Financing the Future: Building AI Ready Digital Foundations for Asia and the Pacific

Detailed Summary

• Antonio Zaballos opened the session, framing the conversation around two “layers” of AI‑ready foundations.
  • Layer 1 – Core Digital & Data Infrastructure – broadband, data‑centres, cybersecurity, and protection of critical information.
  • Layer 2 – Accelerators – AI compute capacity, foundation‑model development, and sector‑specific use‑cases such as early‑warning systems, energy‑transport digitalisation, and social‑services optimisation.
• He stressed that both layers require public‑private involvement and a “quantum leapfrog” in adoption across sectors.
• Antonio highlighted the ADB’s role as a partner for governments, supporting policy, financing and implementation.

2. India’s AI‑Ready Digital Foundations – The India AI Mission

• Khushal Wadhawan (India AI) presented the flagship India AI Mission, a five‑year, US$1.2 billion programme.
• Historical context – India has been building Digital Public Infrastructure (DPI) for over a decade (Aadhaar, Unified Payments Interface, DigiLocker). Key statistics shared:
  • Payments – ≈ 21.7 billion transactions per month, ≈ US$290 billion value.
  • Aadhaar – > 99.5 % population coverage; generating US$42 billion in savings via Direct Benefit Transfer.
  • DigiLocker – 660 million users, 9.5 billion verified documents.
• Why AI now? – AI is reshaping work, productivity, and governance. Integrating AI with DPI can “defrost” traditional development cycles.
• Seven pillars of the Mission (each a digital foundation):
  1. Compute Capacity – scaling from 438 GPUs (2021) to 38,000 GPUs via public‑private partnerships, demand aggregation, and price‑transparency mechanisms.
  2. Foundation Models – development of indigenous, multilingual models suited to India’s linguistic diversity.
  3. AI Co‑Ecosystem (AI Coach) – a central repository of datasets, models and development tools; comparable to an “AI Hub”.
  4. Application Development Initiative – “AI‑for‑good” pilots (e.g., synthetic cyber‑crime database hackathon).
  5. Future‑Skills Pillar – up‑skilling of technical‑workforce in tier‑2/3 cities, establishing AI labs in industrial training institutes.
  6. Safe & Trusted AI – creation of governance tools, benchmarks and technical interventions to avoid over‑burdening bureaucracy.
  7. Startup Financing Pillar – patient‑capital funds for AI start‑ups tackling non‑commercially viable problems.
• Outcome snapshot – by the time of the session, the mission had achieved:
  • 38 k GPUs (≈ 87× increase),
  • 10 k datasets, 250 models, 12 AI start‑ups, 30 AI labs (target 270).
• Key insight – the rapid GPU scale‑up was possible only through PPPs; the government did not fund all hardware outright but leveraged demand‑aggregation and transparent procurement to attract private capital.

3. ADB’s Regional Perspective – “Readiness, Financing, Coordination”

• Carolyn Florey framed the remainder of the panel around three cross‑cutting themes: Readiness, Financing, Coordination.
• She invited each panelist to share country‑specific experiences, noting that the Asian‑Pacific region shows heterogeneous maturity – from advanced digital public infrastructure (India) to nascent foundations (several Central Asian states).

3.1. Financing AI Infrastructure – Private‑Sector & Development‑Finance Blend

• Mayank Chaudhary (ADB, private‑sector financing) highlighted three financing archetypes:
  1. Fully Government‑Funded Model – rare in low‑income settings; exemplified by certain Indian state‑run DPI projects.
  2. Pure Private‑Sector Proprietary Model – large tech‑companies own compute without sharing (e.g., cloud‑only services).
  3. Hybrid PPP / Blended‑Finance Model – most suitable for developing economies.
• He emphasized that private investors need guaranteed demand at affordable tariffs; without it, they cannot secure a return on capital.
• Illustrative analogy – ADB’s experience with a dialysis‑machine PPP in India: the government guaranteed a minimum patient volume, allowing the private operator to invest confidently.
• Proposed approach for AI – governments guarantee a baseline “GPU‑hours” or “data‑centre capacity” at a subsidised price; multilateral banks provide concessional long‑term debt; private firms bring equity and operational expertise.
• ADB can also aggregate demand across multiple countries, creating shared regional data‑centres that achieve economies of scale.

3.2. Government View on Financing – Uzbekistan

• Azizjon Akramov (Uzbekistan Ministry of Finance) reiterated that fiscal space is tight, yet AI‑ready infrastructure should be treated as a long‑term productivity catalyst.
• He identified three macro‑benefits:
  1. Better tax compliance & reduced fraud → higher revenues.
  2. Enhanced competitiveness → private‑sector growth.
  3. Avoiding future cost‑inflation – early investment prevents expensive retrofits later.
• Flagship projects:
  • 12 MW data‑centre in IT‑Park, Tashkent (US$150 m, LEED‑Gold), to become Central Asia’s first AI‑enabled data centre with links to Google, Microsoft, Meta, Amazon.
  • Semiconductor & electronics strategy – early‑stage development supported by ADB technical assistance.
• He stressed the need for regional cooperation (double‑landlocked geography) and blended‑finance (public guarantees + private equity).

3.3. Tajikistan’s Journey – From Pilot Academies to Data‑Center Ambitions

• Firuzjon Sodiqov described Tajikistan’s AI‑first strategy despite limited natural resources and mountainous terrain.
• Key milestones:
  • AI Academy – 50‑person cohort, later expanded; early graduates struggled to apply skills, prompting the launch of AI start‑ups (≈ 10 companies, exporting to 25 countries).
  • Leverage hydropower & solar – Tajikistan’s abundant green energy is positioned as a cheap compute substrate.
  • Regional partnership – Tajikistan participates in a UN‑mandated Central‑Asia AI‑infrastructure partnership, facilitating talent exchange and joint data‑centre development.
• Challenge – scaling compute without a domestic market large enough to sustain it; thus cross‑border PPPs are essential.

3.4. Kazakhstan’s Coordination Blueprint

• Though not on the original speaker list, a Kazakhstan panelist (identified as “Aidana”) shared practical coordination lessons:
  • Data interoperability – creation of a national data‑exchange platform (“Smart Data”) that aggregates agency datasets, enabling AI to act on unified information.
  • Shared digital government infrastructure – common identity, integration layers, and service platforms that avoid siloed ministries.
  • Institutional coordination – clear standards, roles and responsibilities across regulators, IT operators, and data owners; AI KPI’s are embedded in ministerial performance metrics.
  • Use‑case example – e‑Appeal system for citizen complaints: AI classifies, routes and prioritises requests, dramatically reducing manual workload and improving transparency.
• The speaker argued that the primary barrier to AI deployment is not the algorithm but the surrounding ecosystem (data, governance, trust).

4. Cross‑Cutting Themes & Emerging Consensus

Theme	Key Points Raised by Panelists
Readiness	• Robust DPI (identity, payments, cloud) is prerequisite (India, Kazakhstan). • Compute capacity must be affordable and regionally shared (Uzbekistan, Tajikistan).
Financing	• Blended finance (public guarantees + private equity + multilateral concessional debt) is the most realistic model. • Demand‑aggregation and price‑transparency mechanisms reduce private‑sector risk (India’s GPU aggregation, ADB dialysis‑PPP analogy).
Coordination	• High‑level political endorsement (Tajikistan’s presidential decree) creates mandatory KPIs for ministries. • Shared data‑exchange platforms enable AI to work across silos (Kazakhstan).
Policy / Regulation	• Need for regulatory clarity on AI‑use, data‑privacy, and responsible AI (India’s “Safe & Trusted AI” pillar).
Talent & Skills	• AI academies, up‑skilling of technical‑workforce, and creation of AI labs in tier‑2/3 regions (India’s future‑skills pillar; Tajikistan’s academy).
Public‑Sector Role	• Government may act as provider of compute (public cloud), regulator/facilitator, or guarantor of demand – a mix of all three appears in successful cases.
Regional Cooperation	• Cross‑border data‑centres and shared‑infrastructure reduce costs for small or land‑locked economies (Uzbekistan‑Tajikistan partnership, ADB’s multi‑country model).

5. Closing Remarks

• Moderator thanked all panelists and noted that time constraints prevented an audience Q&A, but panelists remained available for follow‑up discussions after the session.

Key Takeaways

• AI‑ready foundations start with solid Digital Public Infrastructure (identity, payments, data exchange) before scaling compute or AI models.
• Blended‑finance PPP models—public demand guarantees, multilateral concessional debt, and private equity—are the most viable way to fund AI compute and data‑centre projects in low‑ and middle‑income countries.
• Regional cooperation (shared data‑centres, cross‑border demand aggregation) can dramatically lower costs for land‑locked or small economies.
• Political buy‑in and KPI‑driven mandates (e.g., Tajikistan’s presidential decree) are essential to push ministries toward AI adoption and to overcome “black‑box” skepticism.
• Talent pipelines (AI academies, up‑skilling programs) and AI‑for‑good pilots help create a domestic ecosystem of startups that can commercialise AI solutions.
• Data interoperability platforms (Kazakhstan’s “Smart Data”) are the critical “glue” that turns isolated digital services into an AI‑ready ecosystem.
• Regulatory clarity and safe‑AI frameworks protect citizens and reduce bureaucratic friction, enabling faster rollout of AI‑enhanced public services.
• Monitoring and evaluation (e.g., guaranteed GPU‑hour usage, demand‑side subsidies) provide the confidence needed for private investors to commit capital.

These insights collectively map a practical road‑map for Asian‑Pacific nations to transition from basic digital services to fully AI‑enabled, inclusive, and resilient economies.

See Also:

• democratizing-ai-resources-in-india
• ai-for-everyone-empowering-people-businesses-and-society
• the-sustainable-digital-infrastructure-accord-driving-sustainability-of-ai-infrastructure-in-the-asia-pacific-region
• scaling-ai-solutions-through-southsouth-collaboration
• powering-the-public-good-aligning-industry-philanthropy-and-government-from-india-to-africa
• ai-for-economic-growth-and-social-good-ai-for-all-driving-economic-advancement-and-societal-well-being
• ai-innovators-exchange-accelerating-innovation-through-startup-and-industry-synergy
• responsible-ai-at-scale-governance-integrity-and-cyber-readiness-for-a-changing-world