To successfully use AI, public and private sector organisations have to “ensure the diffusion happens fast,” according to Satya Nadella, chairman and CEO of Microsoft, who spoke at the World Economic Forum Annual Meeting in Davos, Switzerland. “This has always been the arc of computation, we saw it in the web, mobile and cloud era. AI is of the same class as those arcs.” Using computers amplified productivity for knowledge workers, and Nadella says the same thing will happen with AI.
Nadella highlighted that AI is widely admired right now, but it needs to start making a tangible difference for proper diffusion (the widespread adoption and practical use of the technology). For instance, if AI models are consuming national energy resources, their output – measured in AI tokens – should produce measurable benefits, such as improvements in healthcare.
“We have to make sure the supply is there (everything from the chips and token factories),” said Nadella. “The token factory [the data centres, graphics processing units (GPUs) and software designed to convert data into AI-powered insights] is the first thing that will be diffused – you need a ubiquitous grid of energy and tokens. Then the demand side of this is every firm has to start using it [AI]. It’s going to require real leadership in the private and public sector to ensure that diffusion is successful. Diffusion is also strongly correlated to how broadly people are skilled to using this.”
Technology companies also need to ensure AI diffusion is evenly spread across the world.
“We have the ability to deliver the tokens evenly around the world – more so than in the PC and mobile era,” said Nadella. “As hyperscalers, we are investing all over, and then that will lead to demand.”
However, Nadella warned that investment alone will not be enough without parallel progress in energy and infrastructure. “The grid in most countries is fundamentally driven by the government and the public – that will hold things back,” he said. “A long-term scalable solution is to have all these token factories, connected to the grid, that will drive and add scale across the world. If you look at tokens per dollar per watt, gross domestic product growth of any place will be directly correlated. If you have a cheaper commodity (grid usage), it’s better. It’s not just the production side, that’s why having the grid is important.”
Microsoft’s Satya Nadella speaks with BlackRock’s Laurence Fink at the World Economic Forum Annual Meeting 2026
Governments are increasingly recognising AI as a core utility.
“In 2026, almost every government has recognised AI is no longer a nice–to-have concept but is becoming a real, physical utility that the world is already consuming,” said Peng Xiao, CEO of G42, in an interview with Bloomberg on chip imports and AI development, while at the World Economic Forum Annual Meeting. “We are seeing governments across the globe deploying AI agents to support processes. They are now racing to implement policies to ensure that AI deployed in their country reflects their national policy and agenda. They also want to make sure AI infrastructure they are building out has some degree of independence so nobody else externally can unplug it and cause disorder inside their society.”
Against that backdrop, the United Arab Emirates (UAE) has positioned itself as one of the earliest and most proactive government actors in the AI space.
“The UAE has been in this space for more than five years now,” said Xiao. “Abu Dhabi was already leading AI before AI was cool. We foresaw AI becoming a utility that our world cannot live without. We’ve been forming partnerships around the world, knowing that AI cannot be developed by a single company or nation, including Microsoft.”
The UAE’s Stargate project demonstrates this ambition.
“We announced the buildout of the largest AI campus outside of the United States,” said Xiao, referring to a one-gigawatt compute cluster that will be built by G42 and operated by OpenAI and Oracle. The facility will provide low-latency inferencing to deliver AI that meets the world’s increasing demand.
Since the G42 announcement, the project has moved quickly from ambition to execution.
“In this complex world of AI, nothing is straightforward but through the past half year, we’ve made tremendous progress,” said Xiao. “The first batch of the most advanced AI chips will be shipped to the UAE in the next couple months – they are mostly Nvidia.”
Xiao summarised the challenges of AI as job security, power concern and national security if governments fall behind.
“People are confused, concerned and asking for a pause,” he said. “But I think it’s a mistake [to pause]. Many other nations will not stop their development and there will be AI diffusion around the world if our technology [despite calls for restraint or regulation]. So, I think this year, it’s time to double down. UAE has two great advantages when it comes to AI infrastructure. One is abundance of power and the second is incredible connectivity. We will play a key role in the global AI network. We’ll be a key node on the grid to serve almost four billion people.”
Representatives from G42, Abu Dhabi, Dubai, Nvidia, OpenAI, the UAE, Oracle and more have partnered to build Stargate UAE
The next era of technological growth, will be defined by AI infrastructure, compute power and GPU innovation, according to Jensen Huang, CEO of Nvidia. During his conversation with Fink at the World Economic Forum Annual Meeting, Huang outlined five key layers of AI infrastructure: energy forms the foundation, followed by chips and computing hardware, then cloud, AI itself, and finally the top layer of real-world applications, such as in financial services, healthcare or manufacturing.
He used the example of a radiologist to show how AI is enabling people to do their jobs better by taking away the drudgery of processing scans. “The purpose is enhanced and made more productive because the task has been automated.”
Huang emphasised that AI skills are like other leadership skills. “AI skills are no different than leading people, managing people, things we do all the time. Recognise that AI is likely to close the technology divide,” he said. “I’m fairly optimistic about the potential of AI to lift the countries that are emerging.”
Nvidia’s Jensen Huang at the World Economic Forum Annual Meeting 2026
Organisations must also embrace AI culturally and operationally. Businesses should “think about changing the workflow,” said Nadella. “You can’t be afraid of it – it’s going to be diffused – so learn how to use it.”
Nadella predicts there will be no single dominant AI model across enterprises and public sectors.
“It’s a multi-model world,” he said. “There are going to be multiple models and the trick is how to take advantage of them to build your own model by distilling these. Harness engineering – the IP of any application or firm is how to use these models using context of your data. Feed it your data to change the trajectory of some outcome that I care about. This is a technology that will build on the rails of cloud and mobile, diffuses faster and brings economic growth around the world. It has demand all over the world and it will only be there if there is local surplus.”