A 45%
Managed acceleration
AI delivers much of its promise, energy infrastructure gets unblocked and fragmentation is managed.
3.0–3.5%
Insights · Strategic foresight 2030
The world in 2030: 12 insights to act on today.
Artificial intelligence, geopolitics and energy will reshape business this decade. These are the core findings of “What will the world look like in 2030?”, our strategic foresight report, with the data behind them.
Between 2026 and 2030 the world will produce far more intelligence at near-zero cost, far more energy at uneven cost, and much less free trade. The winning company will be the one that secures energy, controls data no one else has and rebuilds its pricing model before the market rebuilds it for them.
The numbers to watch
Six figures that frame the decade.
- 945 TWh
- projected data center electricity consumption in 2030, up from 415 TWh in 2024 (IEA)
- 31%
- of companies have at least one AI agent in production (market surveys, 2026)
- 88%
- of AI agents that are developed never reach production
- >90%
- of the world’s leading-edge chips are made in Taiwan
- 91%
- of global rare earth separation and refining takes place in China (IEA)
- USD 2.9 trillion
- in projected global military spending by 2030, up from USD 2.6 trillion in 2026
Executive summary
Twelve findings about the world in 2030.
- 01
The constraint of this decade is the electron, not capital or talent.
The IEA projects data center electricity consumption to rise from 415 TWh in 2024 to about 945 TWh in 2030, growing close to 15% a year. The bottleneck is not generation but connection: in the United States interconnection queues reach eight years and large power transformers take three to four years to deliver.
- 02
AI has moved from a technology debate to a cost-structure debate.
The marginal cost of standardized cognitive work —drafting, summarizing, classifying, translating, reconciling— is heading toward zero, and its market price will follow. If a company’s margin depends on selling hours of that work, it has a business-model problem, not a productivity problem.
- 03
Real adoption lags far behind the narrative.
According to 2026 market surveys, around 31% of companies have at least one AI agent in production, and about 88% of the agents being developed never reach production. The gap between companies that industrialize AI and those piling up pilots is today the largest source of competitive advantage available.
- 04
There is a real financial risk, and it is already formally recognized.
The Bank for International Settlements added the sustainability of AI investment to its register of systemic risks. J.P. Morgan estimated the industry needs around USD 650 billion in new annual revenue just to earn a 10% return on the infrastructure being built. Plan on the assumption that the technology keeps advancing and financing gets more expensive.
- 05
Global trade is shifting from “efficient” to “redundant”, and that is inflationary.
Tariffs, export controls, local-content rules and duplicated supply chains raise production costs permanently. The world is moving toward multi-hub supply chains, and redundancy adds between 5% and 15% to unit costs.
- 06
Critical minerals are the asymmetric weapon of the decade.
China accounts for about 60% of global rare earth mining and about 91% of separation and refining. Since January 1, 2026, any product made outside China containing 0.1% or more of Chinese-origin rare earths requires a license; the IEA estimates the rule reaches value chains worth around USD 6.5 trillion.
- 07
Taiwan is the only truly systemic single point of failure in the global economy.
More than 90% of leading-edge chips are made there. TSMC projects a global semiconductor market above USD 1.5 trillion by 2030, with AI and high-performance computing accounting for 55%.
- 08
Global rearmament is a ten-year cycle, not a one-off event.
Global military spending will exceed USD 2.6 trillion in 2026 and is projected to reach about USD 2.9 trillion by 2030. NATO allies committed to spending 5% of GDP on defense by 2035, including 1.5% on critical infrastructure and the industrial base.
- 09
Energy is moving at two speeds.
Solar and wind exceeded 800 GW installed in a single year, and battery storage is set to grow seventeenfold by 2030 according to BloombergNEF. At the same time, oil is plateauing, not collapsing: the transition is additive before it is substitutive.
- 10
Demographics are the silent constraint almost no one puts in the plan.
There will be around 1.4 billion people over 60 in 2030. The result is a structural shortage of trades, care, construction and logistics workers, alongside a surplus of desk work: too many keyboards, not enough hands.
- 11
Value concentrates at the extremes and the middle empties out.
Global scale or deep niches with proprietary data win. Generic services, horizontal software without exclusive data, unbranded retail and undifferentiated consulting get stuck in the middle, where AI compresses prices and global scale compresses costs.
- 12
For SMEs, the asymmetry flips for the first time in thirty years.
Capabilities that required a 200-person team five years ago can now be rented by subscription. Competitive advantage has become decision speed, and a well-run SME consistently beats the corporation there.
Analytical framework
Three forces, six friction zones.
Artificial intelligence, geopolitics and energy are not parallel phenomena: they behave like tectonic plates. What shapes the economic landscape is not how each one moves, but what happens where they collide.
AI × Energy
The price of compute becomes the price of electricity.
Training frontier models could require between 4 and 16 GW per run by 2030, according to Epoch AI: the power contract becomes the tech industry’s number-one strategic input.
AI × Geopolitics
Compute is sovereignty.
Export controls and cross-restrictions between the United States, China and Taiwan decide who can train what. Choosing a cloud and a model now carries country risk for the first time.
Energy × Geopolitics
Critical minerals are the new oil.
The top three producers of copper, lithium, cobalt, graphite and rare earths control 86% of the market, up from 82% in 2020. And there is no deep spot market or meaningful strategic reserve.
AI × Work × Demographics
Shortage and surplus at the same time.
Too many analysts, not enough electricians. The World Economic Forum estimates 92 million jobs displaced and 170 million created by 2030: the problem is not the net figure but that they are different people, places and skills.
Energy × AI × Politics
The power-bill backlash.
When bills rise next to a data center, politics reacts. It is probably the most underestimated regulatory risk of the decade: AI regulated not for what it says, but for what it consumes.
Geopolitics × Capital
The end of easy regulatory and tax arbitrage.
Local content, foreign investment screening, global minimum taxes and secondary sanctions: corporate structure and asset location become strategic decisions again.
Scenarios
Three scenarios for 2030, with probabilities.
They are not predictions: they are frameworks to test whether a decision holds up in more than one possible world. Probabilities are subjective, assigned by the Analytics Town and ATOWNx team.
B 35%
Prolonged friction
Technology advances but infrastructure and financing fall behind. A market correction in AI.
2.3–2.8%
C 20%
Rupture
A major shock: the Taiwan Strait, a cut-off of critical minerals or a systemic accident in financing.
< 1.5%
Eight “no-regret” moves that work in all three scenarios
- Secure long-term energy
- Reduce single-supplier dependence
- Build and protect proprietary data
- Lower the operating break-even point
- Strengthen cash and extend debt
- Industrialize two or three processes with AI
- Train and retain senior talent and trades
- Instrument outcome measurement
Winners and losers
Where margin is moving.
The question is not which industries will grow —almost all grow in nominal terms— but where demand grows faster than supply capacity. Ranked by relative margin attractiveness and urgency of response.
Benefit
- 01Electric power
- 02Grid infrastructure
- 03Semiconductors
- 04Data centers
- 05Defense and security
- 06Critical minerals
- 07Health and longevity
- 08Vertical software
- 09Agriculture, food and water
- 10Multi-hub logistics
- 01BPO and contact centers
- 02Hourly-billed services
- 03Media and advertising
- 04Per-seat SaaS
- 05Legacy automotive
- 06Unbranded retail
- 07Power-intensive manufacturing
- 08Fuel downstream
- 09Heavy retail banking
- 10Traditional education
Main sources cited in the report: International Energy Agency (IEA), International Monetary Fund, BloombergNEF, Bank for International Settlements, NATO, World Trade Organization, TSMC, Epoch AI and World Economic Forum, among others. Projections to 2030 are estimates subject to revision: read them as orders of magnitude and directions of change, not point forecasts.
Special report · Free download
What will the world in 2030 look like?
Strategic foresight on the three forces that will reshape business this decade: artificial intelligence, geopolitics and energy. What will change, who wins, who loses — and what business leaders should do today.
78 pages14 infographics100-day plan
FAQ
What people usually ask.
01 What is the main constraint for businesses through 2030?
Connected electric power. The IEA projects data center consumption to rise from 415 TWh in 2024 to about 945 TWh in 2030, and the bottleneck is not generation but connection: interconnection queues of up to eight years in the United States and three- to four-year lead times for transformers.
02 What scenarios does the report set out for 2030?
Three: managed acceleration (45% probability, 3.0–3.5% global growth), prolonged friction (35%, 2.3–2.8%) and rupture (20%, below 1.5%). These are subjective probabilities assigned by Analytics Town and ATOWNx, not predictions.
03 Which industries benefit most through 2030?
In order: electric power, grid infrastructure, semiconductors, data centers, defense and security, critical minerals, health and longevity, vertical software, agriculture, food and water, and multi-hub logistics.
04 Which industries are hit hardest?
In order: BPO and contact centers, hourly-billed services, media and advertising, per-seat SaaS, legacy automotive, unbranded retail, power-intensive manufacturing, fuel downstream, heavy retail banking and traditional education.
05 What should companies do today?
Start with the eight “no-regret” moves, which improve a company’s position in all three scenarios: secure long-term energy, reduce single-supplier dependence, build and protect proprietary data, lower the operating break-even point, strengthen cash and extend debt, industrialize two or three processes with AI, train and retain senior talent and trades, and instrument outcome measurement.
06 How will AI affect jobs through 2030?
AI reduces demand for junior cognitive work just as demographics reduce the supply of skilled manual labor: too many analysts, not enough electricians. The World Economic Forum estimates 92 million jobs displaced and 170 million created by 2030; the issue is not the net figure but that they are different people, places and skills.
Let's talk
Ready to deploy your AI special forces?
Tell us what you want to solve. We reply within 24 business hours with a concrete proposal for next steps.