By Ammad Aslam Ghumman
We often hear that the United States and China are competing to lead the world in Artificial Intelligence. But perhaps we still do not fully understand what this competition means.
AI is not just about using AI Model to complete the software application development. Behind every operational advanced AI model is a physical infrastructure of powerful processors, specialized chips, data centers, high-speed networks, massive data and, above all, reliable source of energy.
In simple terms, AI computing power is a proper strategic resource.
Just as oil and electricity powered industrialization in the last century, computing power is becoming one of the foundations of economic and national power in the AI era.
This is one reason the US-China technology rivalry has become so important. The United States is restricting China’s access to some advanced AI chips and semiconductor technologies, while China is investing heavily in domestic alternatives. According to IDC data reported by Reuters, Chinese companies shipped approximately 1.65 million AI components in 2025, representing about 41% of China’s AI accelerator server market. Huawei, the largest Chinese manufacturer of chips, alone accounted for around 812,000.
The lesson is clear: China is not simply trying to build better AI models. It is trying to build an alternative complete AI technology ecosystem.
The Gulf has understood this strategic shift as well. Countries such as the UAE and Saudi Arabia are investing heavily in AI infrastructure, cloud computing and large data centers. The UAE has set an ambition for AI to become deeply integrated into government operations and run 50% of operations on AI by 2030, while Saudi Arabia is pursuing some of the world’s largest AI infrastructure investments.
This highlights the importance of Sovereign AI.
Sovereign AI does not mean that every country must manufacture its own GPUs or build every component itself. It means having sufficient domestic capability and strategic options to develop and use critical AI capabilities without becoming completely dependent on security of data and models decisions making. Nations must have the ability to not just procure AI infrastructure but also secure it.
Where does Pakistan stand in this race?
We have some important advantages. Pakistan has a large young population, a sizeable technology workforce, a growing digital economy and a strong base of software professionals and freelancers. IT and ICT export exports crossed $4 billion in last fiscal year.
We are also beginning to build physical AI infrastructure. Sky47’s Karakoram-01 in Islamabad has an announced power capacity of 8.5 MW and was developed with Chinese technology company ZTE as an end-to-end data center infrastructure partner. It incorporates technologies including cold-plate liquid cooling for high-density computing.
These developments are encouraging. But where will the electricity come from?
AI data centers cannot operate on intermittent power. They require highly reliable electricity and cooling systems. The International Energy Agency estimates that global data center electricity consumption could more than double to around 945 TWh by 2030, since countries will be building AI data center to support Sovereign AI.
For countries like us it’s a challenge. We rely on imported fuel to generate a portion electricity, though we have reduced the dependency in recent years. Reality is we are still dealing with electricity shortages, load management and transmission constraints.
At the same time, Pakistan faces a serious water challenge. We can’t provide drinkable water to residents of Karachi from decades. AI data centers generate enormous amounts of heat. Depending on the design, cooling can require significant quantities of water. Modern closed-loop and liquid-cooling technologies can substantially reduce water consumption, but cooling remains an important infrastructure consideration.
This matters in Pakistan because water is not an unlimited resource. The Pakistan Council of Research in Water Resources has warned that Pakistan’s annual water availability has fallen below 1,000 cubic meters per person, putting the country in the water-scarce category.
This creates a difficult but necessary debate.
Should a country struggling to provide reliable electricity and safe water invest in energy-intensive AI infrastructure?
My answer is yes but strategically. Because losing this race is significantly damaging to not just our economy but also to our governance and military strength. I can just present the ideas but execution of these in the best interest of our country will happen only if honest and visionary efforts are made. The execution will require competent people who hold the love of land and its people dear to them.
We should not see AI data centers as competitors to ordinary citizens for electricity or water. We should design them in a way that creates additional economic value without worsening existing resource pressures.
That means locating AI infrastructure where reliable power is available, developing dedicated energy solutions, prioritizing renewable and nuclear generation where appropriate, using advanced low-water cooling, and making water consumption a part of data center planning from the beginning.
Pakistan needs to move beyond the question of how many people we can train in AI to accomplish the real goals.
We must also ask, how much computing power will Pakistani researchers, startups, universities and companies have access to our own AI facilities?
If we train hundreds of thousands of people but they remain dependent on foreign platforms and foreign computing infrastructure, we will have created AI users not a sovereign AI economy.
Our objective should be compute access sovereignty.
Pakistan does not need to manufacture the world’s most advanced GPU tomorrow. But we should build enough domestic AI data center capacity, energy infrastructure, data resources and AI engineering capability to give our companies and researchers a genuine opportunity to compete with world.
This will not happen without good governance.
AI infrastructure will require significant public and private investment. The real test will not simply be how much money we announce, but how much value we generate from every rupee invested. Transparent procurement, professional institutions, accountability and long-term policy continuity will determine whether these investments become productive national assets or another collection of underutilized projects. This is basically all the ingredients of running the affairs of governance with sincerity. I will try to address this national challenge in a different article hopefully soon.
Pakistan should also avoid choosing one technology bloc. The US, China, Gulf states and other technology ecosystems all offer different capabilities and sources of capital. Our strategy should be to work with multiple partners while developing enough domestic capability to preserve our strategic choices. This will require strong diplomatic and trade ties.
About The Author
