate cameras/microphones or networking stacks) tend to collapse into purpose-generic, programmable layers (like smartphones or TCP/IP), unlocking exponential value. This pattern is applied to the Finternet concept.
- Siloed Platforms vs. Open Networks [27:01, 27:18]: Illustrates the shift from multiple closed platforms (creating monopolies) to an interoperable, open-loop network where multiple players can connect and transact, similar to the internet.
- Open Network Examples [27:34, 28:05, 28:35]: Shows logos and brief descriptions of ONDC (Open Commerce Network in India), Uei (Unified Energy Interface in India), and emerging networks in Kenya, Gambia, and Brazil, demonstrating the application of open network principles across different domains.
- AI for Inclusion [29:30]: Features an image of a person using voice commands on a smartphone, illustrating how AI (especially voice interfaces) can overcome literacy and digital literacy barriers to enable participation in the transaction economy.
- DPI Approach to AI [31:00]: Lists key components for leveraging AI with DPI: Democratized compute, Open data repositories, Open Models, and Smaller, use-case-first LLMs.
- DaaS Concept [32:52]: Presents “DPI as a Packaged Solution” (DaaS) as a mechanism for rapid deployment, contrasting the slow “100+ years” timeline of the current approach. Includes a QR code for a related initiative.
- Universal Infrastructure [33:01]: Depicts a network graph emphasizing the need to think in terms of protocols, standards, and the internet model for building truly universal DPI.
Key Questions Addressed or Raised
- Addressed:
- What is the future evolution of DPI beyond foundational identity and payments?
- How can DPI move from siloed systems to interoperable, universal infrastructure?
- How can user control over data be implemented at scale (user-centric data sharing)?
- What role does tokenization play in the future of digital assets?
- How can open network principles be applied to sectors like commerce and energy?
- What is the “Finternet” concept?
- How can AI enhance DPI adoption and inclusion?
- Raised:
- How can we accelerate the global deployment of DPI? (Addressed by proposing DaaS).
- How do we ensure data sharing infrastructure is user-centric, not system-centric? [Implicit challenge]
- How do we move markets from fragmented micro-economies to unified mega-economies using DPI? [Implicit challenge]
- Call for input on working papers related to user-centric data sharing and DaaS. [20:07], [32:52]
Stated or Implied Applications
- Identity: Formalization, KYC, Authentication across borders and sectors. [06:59]
- Payments: Frictionless P2P, P2M, G2P, cross-border payments. [07:44]
- Data Sharing: Open Finance, Open Health, secure credential exchange (education, health, skills). [08:35], [11:13]
- Asset Management: Tokenization and programmable transactability of real-world assets (art, property, vehicles) and digital assets (data, content, IP). [20:01], [25:02]
- Finance: Streamlined sustainability/green financing, carbon/tax credit markets (“Finternet”). [25:49]
- Commerce: Decentralized e-commerce, logistics, and mobility networks (Open Commerce). [26:30], [27:34]
- Energy: Interoperable EV charging, P2P energy trading, microgrid management (Open Energy). [28:05]
- Agriculture: Empowering farmers through learning and agri-services networks. [28:35]
- Skills/Livelihood: Networks for livelihood empowerment, skills training, employment opportunities. [28:35]
- Inclusion: Using AI to overcome language/interface barriers for accessing digital services. [29:30]
Key Terminology Defined
- DPI (Digital Public Infrastructure): Reusable, shared digital systems (tech standards, governance, solutions) enabling diverse applications, analogous to physical infrastructure like roads. [03:14], [04:17]
- Reusability: The core principle of DPI, meaning components can be used by various actors to build different solutions. [04:55]
- User-Centric Data Sharing: A model where individuals control their data (held as verifiable credentials, often in a wallet) and consent to its sharing, shifting control away from systems. [11:13]
- Digital Assets: Includes digitized money, data, credentials, entitlements, property records, IP, etc., which can become economic assets when user-controlled and verifiable. [09:59], [11:18], [12:36]
- Tokenization: Representing real-world or digital assets as digital tokens to restore user control and enable programmable transactability, often utilizing ledgers for assets requiring “singleness”. [20:01], [25:02]
- Credentialing: The process of issuing and verifying digital, cryptographically signed statements (credentials) about an individual or entity (e.g., certificates, licenses). [18:24], [23:51]
- Finternet: A proposed future financial system based on universal, protocol-based infrastructure, generalizing specific financial functions. [14:06]
- Open Networks: Decentralized, interoperable networks for specific sectors (e.g., commerce, energy) built on open protocols, allowing multiple platforms and players to participate without a central intermediary. [27:18]
- Beckn Protocol: An open protocol used for building decentralized open networks, mentioned in the context of ONDC and Uei. [27:58]
- DaaS (DPI as a Service): A proposed model for accelerating the deployment of DPI globally, offering it as a packaged solution. [32:52]
Timestamped Outline / Chapters
- [00:00] Introduction of Pramod Varma
- [02:25] Keynote Start: The Future of DPI
- [03:14] Defining DPI: Reusable Building Blocks
- [04:17] The Three Layers of DPI
- [04:55] Essence of DPI: Reusability
- [05:50] Recap: Foundational DPI Blocks (Identity, Payments, Data Sharing)
- [09:37] User-Centric View: Who Am I? What I Have? What I Do?
- [11:13] Problem: System-to-System Data Sharing Limitations
- [11:13] Solution: User-Centric Approach for Data/Credential Sharing
- [14:06] Introducing Finternet: Future Financial System
- [15:59] Technology Evolution: Purpose-Specific to Generic
- [17:31] Shifting from Platforms to Open-Loop Networks
- [20:01] Digital Assets and Tokenization
- [25:44] Use Case: Sustainability Financing on Finternet
- [26:43] Fragmented Micro-Economies & Open Commerce Networks (ONDC)
- [29:00] Open Energy Networks (Uei) & Other Sectoral Networks
- [29:30] Role of AI in Supercharging DPI for Inclusion
- [31:00] DPI Approach to AI: Democratization
- [32:04] The Problem of Slow Deployment & Need for Universality
- [32:52] Proposing DaaS (DPI as a Service)
- [33:57] Conclusion: Think Protocols, Think Standards, Think Internet
Related Resources Mentioned
- Nandan Nilekani’s Keynote: Referenced from the previous day. [04:23]
- Working Paper: A User-Centric Approach for Universal Credential & Data Sharing: QR code provided for access and contribution. [11:13], [20:07]
- Finternetlab.io: Website for the Finternet initiative/paper co-authored with BIS. [14:06], [21:13]
- Working Paper: DaaS (DPI as a Service): QR code provided for access and contribution. [32:52]
- Beckn Protocol: Open protocol underlying ONDC and Uei. [27:58]
- ONDC (Open Network for Digital Commerce): Example of an open commerce network in India. [27:34]
- Uei (Unified Energy Interface): Example of an open energy network in India. [28:05]
Key Points
- Digital Public Infrastructure (DPI) should be viewed as reusable, general-purpose 'building blocks' (like protocols and standards), not complete, siloed solutions or platforms.
- The essence of DPI is reusability; if a system cannot be reused by others to build diverse solutions, it is not truly DPI.
- Foundational DPI blocks (Identity, Payments, Data Sharing) built over the last decade have driven formalization and frictionless transactions, but data sharing remains nascent.
- The future requires a user-centric approach, especially for data sharing, where individuals control the flow of their trusted, verifiable data and assets using mechanisms like digital wallets and verifiable credentials.
- System-to-system data sharing models are limited, costly, create dependencies, and hinder scalability and cross-border use.
- 'Finternet' represents the next phase for finance: an interconnected financial ecosystem built on universal, protocol-based infrastructure, moving beyond purpose-specific systems.
- Tokenization of real-world and digital assets (user-controlled, attested, registered, regulated) is key to restoring user control and enabling programmable transactability.
- Open, decentralized networks (e.g., Open Commerce, Open Energy) built on common protocols (like Beckn) are emerging globally to break down platform monopolies and foster innovation.
- AI can supercharge DPI by overcoming language and interface barriers, enabling the next billion users to participate in the transaction economy, requiring democratized compute, data, and models (including smaller, use-case-specific LLMs).
- Building DPI one use case, one system, one country at a time is too slow; a faster, universal approach using protocols, standards (like the internet), and potentially 'DPI as a Service' (DaaS) is needed.