Pakistan has placed artificial intelligence (AI) near the centre of its economic agenda. At the Indus AI Summit in February, the government announced a $1 billion AI investment commitment by 2030.
ARTIFICIAL INTELLIGENCE
Pakistan has placed artificial intelligence (AI) near the centre of its economic agenda. At the Indus AI Summit in February, the government announced a $1 billion AI investment commitment by 2030.
The plan covers sovereign compute infrastructure and research, an AI curriculum for schools, 1,000 fully funded PhD scholarships and training for one million non-IT professionals.
Prime Minister Shehbaz Sharif recently pointed to a national backbone of more than 234,000 kilometres of fibre-optic network, six submarine cables, around 58,000 cellular towers, and over 20 modern data centres.
These are substantial commitments. Pakistan is going to create the conditions that allow AI to improve the way businesses operate, workers build skills and public institutions serve people.
GPUs attract attention because they are visible, expensive and closely associated with the global AI race. They are also only one part of the system.
The World Bank, in its Digital Progress and Trends Report 2025, offers a useful framework that identifies four foundations for AI adoption: connectivity, compute, context and competency. The four cover infrastructure, processing power, locally relevant data and the skills to use it. A powerful model creates little value when an SME cannot count on stable internet access. The same applies when a public agency works with disconnected databases or when a teacher receives a new tool without the training needed to use it well.
Running a digital platform makes this easy to see. People experience technology through reliability, speed, cost, and trust -- the same standards by which AI will be judged.
The same applies to our work at MEXC. In the first quarter of 2026, MEXC AI-related features reached around 140,000 daily active users, with more than 1.04 million users over the quarter. No surprise that people are more likely to use AI when it is useful, easy to access, and built into tools they understand.
Pakistan already has important digital foundations in place. Its national identity system and instant-payment infrastructure have created opportunities for more efficient public services. Yet the country still lags behind comparable economies in digital infrastructure, digital governance and the broader environment needed for a strong digital economy.
There is also uneven connectivity across districts and the relatively high cost of fixed broadband. These problems can feel abstract until a business tries to move more of its operations online. AI adoption makes dependable access even more important.
Power is one of the clearest pressure points. World Bank President Ajay Banga described electricity-sector reform as Pakistan’s most urgent near-term priority. Distribution losses and inefficiencies continue to limit growth, even after improvements in generation capacity.
AI will not solve Pakistan’s employment challenge by itself. It can still help companies become more productive, support new businesses and give workers useful tools. The real impact will depend on how widely those tools are available
The underlying problem is that Pakistan has continued to experience outages and load shedding despite maintaining surplus generation capacity. A large part of that capacity remains underused because the transmission network has not kept pace.
AI adds a further dimension of urgency. Data centres and cloud platforms require reliable power. If they need to work around power cuts, backup power charges, or patchy internet speeds, it will be difficult for companies to integrate AI tools into their daily workflows.
The local data-centre investment can place additional pressure on power grids. New computing capacity has to be matched with affordable energy and resilient digital networks.
So before Pakistan can aim to lead in AI, it needs a more dependable digital economy. Grid modernisation, broadband access, fibre networks and data-centre investment belong in the same conversation.
Hardware and connectivity will only take the country so far. The people using AI will ultimately determine its economic value. Pakistan has started to recognise this. The national plan announced at the Indus AI Summit includes training for one million non-IT professionals. A separate initiative is already training 10,000 government officials to strengthen digital capacity across public institutions and improve service delivery.
This is a promising start. People also need the skills to use these tools with confidence. Roads help people get to markets and telecom networks help businesses get to customers. Practical AI skills can help workers find their place in a shifting economy, too.
AI capability must be built as infrastructure, underpinned by learning anchored in real tasks and measured in practical competencies. One-off workshops can raise awareness, but lasting value is in skills that people can use in their day-to-day work.
For a civil servant, this could mean using AI to speed up and improve the quality of public services; for an SME, it may mean reducing repetitive administrative work or serving customers more efficiently. Teachers could use carefully designed tools to support learning and healthcare professionals could work with patient information to save time.
AI’s real value tends to be in practical tools that help people do their jobs better. These use cases may receive less attention, but they are often where economic gains begin. The approach we take at MEXC is that useful AI should make a familiar task easier to complete without adding another layer of complexity.
Pakistan’s AI ambitions also need to be understood in the context of the country’s employment challenges. It needs to generate between 25 million and 30 million jobs over the next decade as millions of young people enter the workforce.
AI will not solve Pakistan’s employment challenge by itself. It can still help companies become more productive, support new businesses and give workers useful tools. The real impact will depend on how widely those tools are available. The race Pakistan needs to win is not for the largest model. It is for the moment a shopkeeper in Multan, a teacher in Quetta and a clerk in Karachi all reach for the same tool without thinking twice.
The countries building the biggest models may not benefit most from AI. And those making these tools reliable, affordable and useful in everyday life will.
Pakistan has already made its ambition clear. Build the grid, the broadband, and the skills and the models will take care of themselves. That is the work Pakistan has left to do.
The writer is the CEO of MEXC.
