Methodology

What the site measures

GINC Trends follows 250 trends: slow, structural forces that shape national power, prosperity and resilience. For each one the site publishes two weekly indices, kept apart:

• The Trend Attention Index looks back. It measures how much people read and search about a trend, compared with the other trends and with its own past.
• The Trend Expectations Index looks forward. It reads prediction markets for what traders expect to happen to the trend.

Attention is not strength. A trend can deepen for years with little notice, and one news story can lift attention for weeks. The indices say where interest and expectations are moving, not how much a trend matters.

How the list was chosen

The list is reviewed once a year and then held steady, so that trends can be followed over time. The list for 2027 was built in October 2026.

1. Candidates. Four independent AI panels, each using a different model (Anthropic, DeepSeek, OpenAI and Mistral), proposed 300 candidate trends. The candidates were balanced across the nine domains of the GINC National Capability Framework.
2. Cross-check. The candidates were tested against 27 trend and risk reports published in 2025 and 2026, listed below.
3. Expert review. People rated each candidate one by one and cut the list to 275. A final pass reached 250, removing overlap, sharpening definitions and keeping only trends with a sustained, measurable signal.

Of the 250, 199 carry over from the previous list, 150 of them under new names, and 51 are new. Of the 32 trends dropped, the 25 closest to the cut are kept on a private watchlist and measured each week, so a sharp rise is noticed. A renamed trend keeps its page and address.

Each trend is also graded against the 20 GINC Scenarios, and linked to the trends most like it. The links make the trend network, whose clusters are the themes.

The Trend Attention Index

Sources.

• Wikipedia page views in 28 languages.
• Google search interest worldwide and in 48 countries, with Google search volumes behind the share of attention.

Each trend has its own set of search terms and pages, chosen to capture it broadly while staying specific to it. Terms driven by a single news event or by another meaning of the words are reviewed and replaced. Sources for China and Russia are being added, where Google and Wikipedia reach few people.

From sources to a score. Each source is set against the typical trend that week. This removes what moves every trend at once, such as Wikipedia’s long decline in readers. The sources are then combined into one score. An index of 100 is the typical trend, 200 is twice its attention and 50 is half.

Places. Each country is read from its own searching and the languages mostly read there. A region or bloc is the average of its member countries, each counting equally, so no group stands in for its largest member.

History. Every trend has three years of weekly history, recalculated with the current method at every run, so a trend can always be compared with its own past.

The Trend Expectations Index

The Expectations Index reads three prediction markets: Polymarket, Kalshi and Manifold, with Manifold at half weight.

1. Mapping. Each market question is matched to the trends it might bear on. Three AI judges then work blind. A question counts towards a trend only when at least two of them agree it does, and in which direction.
2. Reading. A trend’s expectation is the market-implied chance, from 0 to 100%, that the trend strengthens. It is a weighted average of the prices of its live questions.
3. Coverage. Every reading is published with the number of live questions behind it. A trend with no live questions has no reading, rather than a reading of nought.

The index is a prototype.

The four-week rolling average

Both indices are published as four-week rolling averages. Each week’s figure is the average of that week and the three weeks before it. This is the index itself, not a smoothing added to the charts: every figure, ranking, chart and download on the site is worked out from it.

Why four weeks:

• Attention moves on a cycle of weeks, not days. A major event, such as riots in Paris, a cyber attack on a hospital or a surprise election result, does not come and go in a week. It runs through the news cycle for several weeks: first reports, then analysis, follow-ups, hearings and reactions. People keep reading and searching about it for as long as it stays in their minds. A four-week window follows that whole cycle, not one slice of it.
• Information spreads unevenly. Searching often comes first, reading follows, and markets reprice as new information arrives over days and weeks. The sources rarely move in step within a single week; over a month they tell a more consistent story.
• A month suits slow forces. The trends change over years. A month is the shortest span over which a change in attention says something about the force, rather than about one week’s headlines.
• The calendar adds noise of its own. Public holidays, school terms and an event that falls either side of the weekly cut-off can move a single week for reasons unconnected with the trend.
• The latest week is a little short. We publish on the Saturday, before the week has quite ended. Averaged with the three weeks before, a short week cannot swing the figure.
• Rankings should not reshuffle by chance. Ranked on single weeks, the list of fastest-rising trends changes largely at random from one week to the next. Ranked on four-week averages, it holds steady, while a genuine burst of attention still shows in the week it happens.

We tested longer windows too. Six or eight weeks are steadier still, but they show a burst a week or two late and blunt it.

Reading change

Every change on the site compares two four-week averages.

• Week on week: this week’s four-week average against last week’s. It is not the latest four weeks against the four weeks before.
• Month on month: this week’s four-week average against the one four weeks earlier.
• Quarter on quarter: this week’s four-week average against the one 13 weeks earlier.
• Year on year: this week’s four-week average against the one 52 weeks earlier.

For a week or a month, the level shown is this week’s four-week average. For a quarter or a year, it is the average across that period. The same rule applies to the Expectations Index, where change is shown in percentage points.

The week

Each week runs from 00:00 GMT on Sunday to 23:59 GMT on Saturday, and is named by its Saturday: “week ending 31 October”. We publish on the Saturday, so the latest week holds the data available when the run is made, usually a few hours short of the full week.

This does not make the latest week look weaker. Each trend is compared with the typical trend in the same short week, not with a full one.

Comparing over time

Figures published before 24 October 2026 used the earlier list and search terms. Compare them with later figures only for trends that continue, and with care. Figures published before 31 October 2026 were not four-week rolling averages, so week-on-week changes from before that date are not comparable with those after it.

Limitations

• It leans to the connected. Wikipedia and Google are thin where other platforms dominate or access is restricted. China cannot yet be read, and Russia rests on Russian-language reading alone.
• Search terms are a judgement. A term can be searched for reasons unconnected with the trend. The terms are reviewed, but not exhaustive.
• Markets are thin on some trends. An expectation resting on two or three questions can move sharply when one of them settles.
• Grades and estimates are provisional. The scenario grades and the estimates of likelihood and impact come from AI models working to a rubric, and will be replaced by expert judgement.

Reports consulted

These are the 27 trend and risk reports the candidate list was tested against.

• Accenture Strategy. Macro Foresight Brief: 2026 Outlook and Top 10 Macro Trends for Companies. 2025.
• BlackRock Investment Institute. Mega Forces. 2026.
• Capgemini. TechnoVision: Top Tech Trends of 2026. 2026.
• Control Risks. RiskMap 2026: Top Risks. 2025.
• Control Risks. RiskMap 2027 (preview themes). 2026.
• Deloitte. Tech Trends 2026. 2026.
• The Economist. The World Ahead 2026. 2025.
• Eurasia Group. Top Risks 2026. 2026.
• European Commission. 2025 Strategic Foresight Report: Resilience 2.0. 2025.
• EY. Futures Reimagined: EY Megatrends 2026 and Beyond. 2026.
• Gartner. Top Strategic Technology Trends for 2026. 2025.
• International Monetary Fund. World Economic Outlook, October 2026. 2026.
• Ipsos. Ipsos Global Trends, 9th edition (2025) and 10th edition (2026).
• McKinsey & Company. McKinsey Technology Trends Outlook 2026. 2026.
• MIT Technology Review. 10 Breakthrough Technologies 2026. 2026.
• Morgan Stanley Research. Key Investment Themes for 2026 and 10 Thematic Predictions. 2026.
• Munich Security Conference. Munich Security Report 2026, with the Munich Security Index 2026. 2026.
• OECD. OECD Economic Outlook, Volume 2026 Issue 1. 2026.
• OECD. OECD Science, Technology and Innovation Outlook 2025. 2025.
• Stockholm International Peace Research Institute. SIPRI Yearbook 2026. 2026.
• Temasek. Temasek Review 2026. 2026.
• UK Ministry of Defence, Development, Concepts and Doctrine Centre. Global Strategic Trends: Out to 2055. 2024.
• UN Department of Economic and Social Affairs. World Economic Situation and Prospects 2026. 2026.
• UN Economist Network. Shaping the Trends of Our Time. 2020.
• US National Intelligence Council. Global Trends 2040: A More Contested World. 2021.
• World Economic Forum, with Marsh McLennan and Zurich Insurance Group. Global Risks Report 2026. 2026.
• World Economic Forum, with Frontiers. Top 10 Emerging Technologies of 2026. 2026.

Downloads

Everything on the site can be downloaded: the list, its scores against every scenario, the trend network and the weekly attention data for every place. The files are listed below.

Licence

Content and data are published under CC BY 4.0. You may share and adapt them for any purpose, with attribution. Code is published under the MIT licence. Wikipedia page view counts are released by the Wikimedia Foundation under CC0. Google Trends figures may be reused with attribution to Google Trends.

Citation

• The list. GINC (2026). GINC Trends v2.0: 250 trends scored against 20 scenarios. Oxford: Global Institute for National Capability. trends.ginc.org
• A trend. GINC (2026). Tariff barrier normalisation, GINC Trends v2.0. trends.ginc.org/trends/tariff-barrier-normalisation
• Attention data. GINC (2026). Trend attention index v1.2 [dataset]. Oxford: Global Institute for National Capability. trends.ginc.org/methodology/#downloads