Trends 2040

In this scenario, AI becomes a utility, like electricity or the internet. You no longer open a separate application: intelligence sits inside machines, software, buildings and phones. The early signs are here. More AI runs on devices themselves, using dedicated chips, while assistants work within programs employees already use every day. Quantum computers calculate according to the laws of quantum mechanics. They could greatly accelerate specific tasks, such as simulating molecules, though it remains uncertain when they will be reliable enough for everyday use. In this version of 2040, almost everything people work with can think alongside them, including the forklift and the office building.

Robots have long followed programmed movements exactly. AI changes that. Models that understand language are learning to interpret images, space and movement. Given an instruction in plain language, such as "tidy this shelf", a robot works out the steps itself. It learns by copying people and practising a task thousands of times in simulations. Driverless cars already carry passengers in some cities; humanoid robots are being tested in factories and warehouses. In the 2040 scenario, machines share their learning: something one warehouse robot learns could be available to every similar robot the next day. Cost and reliability will determine how quickly this happens.

Sensors and cameras give the digital world its senses. Smaller and more affordable, they measure temperature, vibration, movement and congestion. A digital twin is a virtual copy of a physical asset, such as a machine, building, electricity grid or port, continually updated with these measurements. AI spots patterns that reveal when a pump may falter or where congestion will build. Grid operators and factories already use it. By 2040, digital twins could connect entire supply chains, provided measurements are accurate and organisations share data. Data becomes a raw material: ownership and willingness to share it help determine what a connected supply chain can achieve.

Biotechnology uses living cells and their DNA to make or repair things. CRISPR lets researchers edit specific stretches of DNA, work that earned Emmanuelle Charpentier and Jennifer Doudna the Nobel Prize in Chemistry. Regulators including the FDA have approved the first CRISPR treatment, for an inherited blood disorder. In bioreactors, yeast and bacteria produce food proteins, including milk protein without a cow. AI accelerates this work by predicting protein structures and suggesting molecules for researchers to test in laboratories. In the 2040 scenario, these applications help more people stay healthy for longer. Progress depends on research, price and what society considers ethically acceptable.

The energy system of 2040 draws on many sources. Solar and wind depend on weather; batteries balance fluctuations. New drilling techniques make geothermal energy accessible in more places. Smart grids measure supply, switch automatically and shift consumption to times of plenty. AI forecasts solar and wind output and tomorrow's demand. Demand is rising, with the International Energy Agency expecting sharp growth in data centres' electricity use. Technology companies are already signing contracts for small modular nuclear reactors intended to power data centres from the 2030s. Batteries, grids and turbines need metals such as lithium and copper. AI is helping search for materials that could replace them.

Technology has become geopolitical. The most advanced chips come from a few manufacturers; the largest AI models come from a handful of providers. Governments restrict chip exports and want their own data centres, models and access to raw materials. This ambition is known as digital sovereignty. The European Commission invests in AI factories, and the EU AI Act sets rules for AI development and use. Other power blocs make different choices. In the 2040 scenario, several technology blocs coexist, each with its own chips, models and rules. Organisations operating across borders move between these different worlds.

In many countries, the number of older people is growing faster than the number of people of working age. The United Nations and OECD expect further ageing and, in many places, fewer working-age people. Employment levels also depend on participation and policy. Healthcare, engineering, education, logistics and government compete for the same workers. Automation becomes less about saving money and more about getting the work done. AI handles administration, robots lift and drive, and people stay in the workforce longer and increasingly work part-time. Customers change too: older populations need different products and services. In this 2040 scenario, an organisation's capacity depends more on how people, AI and robots work together than on how many applicants it attracts.

Major changes often happen where different forces meet. In this scenario, AI designing molecules and robotic laboratories testing around the clock could shorten the search for new materials and potential medicines from years to months. Sensors, AI and care robots could help many more older people live safely at home in 2040. A grid that balances itself could coordinate solar, wind, geothermal and nuclear power. Machines with a brain take on heavy and repetitive work, leaving people more time for what they do well: judging, caring and designing. Nobody knows which combination will break through first. The building blocks for 2040, though, are already here.