2026: Growth under hard limits
· 19 min read
Growth will stay positive, but the political environment is fragile. This year will feel tighter than most organisations are prepared for. The key issue is how many challenges arise simultaneously and how quickly systems can adjust when there is little room for error.
AI investment continues to progress, despite infrastructure bottlenecks. The primary constraint is the time required for new projects to connect to the grid, so delivery speed is now a key advantage. In many regions, large projects still take years rather than months to connect, so investors look for places that can move faster. Automation is changing job quality in services and logistics. There has not been a single big labour shock, but pressure is rising as entry-level jobs become harder to find and personal interaction becomes more expensive.
Chokepoints and dependencies are intensifying geopolitical tensions. Energy corridors, refined materials, subsea cables, and Taiwan’s role in semiconductors are now central issues. Ongoing conflicts in Ukraine and the Middle East, along with disputed maritime routes, keep resource security, naval shipping, and semiconductor supply at the top of national agendas. With little slack in the system, countries focus on controlling key inputs and routes to build resilience and maintain leverage.
Financial structure is now a key factor. AI infrastructure, power grids, and chip production rely on private credit, specialist funds, and long-term contracts. The question for 2026 is whether this growth can continue without a major credit crisis that necessitates emergency support and underscores how much risk has shifted beyond traditional banks.
By the end of the year, three clear signals will be most important. First, do grid connection times for large projects get shorter? Second, can the next disruption in minerals, shipping, or chips be managed without a policy crisis? Third, do governments respond to social tensions by investing in capacity, skills, and trust, or do they rely more on digital controls because they cost less than real-world delivery?
These are twelve areas where a bad month can rewrite the year.
Geopolitical intensity: How much state-level competition or conflict shapes this area. 3 = actively contested by major powers; 2 = geopolitical pressure present; 1 = largely outside great power friction.
Systemic interconnection: How fast stress spreads to other sectors. 3 = disruption cascades rapidly; 2 = linkages exist but containable; 1 = problems stay local.
The two are independent; an area can be high on one and low on the other. The scores are judgements, not moods.
A reasonable outlook for 2026 is steady but moderate growth, though resilience remains limited. Global growth is slightly above 3 per cent. Advanced economies hover near 1.5 per cent, while emerging markets and much of the Global South are just over 4 per cent.
That divergence is K-shaped in practice: the headline number holds up while affordability, access to services, and institutional capacity continue to diverge in opposite directions across countries.
The US is set for slow expansion rather than a boom. The UK and the euro area are likely to experience low growth rather than a strong recovery. Germany still faces periods of stagnation, as high energy costs and borrowing limits constrain infrastructure investment.
A poor harvest across multiple central basins, disruption through a contested corridor, a regional grid failure, or a sudden tightening of credit can knock the year off course, and higher debt and refinancing costs leave governments with less room to respond.
Services and housing costs remain stubborn, so political tensions stay high even if headline inflation does not spike again. In many G7 countries, core inflation ends the year above target. Central banks try to keep policy tight enough to avoid a setback, but not so tight that they stall already weak growth. The primary debate shifts to real wages, housing, and affordability, rather than declaring that inflation has been beaten.
Food and water costs quickly consume a larger share of household budgets. In many places, food prices rise faster than wages, and a poor season can push tens of millions into severe food insecurity in emerging economies. Developed countries experience the initial effects, including price spikes and public frustration. Lower-income countries face budgetary stress, emergency imports, and local unrest, which in turn exacerbate migration and border issues in other regions.
People are also a limiting factor. Upgrades to grids, housing, water systems, and data centres all need the same electricians, engineers, planners, and builders. When politics restricts the labour pool or migration, wait times increase, costs rise, and projects fall further behind.
Electricity planning now treats data centres as a core part of demand, not just an afterthought. Demand is concentrated on certain substations and corridors that were not designed for rapid growth. By 2026, data centres, AI, and crypto could consume about 1,000 terawatt-hours of electricity, or nearly 4 per cent of global demand. In Europe, data centre consumption may be 30 per cent higher than in 2023.
In Northern Virginia, Dublin and Frankfurt, connection queues are long enough to change corporate strategy. Developers shuffle sites, phase plans, and sit on partially utilised assets while reinforcement catches up. Idle campuses are not a curiosity. They are a capital allocation signal that tells you where the next wave of compute will not land, regardless of ambition.
Queues form because companies act logically. When a queue starts, firms join early, thereby lengthening and making the line more visible. Equipment such as transformers, switchgear, cables, and power electronics can take years to produce, particularly at high voltages. Skilled workers are also in short supply and cannot be added quickly. In this situation, increasing expenditure does not expedite the process. It only secures a spot in the queue.
Companies look for workarounds. On-site power, storage, and hybrid systems appear attractive when grid connection takes years. Some projects will work if permits are easy to get and grid services can be sold. Emissions regulations, procurement delays, equipment lead times, or local resistance will delay many. The fastest solutions often use fossil fuels. Cleaner options usually take longer. Gas turbines are not a quick fix, as equipment lead times range from 4 to 7 years.
Engineering limits are closely tied to financial limits. If lenders perceive greater risk or reduce funding, transformer orders are delayed, contractors request tougher terms, and utilities slow upgrades because the numbers no longer work. A problem in private credit can quickly become a problem for the financial system, sometimes within a few months. This can leave assets unused, even if they seemed safe on paper.
Efficiency improvements do not solve the problem. Cheaper computing leads to more demand. Longer context windows, heavier workloads, always-on systems, and broader use all use up the gains.
Energy prices are moving in different directions for households and industry. In many countries, policies help shield households from the full impact of rising costs. In contrast, industries experience greater price volatility and higher ongoing expenses. This makes it harder for them to compete and forces difficult choices, especially in energy-intensive sectors.
Oil prices still help ease some of the political pressure from inflation, since people feel fuel costs right away. Policymakers seek to prevent persistent price spikes by employing diplomacy, drawing on reserves, or encouraging local production. These steps can stabilise the situation in the short term, but they do not address the underlying issues. Holding prices down for too long can make it harder to adjust later.
A quiet surcharge persists through shipping and insurance. Elevated risk in contested corridors such as the Red Sea and key choke points in the Taiwan Strait, Gulf and around the Strait of Hormuz keeps premia above pre-2023 levels and encourages rerouting and inventory buffers. The effect is not always dramatic in any one month. It is cumulatively inflationary and operationally slowing.
Problems such as heat, drought, and water shortages do not affect only one area. They impact both farming and electricity production. When there is insufficient water for cooling or hydropower generation, energy supplies can become less reliable. If a region faces a water crisis, it can lead to higher food prices, power outages, and additional costs for governments, thereby slowing economic growth.
Refined materials and specialised equipment remain the other boundary. The build-out of grids, EVs, batteries, and data centres depends on concentrated processing and manufacturing capacity, and this concentration is strategic because it increases the leverage of those upstream activities.
Subsea cables, interconnectors, and orbital systems provide navigation, timing, communications, and weather data. These systems form the backbone of domestic grids. Our reliance on them is significant but often goes unnoticed.
Cable faults occur regularly, and there are insufficient resources to repair them promptly. Most companies have less route diversity than they claim. When multiple faults occur in the same area, such as in Asia or the Middle East, the effects include slower cloud services, payment delays, operational problems, and increased insurance costs.
Space is now an area of active development. The United States and China are treating cislunar space and the lunar surface as critical strategic locations. The Artemis program aims for a crewed lunar flyby in 2026 at the earliest, with a South Pole landing planned later. China and its partners are developing the International Lunar Research Station and plan to establish a nuclear-powered base at the lunar south pole in the 2030s. These bases will use nuclear and solar power to support extended stays and industrial work.
Every heavy launch adds more sensors, communications equipment, and dual-use technology to cislunar space. Defence planners in both countries now see space as an active military domain, not just a shared resource.
As more objects are launched, orbital congestion and debris become increasingly serious problems. Satellites require more avoidance manoeuvres, which shorten their lifespans and increase costs. If positioning and timing signals are disrupted, it can affect logistics, ports, and grid systems that depend on them. People talk about resilience, but it is not always put into practice.
International rules are slow to change. As our reliance on these systems grows faster than our ability to prepare for them, the risk to modern operations increases.
Ongoing conflicts keep defence spending high. This results in less funding for power grids, water, housing, and climate projects. Budgets cannot cover all priorities, so trade-offs are unavoidable.
Conflicts now focus more on who controls energy, raw materials, and key supply routes. Oil and gas still provide countries with short-term leverage; minerals and processing matter for keeping industries running in the medium term. Semiconductors are crucial for secure communications, intelligence, and automation.
Ukraine is still fighting a large-scale industrial war using artillery, drones, and long-range strikes. As 2026 begins, Western support remains uncertain. In Gaza, a fragile ceasefire holds after almost two years of conflict. Winter floods tents and shelters while talks about the future continue.
Yemen and the Red Sea remain hotspots of proxy conflict. Shipping risks remain high, affecting insurance and routes even during periods of relative calm. The Venezuela–Guyana dispute over the Essequibo region is another resource-driven flashpoint. The US uses embargoes and shipping rules to control Venezuelan oil, aiming for both lower fuel prices and political leverage. Other oil-rich countries see this as a sign that access can be conditional, so they hedge their bets and trust between partners weakens.
Taiwan matters because it produces most of the world’s advanced chips. This is not just about territory or symbolism. Countries depend on these chips for national power. When alternatives are scarce and supply chains are disrupted, the risk of losing access shapes military planning and alliances.
The US is returning to a Monroe Doctrine approach, but China’s global influence means the Western Hemisphere is no longer isolated.
Foreign policy depends on how well domestic systems work. People are more willing to support international commitments when domestic affairs run smoothly. If infrastructure fails, public support for foreign policy drops.
Export controls and investment screening have shifted from short-term measures to long-term policies. This shift leads to more fragmentation, repeated efforts, higher costs, and slower spread of new technologies.
Both sides still rely on each other. China lacks access to the most advanced tools in certain areas. The United States and its partners depend on China for early-stage processing and concentrated supplies of materials used in clean technology and electronics. China accounts for most of the world's lithium, cobalt, graphite, and other battery minerals, often managing 70 to 90 per cent of refining and midstream processing for these materials. It also leads in lithium-ion cell manufacturing. When tensions rise, both sides typically target specific areas rather than escalating across the board, since a full escalation would harm both industries and disrupt supply chains with few alternatives.
Countries continue to build up their own capabilities where possible. Even if China is not leading in the most advanced areas, being good enough still counts. Many real-world applications depend on inference, and these applications are often constrained by cost and power constraints.
In November 2025, China’s industrial profits dropped by 13.1 per cent compared to the previous year. This was the most significant decline in over a year. Because of overcapacity, China is more likely to increase exports and bundle products together.
China’s strength comes from its integrated system. This includes its ability to build power infrastructure, its large manufacturing base, and its deep supply chains. In a year with many physical constraints, the ability to execute is as important as generating new ideas.
Inside China, competition is fierce, and profits are low. This drives companies to export more and secure demand from other countries.
Combining energy and digital systems is effective because many countries in ASEAN and the Global South prefer to procure infrastructure as a single package. They want one financing plan, one delivery schedule, and one system to operate. Western companies often split these parts across different groups, but a bundled approach is faster and more precise. Chinese companies lead in the production of solar panels, batteries, and grid equipment, and often hold more than 70-80 per cent of the market in key areas.
Tariffs and checks can slow down imports and give local companies more time, but they do not quickly create strong regional competitors. Replacing imports in areas such as refining, heavy equipment, and clean technology manufacturing takes years because it requires permits, skilled workers, and substantial investment.
Mid-tier countries are developing their own AI and data infrastructure to remain resilient, as technology is fragmenting into distinct camps. The main challenges are securing sufficient grid access, finding skilled workers, waiting for equipment, and completing large projects on time.
Being early in line matters as much as having money. You cannot skip the wait for transformers, chips, or site approvals just by spending more. If you order first, you can start sooner. If you wait, your project might not be ready before the political support runs out.
The risks are both financial and political. If a country spends billions on its own computing but then finds out it cannot use it for years because of power or grid delays, the unused equipment looks wasteful to voters and risky to investors. When money is tight, this can make borrowing more expensive, necessitate cuts in other programs, and increase political tensions. The real problem is spending on flashy projects before fixing the basics, which ends up costing more and hurting trust.
A simple way to track this is to compare the amount spent on computing with the available grid- and site-level power. If the budget for computing is more than about three times what is funded and ready for power, the project risks ending up with expensive equipment that cannot be used yet. The money is spent and visible, but the results are delayed.
Nuclear power reenters this picture as an enabling asset. Tech firms enter into long-term power purchase agreements to restart or extend existing reactors, such as the 20-year deal to revive Three Mile Island Unit 1 to supply Microsoft, and similar contracts that support other US nuclear plants for AI workloads. These deals are early templates for a world in which sovereign stacks and hyperscale compute treat nuclear output as strategic baseload. At the same time, new-build projects and small modular reactors remain further out, beyond most 2026 political cycles.
Critical systems are adopting AI that can operate autonomously, albeit at different speeds. In the West, people still oversee key decisions because high confidence in AI does not always mean it is correct, and mistakes can incur substantial legal or political costs. China is deploying AI agents more rapidly in areas such as grids and logistics, which can boost efficiency but also pose greater risks if things go wrong. Automated systems can fail en masse, not just in small ways. If the AI has the wrong goal, the data changes, or someone hacks the system, it can cause a sudden, major problem. Having humans involved slows things down, but it helps catch errors before they spread.
AI first appears in places with sufficient money and resources to make it work. Education and healthcare are where AI can do the most good, but they are also the hardest to manage fairly. AI tutoring and adaptive learning tools help students perform better, especially in mathematics and language. Still, these tools are primarily available to well-funded schools and universities. This means students who already have advantages receive more personalised support, whereas schools with tight budgets must rely on basic tools that do not address the underlying problems.
In healthcare, AI supports diagnosis, triage, and care coordination, leading to real improvements in areas such as radiology, cancer care, skin conditions, and the management of chronic diseases. However, these benefits are not distributed evenly, as establishing and monitoring these systems requires financial and labour resources. Wealthier healthcare systems adopt AI first and continue to gain an advantage. Tools for assisted living and remote monitoring facilitate older adults’ management of health issues. Still, they are primarily used in Japan, parts of Europe, and wealthier regions of China and North America, where people and organisations can afford them.
The main benefits go to those with money, data, and the ability to use new technology. Downsides such as job loss, increased surveillance, and infrastructure strain are experienced by more people. The real danger is a gradual erosion of essential services, not a single significant crisis.
Automation changes the nature of jobs more than it causes widespread job loss. The first to go are routine tasks done on screens. In many places, customer service, back-office work, and basic logistics are now handled by large language models. As a result, some companies in developed countries hire fewer new front-line workers, since chatbots and workflow tools now perform that work.
Industrial robots are becoming more common. By the mid-2020s, more than 570,000 new robots were installed annually, primarily in automotive factories, electronics plants, and food-processing facilities. Autonomous systems are now used along freight routes and in some city areas. In the US and parts of China, self-driving trucks operate on fixed routes, and in busy cities, delivery robots and small driverless vehicles handle more package deliveries where permitted.
People are willing to pay extra to speak with a real person when they become frustrated. Some banks, utilities, and telecom companies now offer paid plans that guarantee you can reach a human on the phone or in chat, while everyone else waits in AI queues with few options to escalate. Telehealth services also charge more if you want to see a live clinician. This pricing makes sense for businesses, but it creates a divide by making better service available only to those who can pay.
Workers are more likely to push back when automation is introduced if they do not perceive any tangible benefits. Strikes do not have to happen all the time to get attention. They need to be visible and occur in areas such as transport, logistics, or public services, where people notice the disruption and can clearly see that automation is part of the problem.
Rising costs for memory, servers, and other key components are showing up in IT budgets and consumer products. Following sharp price swings in 2024 and 2025, DRAM and NAND prices remain elevated in 2026 due to strong AI and data centre demand. Cloud providers are passing these costs on, thereby increasing the cost of computing and storage. Tariffs and localisation rules are also making it more difficult and more costly to purchase electric vehicles, solar panels, and electronics in the United States and Europe.
How gains and losses are spread is more important than the average. People who own assets or work in AI, defence, or advanced industries are experiencing higher wages and greater capital gains. Those in the middle and lower-income groups feel less secure because a larger share of their income goes to housing, food, and services, and their wages are not keeping up with inflation. Even if the economy grows overall, the recovery is uneven.
Political parties that promise protection through tariffs, migration limits, and economic independence are becoming more popular because they address real concerns. Policies that tax imports, slow some forms of automation, or use AI-generated profits for direct payments are becoming key parts of election campaigns, even if they sacrifice long-term efficiency for short-term benefits.
When it takes too long or costs too much to deliver services in person, governments turn to digital governance instead. Regulating information and behaviour online is less expensive than building new housing, health, or transport systems. Many governments exploit this difference.
By 2026, about 30 per cent of new public online content is primarily AI-generated, with the visible share in mainstream feeds nearer one in five.
As synthetic media grows and trust falls, platforms become more cautious in how they moderate content. Governments respond with new rules focused on safety and disinformation. Together, these changes make people less likely to speak freely, even though there is no direct censorship. Uncertainty and automated systems remove more content, make it harder to discuss sensitive topics, and push people toward official views.
A new economy based on verified human activity starts to take shape. In-person events, physical credentials, and methods of verifying identity become increasingly important as indicators of authenticity. Sectors such as finance, healthcare, legal services, and critical infrastructure are increasingly requiring evidence that a human, not an automated system, made key decisions. The value of human involvement shifts from a marketing point to a formal requirement.
Debates about jobs bring fiscal policy into focus. Some governments are beginning to test small taxes on AI or automation, using the revenue to fund retraining, wage insurance, or support for front-line services. These early programs are limited, but they indicate that many view AI as distinct from past technologies and expect new policies to address its impact.
In 2026, results will matter more than talk. Grid queues, transformer lead times, permitting, water, and skilled labour will determine what gets built and where. These factors will show who is prepared to deliver and who is only funded.
Geopolitics focuses on the same bottlenecks: corridors, cables, refining, and chips. These are slow to replace and can be used as leverage. Finance is the key risk. If private credit becomes more difficult to access, projects will not collapse; rather, they will face delays and higher costs.
By the end of the year, three indicators will be of particular importance: first, whether grid connection times decrease; second, whether the next disruption in minerals, shipping, or chips remains a logistics issue; third, whether governments address stability by increasing capacity or by using digital controls to manage frustration as inequality rises.
In a world with little slack, those who reduce waiting times will succeed. Others will remain in line and call it policy.
illuminem Voices is a democratic space presenting the thoughts and opinions of leading Sustainability & Energy writers, their opinions do not necessarily represent those of illuminem.
Track the real‑world impact behind the sustainability headlines. illuminem’s Data Hub™ offers transparent performance data and climate targets of companies driving the transition.
illuminem briefings
AI · Corporate Governance
illuminem briefings
AI · Ethical Governance
illuminem briefings
Corporate Governance · AI
Axios
AI · Corporate Governance
CNN
AI · Human Rights
The Guardian
Human Rights · AI