2026-01-23 02:05:36
Data centers have quietly become some of the most important infrastructure in the U.S. economy. As artificial intelligence (AI) workloads explode and cloud services proliferate, builders are racing to add capacity at record speed.
This graphic, in partnership with BHP, shows U.S. data center construction spending from 2014 to 2025 using data from the U.S. Census Bureau.
Here is a table showing monthly U.S. data center construction since 2014, seasonally adjusted at an annual rate.
| Month | Billions of Current U.S. Dollars (Seasonally Adjusted Annual Rate) |
|---|---|
| Jul-25 | 41.19 |
| Jun-25 | 41.10 |
| May-25 | 40.07 |
| Apr-25 | 39.60 |
| Mar-25 | 36.88 |
| Feb-25 | 37.13 |
| Jan-25 | 35.77 |
| Dec-24 | 34.98 |
| Nov-24 | 35.62 |
| Oct-24 | 35.86 |
| Sep-24 | 32.96 |
| Aug-24 | 32.94 |
| Jul-24 | 31.74 |
| Jun-24 | 31.24 |
| May-24 | 29.98 |
| Apr-24 | 28.34 |
| Mar-24 | 27.78 |
| Feb-24 | 26.29 |
| Jan-24 | 24.93 |
| Dec-23 | 24.45 |
| Nov-23 | 23.72 |
| Oct-23 | 23.84 |
| Sep-23 | 21.13 |
| Aug-23 | 20.25 |
| Jul-23 | 19.55 |
| Jun-23 | 18.62 |
| May-23 | 18.11 |
| Apr-23 | 18.16 |
| Mar-23 | 18.10 |
| Feb-23 | 17.87 |
| Jan-23 | 15.85 |
| Dec-22 | 14.35 |
| Nov-22 | 13.77 |
| Oct-22 | 12.88 |
| Sep-22 | 14.47 |
| Aug-22 | 13.02 |
| Jul-22 | 12.71 |
| Jun-22 | 12.15 |
| May-22 | 12.08 |
| Apr-22 | 12.26 |
| Mar-22 | 11.36 |
| Feb-22 | 10.53 |
| Jan-22 | 11.08 |
| Dec-21 | 11.39 |
| Nov-21 | 10.51 |
| Oct-21 | 9.74 |
| Sep-21 | 10.77 |
| Aug-21 | 10.43 |
| Jul-21 | 9.73 |
| Jun-21 | 9.30 |
| May-21 | 9.31 |
| Apr-21 | 9.44 |
| Mar-21 | 10.32 |
| Feb-21 | 9.04 |
| Jan-21 | 9.31 |
| Dec-20 | 9.23 |
| Nov-20 | 9.18 |
| Oct-20 | 9.34 |
| Sep-20 | 10.05 |
| Aug-20 | 9.07 |
| Jul-20 | 8.73 |
| Jun-20 | 8.97 |
| May-20 | 8.70 |
| Apr-20 | 8.95 |
| Mar-20 | 9.49 |
| Feb-20 | 9.68 |
| Jan-20 | 9.46 |
| Dec-19 | 7.94 |
| Nov-19 | 9.57 |
| Oct-19 | 9.16 |
| Sep-19 | 8.08 |
| Aug-19 | 8.82 |
| Jul-19 | 8.82 |
| Jun-19 | 8.88 |
| May-19 | 8.74 |
| Apr-19 | 7.86 |
| Mar-19 | 7.67 |
| Feb-19 | 8.03 |
| Jan-19 | 8.12 |
| Dec-18 | 8.05 |
| Nov-18 | 7.10 |
| Oct-18 | 7.86 |
| Sep-18 | 6.29 |
| Aug-18 | 7.00 |
| Jul-18 | 7.45 |
| Jun-18 | 7.06 |
| May-18 | 6.99 |
| Apr-18 | 6.56 |
| Mar-18 | 6.47 |
| Feb-18 | 6.24 |
| Jan-18 | 6.08 |
| Dec-17 | 5.74 |
| Nov-17 | 5.46 |
| Oct-17 | 4.96 |
| Sep-17 | 4.69 |
| Aug-17 | 4.39 |
| Jul-17 | 4.67 |
| Jun-17 | 4.58 |
| May-17 | 5.03 |
| Apr-17 | 4.23 |
| Mar-17 | 3.99 |
| Feb-17 | 4.22 |
| Jan-17 | 3.99 |
| Dec-16 | 4.46 |
| Nov-16 | 3.96 |
| Oct-16 | 3.70 |
| Sep-16 | 4.22 |
| Aug-16 | 4.58 |
| Jul-16 | 4.25 |
| Jun-16 | 4.56 |
| May-16 | 4.15 |
| Apr-16 | 4.42 |
| Mar-16 | 4.12 |
| Feb-16 | 3.80 |
| Jan-16 | 3.16 |
| Dec-15 | 3.09 |
| Nov-15 | 2.98 |
| Oct-15 | 2.75 |
| Sep-15 | 2.80 |
| Aug-15 | 2.84 |
| Jul-15 | 2.68 |
| Jun-15 | 3.93 |
| May-15 | 3.34 |
| Apr-15 | 2.31 |
| Mar-15 | 2.14 |
| Feb-15 | 1.93 |
| Jan-15 | 1.92 |
| Dec-14 | 1.66 |
| Nov-14 | 1.95 |
| Oct-14 | 1.83 |
| Sep-14 | 1.92 |
| Aug-14 | 1.96 |
| Jul-14 | 1.79 |
| Jun-14 | 1.75 |
| May-14 | 1.81 |
| Apr-14 | 1.80 |
| Mar-14 | 1.68 |
| Feb-14 | 1.74 |
| Jan-14 | 1.64 |
In early 2014, U.S. data center construction ran at an annualized rate of roughly $1.6 billion. By July 2025, that number reached about $41.0 billion, more than 25 times the 2014 level.
Consequently, data centers now compete directly with offices, warehouses, and industrial facilities for land, power, and transmission capacity.
A typical data center uses roughly 27 metric tons of copper per megawatt of capacity for power, cabling, and cooling. Because each new megawatt adds more copper-intensive equipment, incremental AI capacity has a disproportionate impact on metal demand.
As operators add AI servers and dense racks, each data center concentrates more copper in cables, transformers, and cooling infrastructure.
BHP projects that the amount of copper used in data centers globally will increase by around 6x from 2025 to 2050, rising from about 0.5 million tonnes a year to around 3 million tonnes.
That uplift is roughly equivalent to the combined annual output of the world’s four largest copper mines today. As this layer of “digital” demand stacks on top of the broader energy transition, data centers could become one of the fastest-growing sources of structural copper demand.
Looking ahead, the U.S. grid will neeed to keep pace with this construction surge as data centers use more electricity.
Data centers’ share of global electricity demand could rise from about 2% today to roughly 9% by 2050, with some markets already seeing data centers account for around one-fifth of national power use.
Overall, surging U.S. data center investment signals rapid AI infrastructure growth. It also demonstrates the central roles of copper and electricity in enabling the next era of digital services.
Projected electricity usage growth and economic development are driving global copper demand—see how energy use is rising from 2024 to 2035.
Projected electricity usage growth in major economies is driving global copper demand—see how energy use is rising from 2023 to 2035.
Global potash demand is projected to rise 65% by 2050. See what’s driving the surge and why potash is key to global food security.
Since 1960, potash demand has outpaced both population growth and crop production.
Copper discoveries are becoming increasingly rare and often found deeper underground.
Copper demand globally is estimated to rise by 70% from 2021 to 2050. What are the main sources of this increase in demand?
China’s steel demand remains robust, but the breakdown on a sectoral level has shifted since 2010. Which sectors are driving steel consumption?
2026-01-22 21:11:04
See visuals like this from many other data creators on our Voronoi app. Download it for free on iOS or Android and discover incredible data-driven charts from a variety of trusted sources.
In the second half of 2025, 16.1% of the global working-age population used AI, indicating substantial room for further adoption.
At the same time, usage varies widely across countries. Adoption rates average 24.7% in the Global North, while they are 14.1% in the Global South. Key countries stand as clear regional outliers including the UAE and Singapore.
This graphic shows AI adoption by country, based on data from the Global AI Adoption in 2025 report from Microsoft.
For the analysis, Microsoft estimated AI usage across 147 countries worldwide.
Below, we show the share of each country’s working-age population that used AI at least once over H2 2025. Additionally, we show the percentage point (p.p.) change compared to H1 2025, representing the absolute difference between the two periods.
| Rank | Country | Share of Working-Age Population Using AI H2 2025 (%) |
Change vs H1 2025 (p.p.) |
|---|---|---|---|
| 1 |
 UAE |
64.0 | 4.6 |
| 2 |
 Singapore |
60.9 | 2.3 |
| 3 |
 Norway |
46.4 | 1.1 |
| 4 |
 Ireland |
44.6 | 2.9 |
| 5 |
 France |
44.0 | 3.1 |
| 6 |
 Spain |
41.8 | 2.1 |
| 7 |
 New Zealand |
40.5 | 2.9 |
| 8 |
 Netherlands |
38.9 | 2.6 |
| 9 |
 UK |
38.9 | 2.5 |
| 10 |
 Qatar |
38.3 | 2.6 |
| 11 |
 Australia |
36.9 | 2.4 |
| 12 |
 Israel |
36.1 | 2.2 |
| 13 |
 Belgium |
36.0 | 2.5 |
| 14 |
 Canada |
35.0 | 1.5 |
| 15 |
 Switzerland |
34.8 | 2.4 |
| 16 |
 Sweden |
33.3 | 2.1 |
| 17 |
 Austria |
31.4 | 2.3 |
| 18 |
 South Korea |
30.7 | 4.8 |
| 19 |
 Hungary |
29.8 | 1.9 |
| 20 |
 Denmark |
28.7 | 2.1 |
| 21 |
 Germany |
28.6 | 2.1 |
| 22 |
 Poland |
28.5 | 2.1 |
| 23 |
 Taiwan |
28.4 | 2 |
| 24 |
 U.S. |
28.3 | 2 |
| 25 |
 Italy |
27.8 | 1.8 |
| 26 |
 Czechia |
27.8 | 2 |
| 27 |
 Bulgaria |
27.3 | 1.9 |
| 28 |
 Finland |
27.3 | 1.7 |
| 29 |
 Jordan |
27.0 | 1.6 |
| 30 |
 Slovenia |
26.5 | 1.4 |
| 31 |
 Costa Rica |
26.5 | 1.9 |
| 32 |
 Saudi Arabia |
26.2 | 2.5 |
| 33 |
 Lebanon |
25.7 | 0.9 |
| 34 |
 Portugal |
24.2 | 1.6 |
| 35 |
 Oman |
24.2 | 1.8 |
| 36 |
 Slovakia |
23.8 | 1.7 |
| 37 |
 Croatia |
23.7 | 1.9 |
| 38 |
 Vietnam |
23.5 | 2.3 |
| 39 |
 Dominican Republic |
22.7 | 0.7 |
| 40 |
 Uruguay |
22.5 | 1.6 |
| 41 |
 Lithuania |
22.4 | 1.4 |
| 42 |
 Jamaica |
22.1 | -0.1 |
| 43 |
 Colombia |
22.0 | 1.6 |
| 44 |
 Serbia |
21.5 | 1.2 |
| 45 |
 Panama |
21.5 | 1.8 |
| 46 |
 South Africa |
21.1 | 1.8 |
| 47 |
 Chile |
20.8 | 1.2 |
| 48 |
 Malaysia |
19.7 | 1.4 |
| 49 |
 Argentina |
19.6 | 1.8 |
| 50 |
 Bosnia andHerzegovina |
19.5 | 1.3 |
| 51 |
 Japan |
19.1 | 1.4 |
| 52 |
 Kuwait |
19.1 | 1.4 |
| 53 |
 Greece |
19.1 | 2.4 |
| 54 |
 Philippines |
18.3 | 1.2 |
| 55 |
 Georgia |
18.2 | 0.9 |
| 56 |
 Mexico |
17.8 | 1.1 |
| 57 |
 Ecuador |
17.7 | 0.7 |
| 58 |
 Brazil |
17.1 | 1.5 |
| 59 |
 Moldova |
17.0 | 0.4 |
| 60 |
 Albania |
16.5 | 0.7 |
| 61 |
 China |
16.3 | 0.9 |
| 62 |
 El Salvador |
16.2 | 0.9 |
| 63 |
 Romania |
16.2 | 1.6 |
| 64 |
 India |
15.7 | 1.5 |
| 65 |
 Azerbaijan |
15.5 | 1.3 |
| 66 |
 Guatemala |
14.8 | 1.1 |
| 67 |
 Peru |
14.7 | 1.3 |
| 68 |
 Türkiye |
14.6 | 1.2 |
| 69 |
 Mongolia |
14.3 | 1.7 |
| 70 |
 Namibia |
13.8 | 0.8 |
| 71 |
 Libya |
13.7 | 1 |
| 72 |
 Kazakhstan |
13.7 | 1 |
| 73 |
 Botswana |
13.7 | 0.9 |
| 74 |
 Gabon |
13.4 | 1.1 |
| 75 |
 Egypt |
13.4 | 0.9 |
| 76 |
 Honduras |
13.1 | 0.7 |
| 77 |
 Nepal |
13.0 | 0.7 |
| 78 |
 Senegal |
12.9 | 0.5 |
| 79 |
 Indonesia |
12.7 | 1 |
| 80 |
 Tunisia |
12.7 | 0.4 |
| 81 |
 Zambia |
12.3 | 0.6 |
| 82 |
 Algeria |
12 | 0.7 |
| 83 |
 Cote D'Ivoire |
11.7 | 0.9 |
| 84 |
 Bolivia |
11.6 | 0.7 |
| 85 |
 Iraq |
11.2 | 0.9 |
| 86 |
 Paraguay |
11.0 | 0.9 |
| 87 |
 Morocco |
10.9 | 0.4 |
| 88 |
 Gambia |
10.9 | 0.3 |
| 89 |
 Thailand |
10.7 | 1.6 |
| 90 |
 Iran |
10.7 | 0.7 |
| 91 |
 Nicaragua |
10.7 | 1.1 |
| 92 |
 Pakistan |
10.3 | 0.6 |
| 93 |
 Angola |
9.7 | 0.8 |
| 94 |
 Madagascar |
9.7 | 0.8 |
| 95 |
 Malawi |
9.7 | 0.8 |
| 96 |
 Mozambique |
9.7 | 0.8 |
| 97 |
 Benin |
9.3 | 0.6 |
| 98 |
 Burkina Faso |
9.3 | 0.6 |
| 99 |
 Ghana |
9.3 | 0.6 |
| 100 |
 Guinea |
9.3 | 0.6 |
| 101 |
 Guinea-Bissau |
9.3 | 0.6 |
| 102 |
 Liberia |
9.3 | 0.6 |
| 103 |
 Mali |
9.3 | 0.6 |
| 104 |
 Mauritania |
9.3 | 0.6 |
| 105 |
 Niger |
9.3 | 0.6 |
| 106 |
 Nigeria |
9.3 | 0.6 |
| 107 |
 Sierra Leone |
9.3 | 0.6 |
| 108 |
 Togo |
9.3 | 0.6 |
| 109 |
 Myanmar |
9.1 | 0.3 |
| 110 |
 Lesotho |
9.1 | 0.7 |
| 111 |
 French Guiana |
9.0 | -0.1 |
| 112 |
 Guyana |
9.0 | 0.7 |
| 113 |
 Suriname |
9.0 | 0.7 |
| 114 |
 Venezuela |
9.0 | 0.7 |
| 115 |
 Ukraine |
9.0 | 0.7 |
| 116 |
 Belarus |
8.4 | 0.8 |
| 117 |
 Kyrgyzstan |
8.2 | 0.6 |
| 118 |
 Kenya |
8.1 | 0.3 |
| 119 |
 Russia |
8.0 | 0.4 |
| 120 |
 Cameroon |
7.8 | 0.8 |
| 121 |
 Central AfricanRepublic |
7.8 | 0.8 |
| 122 |
 Chad |
7.8 | 0.8 |
| 123 |
 Congo |
7.8 | 0.8 |
| 124 |
 DRC Congo |
7.8 | 0.8 |
| 125 |
 Zimbabwe |
7.6 | 0.5 |
| 126 |
 Haiti |
7.6 | 0.7 |
| 127 |
 Papua New Guinea |
7.3 | 0.1 |
| 128 |
 Bangladesh |
7.1 | 0.4 |
| 129 |
 Syria |
7.1 | 0.6 |
| 130 |
 Burundi |
6.8 | 0.4 |
| 131 |
 Eritrea |
6.8 | 0.4 |
| 132 |
 Ethiopia |
6.8 | 0.4 |
| 133 |
 Somalia |
6.8 | 0.4 |
| 134 |
 South Sudan |
6.8 | 0.4 |
| 135 |
 Sudan |
6.8 | 0.4 |
| 136 |
 Tanzania |
6.8 | 0.4 |
| 137 |
 Uganda |
6.8 | 0.4 |
| 138 |
 Laos |
6.7 | 0.7 |
| 139 |
 Armenia |
6.6 | 0.4 |
| 140 |
 Sri Lanka |
6.6 | 0.4 |
| 141 |
 Uzbekistan |
6.3 | 0.6 |
| 142 |
 Rwanda |
6.3 | 0.3 |
| 143 |
 Cuba |
6.1 | 0.4 |
| 144 |
 Afghanistan |
5.6 | 0.5 |
| 145 |
 Tajikistan |
5.6 | 0.5 |
| 146 |
 Turkmenistan |
5.6 | 0.5 |
| 147 |
 Cambodia |
5.1 | 0.5 |
| -- |
 World Average |
16.3 | 1.2 |
With 64.0% of the population using generative AI tools, the UAE not only ranks first globally, but stands as one of the fastest-growing countries in adoption.
Even before ChatGPT launched, AI technology was being used across public services in the UAE. This was supported by early governance frameworks that were established in 2017, as part of its national AI strategy targeting nine key sectors.
Ranking in second is Singapore, where 60.9% of the population uses AI. Like the UAE, Singapore invested early in AI infrastructure and research and development.
In Europe, Norway ranks first along with taking third place globally with a 46.4% adoption rate. It is followed by Ireland (44.6%), and France (44.0%), two countries with robust tech ecosystems.
Meanwhile, adoption rates in the U.S. stood at 28.3%, or 24th overall. Interestingly, although the U.S. develops world-class AI research and is home to some of the world’s largest AI-related firms, trust in AI technology is fairly low.
According to the Edelman Trust Barometer, just 32.0% of the U.S. population trusts AI. In comparison, the figure jumps to 67.0% in the UAE.
At the tail end of the adoption spectrum is Cambodia, where it stands at just 5.1%. While progress is underway, limited investment and infrastructure remain key barriers to wider adoption.
To learn more about this topic, check out this graphic on AI compute hubs by country.
2026-01-22 11:20:30
With a population smaller than a mid-sized American suburb and an economy heavily dependent on Danish subsidies, Greenland would seem an unlikely candidate for the center of a geopolitical firestorm. Yet this autonomous Danish territory—home to just 57,000 people and a GDP of roughly $3.3 billion—has become one of the most talked-about places on Earth.
On conventional maps, where bigger countries stand out, Greenland certainly looks important. The island visually rivals Africa in size, appearing as an imposing landmass stretching across the top of the globe. But that impression is a 500-year-old cartographic illusion.
In 1569, Flemish cartographer Gerardus Mercator created a map projection that would become the default for classrooms, atlases, and eventually Google Maps. The Mercator projection preserves angles and shapes that are essential for navigation, but at a significant cost: it dramatically inflates landmasses as they approach the poles.
The result? Greenland appears roughly the same size as Africa. In reality, Africa is 14 times larger.
According to data from climate scientist Neil Kaye and the interactive mapping tool at Engaging Data, Greenland is the single most exaggerated territory on Earth by percentage. It is actually 73.9% smaller than is shown on a Mercator map.
In absolute terms:
A 2020 study published in ISPRS International Journal of Geo-Information surveyed over 130,000 people worldwide and found that this distortion meaningfully shapes how we perceive geography.
Participants consistently overestimated the size of high-latitude countries like Greenland, Canada, and Russia while underestimating equatorial nations. In other words, they had a cognitive bias baked in by decades of exposure to Mercator maps.
Despite the cartographic exaggeration, Greenland is no small place. At 2.17 million square kilometers, it ranks as the 12th largest country or territory in the world, larger than Saudi Arabia, Mexico, and Indonesia. It’s the world’s largest island, more than three times the size of Texas, and about 26% bigger than Alaska.
Its location adds to the intrigue. Thule Air Base in northwest Greenland sits almost exactly halfway between Washington, D.C. and Moscow along the polar route, a geography that has made the island strategically valuable since the Cold War. In an era of hypersonic missiles and renewed Arctic competition, that position remains critical.
From Nuuk, Greenland’s capital, the straight-line distance to Washington is nearly the same as to Copenhagen.
So yes, your mental map has been distorted. The reality of Greenland does not match the gargantuan size portrayed on the world’s most popular maps.
But it’s still the world’s largest island—rich in rare earths, positioned at the crossroads of a geopolitical power struggle, and increasingly ice-free due to climate change. The Mercator projection may exaggerate Greenland’s size, but its strategic importance is no illusion.
2026-01-22 09:04:58
See more visuals like this on the Voronoi app.
See visuals like this from many other data creators on our Voronoi app. Download it for free on iOS or Android and discover incredible data-driven charts from a variety of trusted sources.
Iceland and Greenland are often confused due to their similar names, but they differ dramatically in geography, climate, population, and political status.
This map compares the two islands across a variety of fundamental metrics. Despite being located just a few hundred miles apart in the North Atlantic, the two places differ dramatically in scale and development.
The data for this visualization comes from Wikipedia, and the World Bank.
Greenland is enormous, covering 2.16 million square kilometers—roughly the combined area of Texas, California, and Montana.
Iceland, by comparison, spans just over 103,000 square kilometers, similar in size of Kentucky. Yet Greenland is home to only 57,000 people, while Iceland’s population is nearly 393,000.
| Category |
 Greenland |
 Iceland |
|---|---|---|
| Local name | Kalaallit Nunaat | Ísland |
| Land area | 2.16 million km² | 103,125 km² |
| Avg Annual Temp | −1°C / 30°F | 5° / 41°F |
| Population | 57K | 393K |
| Capital | Nuuk | Reykjavík |
| Political status | Territory of Denmark | Sovereign nation |
| Ice coverage (land area) | 80% | 11% |
| GDP (USD, 2023) | $3.3B | $33.3B |
| GDP per capita (USD, nominal, 2023) | $58K | $82K |
| Life expectancy | 72 years | 83 years |
| Economy | Fisheries, public administration, subsidies from Denmark | Fisheries, tourism, aluminum smelting, data centers |
| Resources | Minerals, fish | Geothermal power, hydropower, fish |
Contrary to what their names suggest, Iceland has a much milder climate and far less ice coverage.
About 11% of Iceland’s land area is covered by ice, compared with roughly 80% of Greenland.
According to medieval sagas, Viking explorer Erik the Red named the icy island “Greenland” around 985 AD to make it sound more appealing to settlers from overcrowded Iceland. Today, Iceland’s average annual temperature sits around 5°C, while Greenland averages closer to −1°C.
Iceland’s economy is significantly larger and more diversified. In 2023, its GDP reached $33.3 billion, compared with Greenland’s $3.3 billion. On a per-capita basis, Iceland also comes out ahead, with nominal GDP per person around $82,000 versus $58,000 in Greenland.
Life expectancy reflects this gap as well, at 83 years in Iceland—among the highest globally—compared with 72 years in Greenland.
Iceland’s economy benefits from geothermal power, tourism, aluminum smelting, and data centers, while Greenland relies more heavily on fisheries and subsidies from Denmark, despite vast mineral resources.
If you enjoyed today’s post, check out How Venezuela’s Oil Reserves Compare to the Rest of the World on Voronoi, the new app from Visual Capitalist.
2026-01-22 05:17:54
See visuals like this from many other data creators on our Voronoi app. Download it for free on iOS or Android and discover incredible data-driven charts from a variety of trusted sources.
In just five years, the companies competing for AI chips and data center market share were reshuffled significantly.
This chart visualizes the changing market share of AI and data center revenue over time between Intel, Nvidia, and AMD.
The data comes from Bloomberg and company-reported segment revenue from Nvidia, Intel, and AMD, with the chart showing each company’s share of the combined (peer-set) total from 2021–2025.
At the start of the decade, Intel was the undisputed king, capturing over two-thirds of AI chip and data center market share when compared to Nvidia and AMD.
In 2021, Nvidia only had about 25% of market share and was known primarily for gaming GPUs, while AMD was a distant third with just 7% market share.
As seen in the data table below, revenues have shifted significantly since 2021, with Nvidia as the market share leader at 86% as of late 2025.
| Quarter | Intel | AMD | Nvidia |
|---|---|---|---|
| 2021 Q1 | 68% | 7% | 25% |
| 2021 Q2 | 64% | 9% | 27% |
| 2021 Q3 | 59% | 11% | 30% |
| 2021 Q4 | 59% | 11% | 30% |
| 2022 Q1 | 55% | 12% | 34% |
| 2022 Q2 | 47% | 15% | 38% |
| 2022 Q3 | 44% | 17% | 40% |
| 2022 Q4 | 45% | 17% | 38% |
| 2023 Q1 | 40% | 14% | 46% |
| 2023 Q2 | 26% | 8% | 66% |
| 2023 Q3 | 19% | 8% | 73% |
| 2023 Q4 | 16% | 9% | 75% |
| 2024 Q1 | 13% | 8% | 79% |
| 2024 Q2 | 12% | 9% | 80% |
| 2024 Q3 | 11% | 9% | 80% |
| 2024 Q4 | 8% | 9% | 83% |
| 2025 Q1 | 9% | 8% | 83% |
| 2025 Q2 | 8% | 7% | 85% |
| 2025 Q3 | 7% | 7% | 86% |
| 2025 Q4 E | 6% | 7% | 86% |
The viral rise of AI chatbots like OpenAI’s ChatGPT took the world by storm after launching in late 2022, turning the tide quickly as Big Tech and governments rushed to build “AI factories”—huge data centers designed to train and run large language models (LLMs)—driving demand toward GPU-heavy infrastructure.
Nvidia capitalized on this shift by improving not just the GPU (making it faster and more power-efficient) but the whole AI system.
This includes chips, networking, and software—so gains compounded at the platform level rather than relying on traditional CPU scaling.
CEO Jensen Huang noted that while traditional Moore’s Law had slowed for CPUs, Nvidia’s AI computing performance was doubling nearly every year.
He explained that Nvidia can push performance faster because it builds “the architecture, the chip, the system, the libraries, and the algorithms” together in parallel.
Beyond the silicon, Nvidia’s advantage was its complete software and hardware ecosystem, which created a moat that raised switching costs.
Intel’s Data Center & AI share fell for one primary reason amidst repeated delays in its 2021 and 2022 CPU chip iterations.
Intel was CPU-focused while competition intensified, and after ChatGPT’s launch (Q4 2022), data-center spending shifted toward GPU-heavy AI systems.
The company failed to adapt and scale, as its AI-chip deals fell short of initial expectations.
Management even dropped its 2024 target of $500M+ in AI-accelerator revenue, citing a software platform transition.
All those missteps left it underexposed to the fastest-growing slice of AI data-center spend, while Nvidia ran away with the lead.
If you enjoyed today’s post, explore more insights about the AI chips market on Voronoi, including Nvidia vs. AMD vs. Intel: Comparing AI Chip Sales.
2026-01-22 02:44:31
See visuals like this from many other data creators on our Voronoi app. Download it for free on iOS or Android and discover incredible data-driven charts from a variety of trusted sources.
As electric vehicle adoption accelerates, charging infrastructure is becoming a critical bottleneck. Countries that scale public chargers efficiently can reduce range anxiety, support faster EV adoption, and ease pressure on urban transport systems.
This visualization ranks major countries by EV charger density, measured as the number of electric vehicles per public charger as of Q3 2025. The data for this visualization comes from Benchmark Mineral Intelligence.
The Netherlands ranks first by a wide margin, with just five EVs per public charger. This reflects a highly coordinated infrastructure strategy, where chargers are often installed based on direct user requests. The result is an efficient, demand-driven network that minimizes congestion and maximizes charger utilization.
Despite having a low share of fast chargers today, the country is steadily expanding capacity. By 2030, fast chargers are expected to play a larger role as EV adoption continues to rise.
| Country | EVs per Charger (2025) | Fast Chargers (2025) | Fast Chargers (2030P) |
|---|---|---|---|
|
Netherlands |
5 | 3% | 5% |
|
China |
9 | 49% | 51% |
|
Italy |
10 | 26% | 32% |
|
Spain |
11 | 31% | 36% |
|
France |
13 | 21% | 33% |
|
India |
13 | 26% | 30% |
|
Sweden |
15 | 14% | 19% |
|
Germany |
19 | 25% | 30% |
|
UK |
26 | 20% | 28% |
|
USA |
31 | 28% | 33% |
China ranks second in charger density, with nine EVs per public charger, but leads decisively in fast-charging deployment. Nearly half of China’s public chargers are already direct current fast chargers, a figure projected to exceed 50% by 2030.
Fast chargers help support dense urban populations and long-distance travel across regions, reinforcing China’s dominance in the global EV adoption.
Several European countries cluster in the middle of the rankings, with roughly 10–13 EVs per public charger.
These countries are also rapidly expanding fast-charging infrastructure, with fast chargers projected to account for around one-third of networks by 2030.
By contrast, the U.S. trails the group, with 31 EVs per public charger.
If you enjoyed today’s post, check out Top 20 Countries by Battery Storage Capacity on Voronoi, the new app from Visual Capitalist.
