In 2015, a small gathering in San Francisco made a decision that barely registered in the global media cycle. There were no headlines, no speeches and no sense of historical consequence. Yet, in retrospect, that moment marked the beginning of a structural shift in the global economic order. The founders of OpenAI were not building another technology company. They were attempting to industrialise intelligence.
Elon Musk saw artificial intelligence as both a civilisational risk and a transformative force. Sam Altman approached it as an execution challenge: scale models, secure compute and deploy intelligence as an economic utility. Different styles, same strategic conclusion — scalable intelligence would shape productivity, capital flows and geopolitical leverage in the 21st century.
At the same time, in Santa Clara, Jensen Huang was making a quieter but far more consequential bet. Nvidia was still viewed as a gaming chip company. Huang had already pivoted towards parallel computing, CUDA architecture and GPU acceleration. Markets did not immediately understand the shift.
Deep learning remained a niche field. Venture capital chased consumer apps. Policymakers debated social media while the computational foundations of the AI age were being built with little public attention.
When Nvidia introduced purpose-built AI systems, commercial demand was limited.
The machines were expensive, complex and ahead of their time. Yet they would soon become the core infrastructure behind nearly every major breakthrough in artificial intelligence. The real ‘picks and shovels’ of the AI revolution were not apps but compute clusters measured in megawatts, with high capital intensity and significant energy consumption.
What followed was exponential, not linear. The partnership between OpenAI and Microsoft transformed frontier research into industrial-scale deployment. GPT-3 demonstrated emergent capabilities that surprised even its developers.
Then, in November 2022, ChatGPT entered the public domain and triggered one of the fastest adoption curves in technological history. Within months, hundreds of millions of users were interacting with machine intelligence capable of writing, coding, analysing and synthesising information at scale.
This was not just a software milestone. It marked the beginning of an AI-driven capex supercycle. Hyperscalers committed tens of billions of dollars to AI infrastructure.
Data centre investments surged. Energy demand projections were revised upward. Training frontier models began to require computational resources comparable to national research facilities of earlier decades. Nvidia’s rise to multi-trillion-dollar valuation levels reflected a deeper shift: compute had become a strategic commodity.
Across Asia, policymakers and corporations read these signals with clarity. India accelerated investments in digital infrastructure, AI ecosystems and IT services scaling. Vietnam deepened manufacturing integration, export competitiveness, and technology-enabled supply chains.
China intensified investments in semiconductors, AI models and industrial automation despite external constraints. The region increasingly treats AI as a productivity multiplier embedded in national economic strategy.
Now consider, in parallel, Pakistan’s intellectual and policy trajectory over the same decade. While the world was industrialising intelligence, Pakistan’s national discourse remained dominated by political scandals, institutional confrontation and ideological flashpoints.
Media cycles recycled accusations. Social media amplified outrage. Strategic debate on compute infrastructure, AI adoption, industrial automation and technological competitiveness remained peripheral.
Pakistan’s R&D expenditure remains around 0.16% of GDP, among the lowest globally. Innovation rankings lag behind regional peers. Network readiness remains weak. AI preparedness indicators place the country behind not only India and China but also increasingly Bangladesh and Vietnam. These are structural indicators of technological underinvestment, not perception gaps.
The frequently cited claim of a large STEM pipeline also weakens when a quality filter is applied. The effective high-skill pool relevant to an AI economy is concentrated within a narrow set of institutions — NED, GIKI, LUMS, FAST and a handful of credible engineering departments.
Even within this group, exposure to advanced computing, research ecosystems and frontier AI tools is limited compared to regional competitors. Headline graduation numbers suggest scale; the underlying talent depth remains uneven and thin in critical areas such as advanced computing, applied AI and high-end engineering.
This structural weakness is now intersecting with Pakistan’s core export sectors, particularly textiles and IT services. Textiles, the backbone of export earnings, are entering an era defined by automation, AI-driven optimisation and digitally integrated supply chains.
Vietnam is embedding smart manufacturing, predictive logistics and automation into production networks. India is leveraging scale, policy support and technology adoption to move up the textile value chain.
Pakistan’s textile sector, in contrast, remains energy-constrained, technologically underinvested and vulnerable to efficiency shocks. As automation reduces the importance of labour cost advantages alone, countries that fail to modernise risk losing competitiveness even in traditional labour-intensive sectors.
The threat is sharper in IT exports. India’s IT services sector is rapidly integrating AI tools across software development, enterprise services and customer operations, significantly improving productivity per worker. Vietnam is positioning itself as a credible technology outsourcing hub with coordinated policy support and skill alignment.
Pakistan’s IT sector, despite pockets of excellence, risks stagnation if AI adoption remains shallow and infrastructure gaps persist. In a world where AI increasingly automates coding, testing and documentation, low-end outsourcing models face structural compression.
Yet policy discourse continues to revolve around IMF programmes as if they constitute an economic strategy. In practice, repeated IMF arrangements function primarily as creditor-stabilisation mechanisms designed to ensure external repayment capacity and short-term macroeconomic stability.
They are not development frameworks and do little to address the structural transformation needs of a rapidly growing, young population. Stabilisation without productivity enhancement merely postpones crises while protecting creditor confidence more than long-term industrial competitiveness.
The deeper issue is that external financing has repeatedly substituted for internal reform. Instead of sustained investment in technological capability, industrial upgrading and research infrastructure, policy responses have defaulted to short-term stabilisation cycles. Public discourse often frames economic distress through political or conspiratorial narratives rather than confronting low productivity, weak export diversification and technological lag.
In the modern economy, focus is a strategic resource. Over the past decade, leading economies have focused on AI, automation and digital infrastructure. Pakistan allocated disproportionate attention to political theatrics and institutional disputes. The opportunity cost of this misallocation is now compounding.
The global economy is entering a phase where intelligence is embedded across production systems — manufacturing, logistics, finance, healthcare, and services. Firms deploying AI will reduce costs, improve efficiency, and capture market share at scale. Countries that fail to integrate AI into their industrial base will experience a gradual erosion of competitiveness, even in sectors where they historically held advantages.
This is the emerging risk for Pakistan: not sudden collapse, but steady strategic erosion. A slow weakening of textile competitiveness as automated factories in Vietnam and technologically scaled producers in India outperform in efficiency, compliance, and delivery timelines.
A plateau in IT exports as AI-enabled competitors offer higher productivity and greater value addition. A widening technological gap that translates into slower export growth, currency pressure and deeper dependence on external financing.
Globally, the decade from 2015 to 2025 will likely be remembered as the period when intelligence became industrial infrastructure — as foundational as electricity or the internet in earlier eras.
Countries that recognised the shift early generated productivity gains, attracted capital and strengthened strategic leverage. Countries that remained intellectually inward-looking widened their structural vulnerabilities.
Technology compounds silently and asymmetrically. While others invested in computing, research and industrial automation, Pakistan remained absorbed in cyclical political narratives. While competitors built AI ecosystems and upgraded industrial capabilities, Pakistan overstated readiness without matching infrastructure or policy urgency.
The real danger is not exclusion from the AI revolution. It is far more insidious: participation at the lowest-value-added tiers while regional competitors capture higher-value segments in textiles, manufacturing and IT services. In an AI-driven global economy, competitiveness will be determined less by labour costs and more by computing access, technological depth and institutional focus.
On each of these fronts, the gap with countries such as India and Vietnam is not static. It is widening — and in a compounding technological era, widening gaps tend to harden into long-term structural disadvantages that become progressively harder to reverse.
The writer is former head of Citigroup’s emerging markets investments and author of ‘’he Gathering Storm’.
Disclaimer: The viewpoints expressed in this piece are the writer’s own and don’t necessarily reflect Geo.tv’s editorial policy.
Originally published in The News
