Times Higher Education3,118
+1,459%Change over time
A rank moves. So does the frame around it.
Trajectories only become useful when table expansion, ties, coverage, and measurement windows remain visible. Follow the line—then inspect the universe that produced it.
Institution trajectories
Follow one method at a time.
Choose an institution and provider. Rank 1 is at the top; bands use their lower bound. A line is not drawn across providers because each source defines and sizes its table differently.
Rank 1 is at the top. Each provider is viewed separately because table size, methodology, and rank bands differ. Banded ranks use their lower bound.
The denominator changed
Ranking universes expanded—sometimes dramatically.
A stable institution can move as hundreds or thousands of entrants are added. Overall-table row counts are shown exactly as published; THE’s latest total separates ranked entries from reporters.
QS1,504
+50.7%CWUR2,000
+1,900%SCImago SIR5,491
+148.8%Method changes the map
The same “top 100” tells six geographic stories.
Latest broad publisher editions. Counts include every institution whose displayed rank lower bound is 100 or better, so QS has 102 records because rank 100 is tied.
U.S. News2026 · 2,250 ranked
100- US
- 39
- GB
- 9
- CN
- 11
- Other
- 41
Times Higher Education2026 · 2,191 ranked
100- US
- 35
- GB
- 11
- CN
- 7
- Other
- 47
QS2027 · 1,504 ranked
102- US
- 26
- GB
- 16
- CN
- 6
- Other
- 54
CWUR2026 · 2,000 ranked
100- US
- 47
- GB
- 8
- CN
- 9
- Other
- 36
NTU Ranking2025 · 1,226 ranked
100- US
- 34
- GB
- 9
- CN
- 14
- Other
- 43
ShanghaiRanking2025 · 1,000 ranked
100- US
- 37
- GB
- 8
- CN
- 13
- Other
- 42
Hong Kong remains separate from mainland China. These are placements, not shares of national systems and not a pooled global league table.
Research geography
Two long-run views point toward wider Asian representation.
Nature Index academic Share and ARWU top-100 composition measure different things. Their direction is informative; their magnitudes are not directly comparable.
2015 output2025 output
- China+733.8%179 in table
- United States+17.1%105 in table
- United Kingdom+8.5%24 in table
- Germany+12.6%34 in table
- Japan-3.5%15 in table
- South Korea+75.5%13 in table
- Canada+8.6%12 in table
- Switzerland+26.1%7 in table
- Australia+39.8%11 in table
- Netherlands+57.6%10 in table
- India+149.2%15 in table
- France-9.5%11 in table
Across two decades the ARWU top 100 grew less concentrated. The U.S. share of seats declined while mainland China rose from none to a double-digit presence—visible as a single continuous trajectory before the two-edition snapshots below.
United States37%China13%United Kingdom8%Australia5%Germany4%Japan2%
Country concentration (Herfindahl–Hirschman index of top-100 seats) fell from 0.348 in 2003 to 0.174 in 2025 — a more distributed field. Hong Kong is counted separately from mainland China. The 2018 overall edition was unavailable, so it is omitted from the line.
ARWU 2003101 seats
Country HHI0.348
ARWU 2025100 seats
Country HHI0.174
2003 → 202558 → 37
U.S. institutions in the ARWU top 100.
0 → 13Mainland Chinese institutions; Hong Kong remains separate.
The 2003 denominator is 101 because the published rank 99 was tied. Lower HHI indicates less country concentration.Indexed-output momentum
OpenAlex coverage shows a changing publication footprint.
For the fixed cohort of 9,559 institutions present in all ten annual files, summed institution-work associations rose +95.1% through 2022.
United States+82.8%China+211.3%United Kingdom+62.3%India+146.6%Indonesia+366%Germany+59%
Institution-work associations, not unique papers. The dashed 2023–2025 segment is shown for continuity but is excluded from growth percentages because recent OpenAlex years are less mature.
Institution output growth
A handful of institutions multiplied their annual output.
Consensus-leading universities ranked by absolute growth in OpenAlex institution-work associations from 2016 to 2025. This is the institution-level counterpart to the country momentum above—scale, not citation impact.
2016 works2025 works
- Zhejiang UniversityChina11K → 33.5K13.1% CAGR
- Shanghai Jiao Tong UniversityChina11.8K → 33.4K12.2% CAGR
- Yale UniversityUnited States11.3K → 31.1K11.9% CAGR
- Peking UniversityChina13.4K → 32.9K10.5% CAGR
- Tsinghua UniversityChina13.6K → 31K9.6% CAGR
- University of OxfordUnited Kingdom17.6K → 33.6K7.4% CAGR
- Fudan UniversityChina7.5K → 21.6K12.6% CAGR
- University College LondonUnited Kingdom20.1K → 34.1K6.1% CAGR
- University of CopenhagenDenmark10.8K → 20.1K7.2% CAGR
- Stanford UniversityUnited States16.7K → 25.5K4.8% CAGR
- Johns Hopkins UniversityUnited States16K → 24.9K5% CAGR
- National University of SingaporeSingapore10.1K → 18.8K7.1% CAGR
- Cornell UniversityUnited States13.6K → 22.2K5.6% CAGR
- The University of MelbourneAustralia12.9K → 21.5K5.8% CAGR
- ETH ZurichSwitzerland8.6K → 16.2K7.2% CAGR
Institution-work associations reconstructed from OpenAlex, not unique papers. CAGR is the compound annual growth rate across the nine-year window; rank is the 2025 output position. Recent OpenAlex years are less mature than earlier ones, so treat late-window growth as indicative.
Scale is not impact
Publishing more and being cited more intensely are different signals.
Across 2,773 Leiden 2025 institutions, output rank and field-normalized impact rank have a Spearman correlation of 0.499. Only 24 institutions overlap their two top 100s.
Hover a point for institution detail. Larger coral points have at least 20% of publications in the field-normalized top 10%.
Web presence vs. academic standing
A large web footprint is its own signal, distinct from scholarly rank.
Webometrics ranks 32,053 institutions by web visibility. Against the academic consensus for the 752 shared universities, some punch far above their scholarly weight online while others stay comparatively quiet.
Webometrics 2025
Where web presence and academic standing diverge.
Webometrics ranks 32,053 institutions by the size and visibility of their web footprint. Comparing that with the academic consensus across the 752 shared universities (66% of the 1,140-strong cohort) shows who is over- and under-represented online relative to scholarly standing.
- 01Harvard University
- 02Stanford University
- 03Massachusetts Institute of Technology
- 04University of Oxford
- 05University of Washington
- 06University of Michigan
Web-forward
Louder online than the rankings suggest
Universities whose Webometrics web-presence rank sits far ahead of their academic consensus rank.
- Brigham Young UniversityWeb #207Academic #644+437
- Loyola University ChicagoWeb #302Academic #650+348
- Michigan Technological UniversityWeb #337Academic #678+341
- North Dakota State UniversityWeb #404Academic #734+330
- University of IdahoWeb #373Academic #697+324
- Portland State UniversityWeb #339Academic #654+315
- University of New BrunswickWeb #358Academic #659+301
- Kent State UniversityWeb #327Academic #624+297
Web-quiet
Academic strength the web understates
Universities the consensus rates highly but whose public web footprint trails their peers.
- Institut Polytechnique de ParisWeb #748Academic #79-669
- Sungkyunkwan UniversityWeb #752Academic #109-643
- Western UniversityWeb #750Academic #133-617
- Queen’s UniversityWeb #751Academic #179-572
- Institut AgroWeb #749Academic #294-455
- Rush UniversityWeb #746Academic #296-450
- Yeshiva UniversityWeb #668Academic #273-395
- Institute of Science TokyoWeb #561Academic #171-390
Positions are ranks within the 752 universities that appear in both Webometrics and the multi-source academic consensus, so the two scales are directly comparable. A positive delta means the web footprint ranks ahead of academic standing; a negative delta means the reverse.