Pakistan Just Put $1 Billion Behind Its AI Plan, And South Asia's Second Tier Is Waking Up

Pakistan Just Put $1 Billion Behind Its AI Plan, And South Asia's Second Tier Is Waking Up

Pakistan's Federal Cabinet approved the Artificial Intelligence (AI) Policy 2025 in late 2025, and the implementation numbers are beginning to show. A $1 billion national AI investment target by 2030, announced at Indus AI Week, has been followed by the AI Seekho 2026 training initiative launched in April, the Islamabad AI Declaration, and the first tranche of Centres of Excellence across seven cities. South Asia's AI story has been India for so long that most regional observers missed Pakistan stepping up.

The Quiet Policy Turn

The architecture is ambitious for a country of Pakistan's fiscal position. The policy is built on six pillars, including a National AI Fund (NAIF), a national compute^✦ grid, Centres of Excellence in seven cities, and formal AI curricula in federal schools. The Ministry of IT and Telecommunications (MoITT) is running it alongside the Special Investment Facilitation Council (SIFC) and the Pakistan Software Houses Association (P@SHA), with implementation partners including Google for Developers and Telenor Pakistan.

The domestic politics matter. Pakistan's IT exports hit $2.2 billion in the first half of fiscal year 2026, up 20% year-on-year. That is a real foreign-exchange story, and it gives the policy a base constituency beyond the usual technocratic enthusiasm. AI is being framed not as a luxury, but as the mechanism through which Pakistan can double IT exports by 2030.

How This Compares To India's Run

The scale gap is real. India's IndiaAI Mission will mobilise tens of billions of dollars in compute subsidy and will anchor a sovereign foundation model^✦. Pakistan's plan is one to two orders of magnitude smaller. But the timing and focus are meaningful. Pakistan is concentrating on workforce development, language-specific models for Urdu and regional dialects, and public-service AI, rather than trying to compete directly on compute.

That is the right play. In South Asia's second tier, the first winners are likely to be those who ship localised, low-cost, narrow AI systems into government service and SME markets, not those who attempt to stand up a frontier lab. Bangladesh and Sri Lanka should be watching. Both have talent pools and call-centre economies that could move into AI-enabled services faster than a capital-intensive compute strategy.

Our approach is deliberately human-capital-first. We do not believe Pakistan needs to compete with India on compute. We need to compete on trained workforce and localised AI utility.

The AI Seekho 2026 programme reaches students in under-served regions that would otherwise never touch an AI curriculum. That is where the compounding happens.

What To Watch In The Next 12 Months

Three things will decide whether the policy sticks. First, funding continuity. A $1 billion target by 2030 requires sustained budget allocations through multiple IMF review cycles, and fiscal headroom remains the single biggest domestic risk. Second, execution capacity. Seven Centres of Excellence is ambitious, and past Pakistani federal initiatives have struggled with uneven provincial delivery. Third, foreign partnership balance. Pakistan's AI posture sits between Chinese cloud players (Huawei, Alibaba Cloud) and Western ones (Google, Microsoft), and the policy does not yet declare a preferred partner architecture.

1. Watch the National AI Fund's first disbursements in Q3 2026.
2. Watch whether the Centres of Excellence actually open on time and with listed faculty.
3. Watch which foreign cloud provider signs the first multi-year compute partnership.
4. Watch how Urdu and regional-language foundation models perform at the first public benchmark^✦ in late 2026.

The Broader South Asian Signal

Bangladesh has a parallel story brewing. Its freelancer economy, the third largest in the world by gig volume, is now a meaningful consumer of AI tooling, and Dhaka is working on a national AI policy draft. Sri Lanka's ICTA has a draft AI strategy that is expected to be published in Q2 2026. Nepal and Bhutan remain behind, but both have begun public-sector AI pilots with regional donors. Pakistan's move is the first real sign that the second-tier South Asian AI story is not going to wait for India to export its approach.

This is part of a pattern we have been tracking. See our detailed write-up on India's IndiaAI Mission and sovereign AI, our analysis of Bengaluru AI salaries surpassing London, and our reporting on India's April 2026 IT Rules on AI content. For the broader South and Southeast Asian policy landscape, the Vietnam AI law phased rollout provides a useful contrast.

Frequently Asked Questions

What is the National AI Fund (NAIF)?

NAIF is a public investment vehicle established under the AI Policy 2025 to finance Centres of Excellence, the national compute grid, and grants to Pakistani AI startups. It is being operationalised by MoITT with support from the Special Investment Facilitation Council.

How does Pakistan's AI policy compare to India's IndiaAI Mission?

Pakistan's plan is one to two orders of magnitude smaller and is workforce-first rather than compute-first. India is building sovereign GPU^✦ capacity and foundation models. Pakistan is building people, curricula, and sector-specific AI applications.

What role are Chinese and Western firms playing?

Both ecosystems are engaged. Google for Developers, Microsoft, and Telenor Pakistan are partners on AI Seekho. Huawei and Alibaba Cloud are competing for compute and enterprise AI deals. Pakistan has not yet declared a preferred architecture.

Is there a comparable Bangladesh policy?

Bangladesh is drafting a National AI Policy expected in 2026. It is not as detailed as Pakistan's current document, but the freelancer-economy angle gives Dhaka a distinct leverage^✦ point that is worth watching closely.

Will Pakistan's workforce-first approach prove more durable than India's compute-first play over the next five years? Drop your take in the comments below.