Visual CapitalistModify

2026-04-23 02:06:33

Where Populations Are Booming and Shrinking by 2050

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.

The world’s population is projected to grow by 1.4 billion people by 2050—but that growth is becoming increasingly concentrated in a handful of regions.

Using data from the United Nations’ World Population Prospects 2024, this map shows where populations are rising fastest—and where they are entering long-term decline. The contrast is stark: parts of sub-Saharan Africa are set to nearly double in size, while several of the world’s largest economies are projected to shrink significantly.

These shifts will reshape labor markets, economic growth, and global influence over the coming decades.

The Fastest-Growing Countries in the World

The most dramatic population increases are concentrated in sub-Saharan Africa, where several countries are on track to nearly double in size by 2050.

The Democratic Republic of Congo leads globally, with its population projected to surge by over 100 million people (+93%). Close behind are countries like Niger, Angola, and Somalia.

The table below shows population forecasts across 195 countries worldwide:

Rank	Country	Population 2025 (M)	Population 2050 (M)	Change 2020- 2025	% Change 2020- 2025
1	DR Congo	112.8	218.2	+105.4M	93.4%
2	Central African Republic	5.5	10.6	+5.1M	92.6%
3	Angola	39	74.3	+35.3M	90.3%
4	Somalia	19.7	37.2	+17.6M	89.3%
5	Niger	27.9	52.5	+24.6M	88.1%
6	Chad	21	38.9	+17.9M	85.0%
7	Tanzania	70.5	129.6	+59.1M	83.7%
8	Mali	25.2	46.2	+21.0M	83.2%
9	Mozambique	35.6	63.5	+27.9M	78.3%
10	Mauritania	5.3	9.4	+4.1M	77.1%
11	Afghanistan	43.8	76.9	+33.0M	75.4%
12	Zambia	21.9	38.1	+16.2M	73.8%
13	Cameroon	29.9	51.1	+21.2M	71.0%
14	Cote d'Ivoire	32.7	55.7	+23.0M	70.4%
15	Yemen	41.8	71	+29.2M	69.9%
16	Congo	6.5	11	+4.5M	69.7%
17	Malawi	22.2	37.4	+15.1M	68.2%
18	Burundi	14.4	24.1	+9.7M	67.7%
19	Uganda	51.4	85.4	+34.0M	66.3%
20	Ethiopia	135.5	225	+89.5M	66.1%
21	Sudan	51.7	85.2	+33.5M	64.9%
22	Benin	14.8	24.4	+9.6M	64.9%
23	Madagascar	32.7	53.2	+20.4M	62.4%
24	Equatorial Guinea	1.9	3.1	+1.2M	62.2%
25	Senegal	18.9	30.4	+11.4M	60.4%
26	Togo	9.7	15.6	+5.9M	60.3%
27	Vanuatu	0.3	0.5	+199K	59.4%
28	Eritrea	3.6	5.7	+2.1M	57.9%
29	Gabon	2.6	4.1	+1.5M	57.5%
30	Solomon Islands	0.8	1.3	+470K	56.1%
31	Rwanda	14.6	22.7	+8.1M	55.9%
32	Liberia	5.7	8.9	+3.2M	55.5%
33	Burkina Faso	24.1	37.3	+13.2M	55.0%
34	Guinea	15.1	23.4	+8.3M	55.0%
35	Iraq	47	71.9	+24.9M	53.0%
36	Guinea-Bissau	2.2	3.4	+1.2M	52.9%
37	Zimbabwe	17	25.9	+8.9M	52.6%
38	Gambia	2.8	4.3	+1.5M	52.4%
39	Sao Tome and Principe	0.2	0.4	+125K	52.0%
40	Nigeria	237.5	359.2	+121.7M	51.2%
41	Palestine	5.6	8.5	+2.9M	51.2%
42	South Sudan	12.2	18.3	+6.2M	50.5%
43	Comoros	0.9	1.3	+425K	48.1%
44	Syria	25.6	37.8	+12.2M	47.5%
45	Sierra Leone	8.8	12.9	+4.1M	46.8%
46	Namibia	3.1	4.5	+1.4M	45.9%
47	Pakistan	255.2	371.9	+116.6M	45.7%
48	Kenya	57.5	83.6	+26.1M	45.3%
49	Tajikistan	10.8	15.6	+4.8M	44.4%
50	Ghana	35.1	50.6	+15.5M	44.2%
51	Vatican City	0.001	0.001	213	42.5%
52	Oman	5.5	7.8	+2.3M	42.4%
53	Jordan	11.5	16.4	+4.8M	42.1%
54	Uzbekistan	37.1	52.2	+15.2M	40.9%
55	Papua New Guinea	10.8	14.9	+4.1M	38.5%
56	Saudi Arabia	34.6	47.7	+13.1M	38.0%
57	Israel	9.5	13.1	+3.6M	37.6%
58	Egypt	118.4	161.6	+43.3M	36.6%
59	United Arab Emirates	11.3	15.4	+4.0M	35.4%
60	Honduras	11	14.8	+3.8M	34.9%
61	Botswana	2.6	3.4	+875K	34.2%
62	Kiribati	0.1	0.2	+46.1K	33.8%
63	Qatar	3.1	4.2	+1.0M	33.7%
64	Timor-Leste	1.4	1.9	+471K	33.2%
65	Kyrgyzstan	7.3	9.6	+2.3M	32.2%
66	Guatemala	18.7	24.7	+6.0M	32.0%
67	Nauru	0	0	+3.7K	31.0%
68	Bahrain	1.6	2.1	+496K	30.2%
69	Djibouti	1.2	1.5	+346K	29.3%
70	Bolivia	12.6	16.1	+3.5M	28.0%
71	Mongolia	3.5	4.5	+984K	28.0%
72	Kazakhstan	20.8	26.5	+5.7M	27.3%
73	Kuwait	5	6.4	+1.3M	26.7%
74	Lesotho	2.4	3	+630K	26.6%
75	Turkmenistan	7.6	9.6	+2.0M	26.5%
76	Algeria	47.4	59.6	+12.1M	25.6%
77	Nicaragua	7	8.8	+1.7M	25.0%
78	Samoa	0.2	0.3	+53.4K	24.4%
79	Libya	7.5	9.3	+1.8M	24.2%
80	Laos	7.9	9.8	+1.9M	23.9%
81	Haiti	11.9	14.7	+2.8M	23.6%
82	Paraguay	7	8.6	+1.6M	23.2%
83	Panama	4.6	5.6	+1.1M	23.2%
84	Malaysia	36	44.3	+8.3M	23.1%
85	Cambodia	17.8	21.9	+4.1M	22.9%
86	South Africa	64.7	79.2	+14.4M	22.3%
87	Bangladesh	175.7	214.7	+39.0M	22.2%
88	Belize	0.4	0.5	+93.7K	22.2%
89	Australia	27	32.5	+5.5M	20.5%
90	Eswatini	1.3	1.5	+249K	19.8%
91	Lebanon	5.8	7	+1.1M	19.7%
92	Peru	34.6	40.6	+6.0M	17.4%
93	Nepal	29.6	34.6	+5.0M	17.0%
94	Ecuador	18.3	21.3	+3.0M	16.7%
95	Luxembourg	0.7	0.8	+111K	16.3%
96	Philippines	116.8	134.4	+17.6M	15.1%
97	India	1,463.90	1,679.60	+215.7M	14.7%
98	Suriname	0.6	0.7	+94.1K	14.7%
99	Canada	40.1	45.6	+5.5M	13.7%
100	Morocco	38.4	43.4	+5.0M	13.0%
101	Mexico	131.9	148.9	+17.0M	12.9%
102	Dominican Republic	11.5	13	+1.5M	12.8%
103	Ireland	5.3	6	+662K	12.5%
104	Guyana	0.8	0.9	+105K	12.5%
105	Indonesia	285.7	320.7	+35.0M	12.2%
106	Maldives	0.5	0.6	+60.3K	11.4%
107	Brunei	0.5	0.5	+53.2K	11.4%
108	Colombia	53.4	59.4	+6.0M	11.2%
109	Micronesia	0.1	0.1	+12.7K	11.2%
110	Bhutan	0.8	0.9	+85.7K	10.8%
111	Iran	92.4	101.9	+9.4M	10.2%
112	Cyprus	1.4	1.5	+138K	10.0%
113	United States	347.3	380.8	+33.6M	9.7%
114	New Zealand	5.3	5.8	+503K	9.6%
115	Venezuela	28.5	31.1	+2.6M	9.0%
116	Iceland	0.4	0.4	+34.7K	8.7%
117	United Kingdom	69.6	75.5	+6.0M	8.6%
118	Vietnam	101.6	110	+8.4M	8.3%
119	Azerbaijan	10.4	11.2	+827K	8.0%
120	Cabo Verde	0.5	0.6	+38.8K	7.4%
121	Fiji	0.9	1	+67.1K	7.2%
122	Liechtenstein	0	0	+2.9K	7.2%
123	Myanmar	54.9	58.6	+3.8M	6.9%
124	Sri Lanka	23.2	24.8	+1.6M	6.8%
125	Seychelles	0.1	0.1	+9.0K	6.8%
126	Tunisia	12.3	13.1	+797K	6.5%
127	Sweden	10.7	11.3	+653K	6.1%
128	Argentina	45.9	48.3	+2.5M	5.4%
129	Bahamas	0.4	0.4	+21.2K	5.3%
130	Norway	5.6	5.9	+277K	4.9%
131	El Salvador	6.4	6.7	+298K	4.7%
132	Switzerland	9	9.3	+375K	4.2%
133	Turkey	87.7	91.3	+3.6M	4.1%
134	Costa Rica	5.2	5.4	+201K	3.9%
135	Singapore	5.9	6.1	+211K	3.6%
136	Netherlands	18.3	19	+612K	3.3%
137	Tuvalu	0	0.01	289	3.0%
138	France	66.7	68.2	+1.6M	2.4%
139	Chile	19.9	20.3	+460K	2.3%
140	Brazil	212.8	217.5	+4.7M	2.2%
141	Denmark	6	6.1	+122K	2.0%
142	Tonga	0.1	0.1	+1.5K	1.4%
143	Belgium	11.8	11.9	+112K	1.0%
144	Antigua and Barbuda	0.1	0.1	846	0.9%
145	San Marino	0.034	0.034	124	0.4%
146	Andorra	0.083	0.082	-709	-0.9%
147	Malta	0.55	0.54	-9.7K	-1.8%
148	North Korea	26.6	25.8	-784K	-3.0%
149	Grenada	0.12	0.11	-4.1K	-3.5%
150	Georgia	3.8	3.7	-143K	-3.7%
151	Uruguay	3.4	3.3	-130K	-3.9%
152	Monaco	0.04	0	-1.6K	-4.1%
153	Dominica	0.07	0.06	-2.7K	-4.1%
154	Austria	9.1	8.7	-389K	-4.3%
155	Saint Lucia	0.18	0.17	-8.1K	-4.5%
156	Finland	5.6	5.4	-272K	-4.8%
157	Russia	144	136.1	-7.9M	-5.5%
158	Saint Kitts and Nevis	0.05	0.04	-2.7K	-5.7%
159	Portugal	10.4	9.8	-642K	-6.2%
160	Spain	47.9	44.9	-3.0M	-6.2%
161	Slovenia	2.1	2	-136K	-6.4%
162	Barbados	0.28	0.26	-18.4K	-6.5%
163	Germany	84.1	78.3	-5.8M	-6.9%
164	Thailand	71.6	66.4	-5.2M	-7.3%
165	Trinidad and Tobago	1.5	1.4	-111K	-7.4%
166	Czechia	10.6	9.8	-784K	-7.4%
167	Hungary	9.6	8.7	-907K	-9.4%
168	Slovakia	5.5	4.9	-538K	-9.8%
169	Saint Vincent and the Grenadines	0.1	0.09	-11.0K	-11.0%
170	China	1,416.10	1,260.30	-155.8M	-11.0%
171	Greece	9.9	8.8	-1.1M	-11.3%
172	Palau	0.02	0.02	-2.1K	-12.1%
173	Italy	59.1	51.9	-7.3M	-12.3%
174	Estonia	1.3	1.2	-170K	-12.6%
175	South Korea	51.7	45.1	-6.5M	-12.6%
176	Mauritius	1.3	1.1	-161K	-12.7%
177	Jamaica	2.8	2.5	-382K	-13.5%
178	Poland	38.1	32.8	-5.3M	-14.0%
179	Cuba	10.9	9.4	-1.6M	-14.2%
180	Japan	123.1	105.1	-18.0M	-14.6%
181	Romania	18.9	16	-2.9M	-15.2%
182	Armenia	3	2.5	-457K	-15.5%
183	Montenegro	0.63	0.53	-99.4K	-15.7%
184	Croatia	3.8	3.2	-614K	-16.0%
185	North Macedonia	1.8	1.5	-301K	-16.6%
186	Belarus	9	7.5	-1.5M	-17.2%
187	Serbia	6.7	5.5	-1.2M	-17.3%
188	Ukraine	39	32	-7.0M	-17.9%
189	Latvia	1.9	1.5	-340K	-18.3%
190	Albania	2.8	2.2	-531K	-19.2%
191	Bulgaria	6.7	5.4	-1.3M	-19.5%
192	Lithuania	2.8	2.3	-571K	-20.2%
193	Moldova	3	2.4	-644K	-21.5%
194	Bosnia and Herzegovina	3.1	2.5	-685K	-21.8%
195	Marshall Islands	0.04	0.03	-11.1K	-30.6%

All 10 of the fastest-growing sovereign states are in sub-Saharan Africa, where fertility remains high and child mortality has fallen sharply—a demographic lag that East Asia passed through decades ago.

This surge will place increasing pressure on infrastructure, education systems, and job markets, while also creating opportunities for economic expansion.

Where Populations Are Shrinking the Fastest

At the other end of the spectrum, several major economies are entering sustained population decline—driven by low birth rates and aging populations.

China alone is projected to lose more than 150 million people by 2050, while Japan, Italy, and Russia are also facing steep contractions. This shift could have significant implications for economic growth, labor supply, and public finances. Overall, European countries make up 11 of the 20 largest absolute declines.

Even Thailand is projected to shrink by 4.2 million people, highlighting how population decline is spreading beyond traditionally aging regions. Like many countries in East Asia, Thailand faces persistently low fertility rates (around 1.2 births per woman) and a rapidly aging population.

By mid-century, global population trends will be defined less by overall growth and more by divergence.

A small group of countries will account for the vast majority of new people, while many others shrink. This widening gap between fast-growing and shrinking populations is set to reshape migration flows, economic power, and the global workforce over the coming decades.

Learn More on the Voronoi App

To learn more about this topic, check out this graphic on the countries with the most births per hour.

2026-04-23 00:37:44

Published

12 hours ago

on

April 22, 2026

By

Julia Wendling

• Twitter
• Facebook
• LinkedIn
• Reddit
• Pinterest
• Email

The following content is sponsored by Adobe

AI Use by Students Across Major Economies

   

Key Takeaways

• Brazil (11.6%) and India (11.5%) lead global student AI adoption.
• India’s young population supports stronger uptake than many developed peers.
• Adoption gaps persist, with the UK (4.6%) and Japan (5.6%) trailing significantly.

Student AI use is becoming increasingly common in classrooms worldwide, reshaping how students complete assignments, conduct research, and manage workloads. New data from Adobe Digital Insights (September 2025) highlights how student AI use for schoolwork varies significantly across major economies.

This visualization, created in partnership with Adobe, highlights global trends in student AI use and marks the third post in our series on AI use in India. While usage rates remain relatively close among top countries, a clear gap is emerging between early adopters and slower-moving markets. This points to broader differences in digital readiness and education systems.

Leaders in Student AI Use: Brazil and India Set the Pace

Brazil leads globally, with 11.6% of students using AI for schoolwork, closely followed by India at 11.5% and Italy at 11.1%. Canada (10.6%) and the United States (9.9%) also rank among the top adopters, showing that AI is gaining traction across both emerging and developed economies.

Country	Student AI Use
Brazil	11.6%
India	11.5%
Italy	11.1%
Canada	10.6%
United States	9.9%
Germany	8.8%
France	7.4%
Japan	5.6%
United Kingdom	4.6%

India is particularly notable in student AI use given its demographic advantage. With a median age of just 30, it has one of the youngest populations among major economies, helping drive faster adoption of AI study tools. School and university students across India are already using AI assistants for homework and revision, turning this youth-driven familiarity with technology into a tangible academic edge.

Lower Adoption Markets: Developed Markets Fall Behind

At the lower end of the spectrum, student AI use is notably weaker in advanced economies such as the United Kingdom (4.6%) and Japan (5.6%). These countries fall well behind the global leaders, highlighting a clear adoption gap.

Even mid-tier adopters like France (7.4%) and Germany (8.8%) lag behind top performers. This disparity may reflect stricter academic policies, slower institutional adoption, or cultural hesitancy toward AI in education. These factors could impact long-term digital competitiveness.

From AI Adoption to AI Study Tools

As student AI use grows globally, tools that bridge casual usage and focused learning will define which students stay ahead. They need purpose-built AI tools for students that turn course materials into real learning outcomes. For students in high-adoption markets like India and Brazil, the next step is turning that AI familiarity into real academic advantage. 

Free AI study tools like Adobe Acrobat Student Spaces let students upload class notes and instantly generate flashcards, quizzes, study guides, and even audio summaries, turning scattered materials into a structured study hub. As AI adoption grows globally, tools that bridge casual usage and focused learning will define which students stay ahead.

Explore AI-powered Document Workflows.

You may also like

• 

  AI4 weeks ago

  India on Top: AI Adoption by Country

  India’s tech-savvy workforce is leading the world in AI adoption, fueling the next wave of economic growth.

• 

  AI1 month ago

  How AI Could Add $600B to India’s Economy by 2035

  India’s AI boom is poised to reshape its economy, with AI expected to boost productivity by between $550B and $607B by 2035.

Subscribe

Please enable JavaScript in your browser to complete this form.

Join 375,000+ email subscribers: *

Sign Up

2026-04-22 22:49:12

Ranked: The World’s Biggest Memory Chip Makers

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.

Memory chip makers’ market capitalizations have surged as AI infrastructure spending reshapes the semiconductor industry.

This visualization is part of Visual Capitalist’s AI Week, sponsored by Terzo. It ranks the world’s publicly traded DRAM and NAND producers using data from CompaniesMarketCap and StockAnalysis.

Samsung Is the World’s Biggest Memory Maker

The data of memory chip maker valuations shows an industry that has rebounded sharply as AI server buildouts drive memory demand higher.

The data below shows the world’s leading publicly traded memory chip makers by market cap:

Company	Market Capitalization (Billions, USD)	Country	Memory Type
Samsung	897.3	South Korea	Both (DRAM + NAND)
SK Hynix	498.4	South Korea	Both (DRAM + NAND)
Micron	481.0	United States	Both (DRAM + NAND)
Sandisk	140.6	United States	NAND
Kioxia	106.8	Japan	NAND
Nanya	22.1	Taiwan	DRAM
Winbond	13.4	Taiwan	Both (DRAM + NAND)
Macronix	8.6	Taiwan	NAND
Powerchip Semiconductor Manufacturing	7.3	Taiwan	DRAM

Samsung ranks first by a wide margin, with a market capitalization of $897 billion. SK Hynix ($498B) and Micron ($481B) follow close behind, forming a clear top tier among memory producers.

Further down the ranking, Sandisk ($141B) and Kioxia ($107B) stand out as sizable second-tier players. Four Taiwanese companies round out the list, Nanya ($22B), Winbond ($13B), Macronix ($9B), and Powerchip Semiconductor Manufacturing ($7B).

Why Memory Stocks Have Climbed

Memory prices jumped in 2025 as suppliers kept output disciplined while AI server demand tightened supply. That combination helped lift both pricing power and investor expectations for memory producers.

Several names on the list posted especially dramatic gains. Nanya rose 560%, Sandisk climbed 559%, and Kioxia advanced 536%, while Winbond, SK Hynix, and Micron also saw major stock gains.

2026 has seen share prices continue to rise for memory makers. Samsung is already up 80% as of April 17, while SK Hynix is up 73% and Micron is up 59%.

What DRAM and NAND Memory Actually Do

DRAM is short-term working memory that holds the data apps need right now, but clears when power is off. NAND is long-term storage memory that keeps files and software even when a device is shut down.

Both are essential to modern computing, but AI data centers are making high-performance memory even more strategically important.

As a result, memory chip makers are increasingly tied not just to consumer electronics cycles, but also to the buildout of AI data centers.

Learn More on the Voronoi App

If you enjoyed today’s post, check out Comparing Major American Chipmakers in One Chart on Voronoi.

2026-04-22 19:38:16

How Much U.S. Population Growth Comes From Immigration

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.

Immigration is now the primary engine of U.S. population growth, and in some places, the only one.

From 2021 to 2025, over four out of every five new U.S. residents came from international migration, according to data from the Harvard University Joint Center for Housing Studies. In 14 states, immigration accounted for 100% of population gains, meaning growth would have been negative without it.

This map shows how much each state relies on immigration, revealing a divide between states gaining residents organically and those sustained almost entirely by global inflows.

Where Immigration Is the Only Source of Growth

In many states, population growth depends entirely on immigration.

This table shows immigration’s share of population change by state from 2021–2025. If immigration exceeds total population growth, the share is capped at 100%:

Florida and Texas led the nation in population growth, but for different reasons. Both gained more than one million international migrants between 2021 and 2025.

But their growth drivers differ. Florida combined strong immigration with large domestic inflows, despite negative natural change. Texas saw growth across all fronts, including a strong natural increase.

This contrast highlights a broader trend. While every state recorded net international migration during this period, many also faced domestic outflows or aging populations. In fact, 25 states saw net domestic outflows, while 21 recorded more deaths than births, making immigration the decisive factor separating growth from decline.

Texas added over 691,000 people through natural growth alone, more than California and New York combined.

When Growth Isn’t Enough: The California Example

California highlights the imbalance: despite nearly one million international arrivals and more births than deaths, the state still saw overall population decline driven by domestic outflows.

Seven states in total lost population over this period, underscoring how internal migration can outweigh both natural change and immigration.

The Future of Immigration and U.S. Population Growth

A sharp slowdown could reshape this map.

In 2026, U.S. immigration is expected to fall to 321,000, less than a fifth of the level seen in 2025. At the same time, natural population change is projected to remain flat.

For states highly dependent on immigration, this may mean slower growth or even population decline.

Over the past five years, six states, including Oregon and Michigan, experienced both domestic outmigration and negative natural change, leaving immigration as their primary source of growth.

States where immigration plays the largest role in population gains are also the most exposed to a slowdown, with potential ripple effects across:

• Tax receipts
• Consumer spending
• Housing demand
• Labor force growth

As natural growth fades, migration, both domestic and international, will determine which states continue to grow and which begin to fall behind.

Learn More on the Voronoi App

To learn more about this topic, check out this graphic on America’s fastest-growing states from 2025 to 2050.

2026-04-22 01:48:00

Ranked: Education Spending Per Student by Country

See visuals like this from many other data creators on our Voronoi app. Download it for free on iOS or Android and discover data-driven charts from a variety of trusted sources.

Education spending per student varies widely across countries, reflecting differences in national priorities and resources.

At the top of the ranking, Luxembourg stands alone, spending far more per student than any other country. Beyond this upper tier, spending levels diverge quickly across both advanced and emerging economies.

This chart ranks countries by annual education spending per student, using PPP-adjusted data from the OECD’s Education at a Glance 2025.

These differences influence everything from class sizes and teacher pay to access to technology and higher education outcomes.

Top Spenders Cluster in Europe and North America

Luxembourg stands far above all peers, spending over $31,000 per student, nearly $9,000 more than second-place Norway and several times higher than lower-ranked countries. The country also leads in teacher salaries.

Rank	Country	Expenditure per student (in USD PPP)
1	Luxembourg	31,439
2	Norway	22,558
3	Austria	20,942
4	United States	20,387
5	South Korea	19,805
6	Denmark	19,229
7	Netherlands	19,186
8	United Kingdom	19,072
9	Belgium	19,024
10	Canada	18,733
11	Iceland	18,707
12	Germany	17,960
13	Sweden	17,804
14	Australia	17,529
15	Ireland	15,915
16	France	15,427
17	OECD average	15,023
18	Finland	15,000
19	Slovenia	14,454
20	Japan	14,130
21	Italy	13,750
22	Spain	13,385
23	Portugal	12,956
24	Israel	12,877
25	Czechia	12,844
26	New Zealand	12,389
27	Estonia	12,362
28	Poland	11,488
29	Lithuania	11,313
30	Slovakia	11,259
31	Hungary	10,097
32	Latvia	9,204
33	Croatia	9,033
34	Bulgaria	8,703
35	Chile	8,068
36	Romania	7,221
37	Greece	7,137
38	Türkiye	5,305
39	China	5,161
40	South Africa	4,395
41	Mexico	4,066
42	Peru	2,612
--	OECD Average	15,022

The U.S. and several Western European countries also rank near the top, typically spending between $18,000 and $21,000. These high levels reflect both strong public funding and the higher costs of education systems.

Canada and the United Kingdom also fall within this upper tier, underscoring consistent investment across advanced economies.

OECD Average Masks Wide Gaps

While the OECD average is about $15,000 per student, most countries fall far from this midpoint, clustering either well above it in Western Europe and North America or far below it in emerging economies.

Countries like Japan, Italy, and Spain fall below the average despite being advanced economies. Meanwhile, emerging European economies such as Poland and Hungary spend closer to $10,000–$11,000.

Spending Drops Sharply Outside Advanced Economies

Outside the OECD’s highest spenders, education investment drops off rapidly.

Türkiye and China spend just over $5,000 per student, while Mexico and South Africa are closer to $4,000. Peru sits at the bottom of the ranking at roughly $2,600 per student.

Learn More on the Voronoi App

If you enjoyed today’s post, check out Comparing Education Levels Across 45 Countries on Voronoi, the new app from Visual Capitalist.

2026-04-22 00:28:26

Charted: Compute Costs More Than Talent in AI

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.

For leading AI companies, the biggest expense is not talent. It is compute.

This chart from Visual Capitalist’s AI Week, sponsored by Terzo, uses Epoch AI data to compare spending at Anthropic, Minimax, and Z.ai across R&D compute, inference compute, and staff plus other costs.

In every case, compute accounts for the majority of total spending, underscoring how capital-intensive it has become to build and serve frontier AI models.

How AI Company Costs Break Down

Despite differences in scale, all three companies allocate the largest share of their budgets to a single category: compute.

The data below compares spending composition across Anthropic, Minimax, and Z.ai. Anthropic’s figures are for 2025, while Minimax’s are from Q1 to Q3 of 2025 and Z.ai’s are for H1 2025.

Across all three AI companies, compute is the main cost center. Epoch AI estimates that R&D compute and inference compute together account for 57% to 70% of total spending, making infrastructure more expensive than staff and other costs in every case.

Among the three, Z.ai has the most R&D-heavy profile, with 58% of spending tied to compute powering model development and training.

Anthropic stands out for sheer scale. Epoch AI estimates the company spent $9.7 billion in 2025, including $6.8 billion on compute alone across training and inference.

Its costs are significantly higher than Minimax’s and Z.ai’s, even if the two Chinese AI companies’ figures were annualized to match Anthropic’s full-year period.

Both Chinese companies release many of their models as open source, meaning the model weights are freely available for anyone to download, modify, and run. This strategy helps them compete with better-funded U.S. labs by building developer adoption at a fraction of the cost.

AI Talent Costs Less Than Chips and Compute

One of the clearest takeaways is that talent costs less than compute in this comparison. Even though top AI labs pay some of the highest salaries in tech, staff and other costs still account for less than half of total spending at each of the three firms.

While the chart focuses on costs, Epoch AI estimates these labs are currently spending around 2–3x more than they generate in revenue, even as some expect economics to improve over time.

How These Estimates Were Built

This dataset comes with a few important caveats. Anthropic’s figures are based on reporting from The Information and are more speculative, while Minimax and Z.ai figures come from IPO filings released in January 2026.

The time periods also differ: Anthropic data is for the full year of 2025, Minimax covers 2025 Q1–Q3, and Z.ai covers 2025 H1. Epoch AI says its expense totals include operating expenses, cost of goods and services, and non-cash items such as stock-based compensation.

Learn More on the Voronoi App

If you enjoyed today’s post, check out The Soaring Revenues of AI Companies on Voronoi.