World Tech Rankings
All technologies

AI chips

Hardware built to run neural networks: accelerators, neuromorphic chips and computing inside memory.

China leads AI chips research, with 46.2% of the 10% most cited papers published 2021–2025; South Korea is second with 12.1%.

Added 14 October 2026. Its top papers were checked for relevance by AI, but its keywords and topic list have not yet been reviewed by hand.

CN
FirstChina46.2%
KR
SecondSouth Korea12.1%
3.8×lead, by share of the 10% most cited papers, 2021–2025

Standings

1CNChina46.2%2KRSouth Korea12.1%3USUnited States11.6%4–5SGSingapore3.4%4–5DEGermany3.3%6INIndia2.3%7HKHong Kong2.2%8–9CHSwitzerland1.9%8–9GBUnited Kingdom1.9%10ITItaly1.8%

Share of the 10% most cited papers, 2021–2025; 689 papers. Source: OpenAlex, pulled 18 October 2026. Method

Year by year

About 138 papers a year. Single-year moves are noise.

China

0204060% China 51.6%

Everyone else, on a larger scale

01020% 20212022202320242025 South Korea 13.3%United States 5.5%India 3.1%Singapore 2.3%Germany 1.8%

Share of the 10% most cited papers published in each year. Source: OpenAlex, pulled 18 October 2026. Method

Inside the field

Always the top 10%Pilot

Each paper sorted into one sub-area by an AI model reading its title and abstract, then credited as usual. On 200 hand-labelled papers it chose the same sub-area, or a defensible second, 92% of the time.

Sub-areaPapersFirstSecondThird
Memristors and resistive-switching synapses178CNChina43.6%KRSouth Korea17.2%SGSingapore6.7%
Compute-in-memory and AI accelerator chips114CNChina35.2%USUnited States27.1%KRSouth Korea6.1%
Optoelectronic synapses and in-sensor computing104CNChina75.8%KRSouth Korea8.6%SASaudi Arabia3.8%
Spiking neural networks and neuromorphic systems66CNChina35.0%USUnited States17.9%NLNetherlands6.1%
Synaptic transistors (electrochemical, ionic, 2D, organic)61CNChina52.3%KRSouth Korea26.1%USUnited States6.9%
Ferroelectric memory and synaptic devices49CNChina53.5%KRSouth Korea10.2%DEGermany8.0%
Photonic AI accelerators31USUnited States28.7%CNChina26.8%DEGermany7.1%
Reservoir and physical computing30CNChina49.2%KRSouth Korea14.1%ITItaly12.9%

Share of each sub-area's papers in the top 10%. Source: OpenAlex, pulled 18 October 2026. Method

Earlier and recent

The same top 10%, split into 2021–2023 and 2024–2025.

FallingRising2021–20232024–2025Change
CNChina 42.8%49.3%+6.5 pts
KRSouth Korea 9.5%14.4%+5.0 pts
USUnited States 15.7%7.9%−7.9 pts
SGSingapore 4.6%2.3%−2.4 pts
DEGermany 4.7%1.9%−2.8 pts
INIndia 1.1%3.4%+2.3 pts
HKHong Kong 1.6%2.7%+1.2 pts
CHSwitzerland 2.8%1.2%−1.6 pts

Greyed rows rest on fewer than five papers of credit in one period.

Share of the 10% most cited papers: 327 papers in 2021–2023, 362 in 2024–2025. Source: OpenAlex, pulled 18 October 2026. Method

2026 so far: a different count

Not the most cited papers: every 2026 paper matched so far (1,805), and a paper counts in full for each country on it. This year's papers are too new to rank by citations, so these shares cannot be compared with the ones above.

  1. CNChina47.8%of 2026 papers have an author here
  2. KRSouth Korea14.3%of 2026 papers have an author here
  3. USUnited States13.2%of 2026 papers have an author here
  4. INIndia8.4%of 2026 papers have an author here
  5. DEGermany5.4%of 2026 papers have an author here

Who cites the top papers

How much more often a country's top papers are cited from home than its size predicts. Expected is the country's share of all papers in the field; the bar is the excess, in percentage points. Home citers include authors citing their own earlier work.

HomeExpectedExcess
CNChina 63.9%44.6%+19.3
USUnited States 27.1%18.0%+9.1
KRSouth Korea 27.9%14.0%+13.9

Based on up to 100 of each country's most cited single-country papers in the top 10%. Source: OpenAlex, pulled 18 October 2026. Method

Behind the share

Institutions with the most credit

  1. 1CNFudan University2.5%
  2. 2KRDongguk University2.1%
  3. 3CNChinese Academy of SciencesGovernment lab2.0%
  4. 4CNPeking University1.9%
  5. 5SGNational University of Singapore1.9%
  6. 6CNTsinghua University1.8%
  7. 7CNZhejiang University1.7%
  8. 8KRSeoul National University1.6%
  9. 9SAKing Abdullah University of Science and Technology1.2%
  10. 10CNHuazhong University of Science and Technology1.1%

Share of the top 10%. Institutions across all 45 technologies

Who does the work

SectorAllCNKRUS
Universities79.2%
81.8%
90.9%
76.1%
Research facilities14.2%
12.2%
6.0%
12.3%
Government labs3.4%
5.0%
0.1%
1.3%
Companies2.5%
0.2%
2.6%
10.3%

Share of credit by type of institution, overall and for the three leading countries.

Across borders

41.1% of the top 10% have authors in more than one country.

  1. CNHKChina with Hong Kong50
  2. CNSGChina with Singapore33
  3. CNUSChina with United States31
  4. KRUSSouth Korea with United States28
  5. CNGBChina with United Kingdom23

Most frequent pairings, by number of shared papers.

The papers behind the numbers

The five most cited papers this definition catches. If one looks out of place, the definition needs work.

  1. 1,877citations
  2. 1,734citations
  3. 1,108citations
  4. 924citations
  5. 709citations

More in computing and ai

The most technologically advanced countries, and is China ahead of the US?

How this technology is defined

A paper counts if its title or abstract matches the search below. Articles in core journals and proceedings only; retracted papers are excluded.

"AI accelerator" OR "neural network accelerator" OR "DNN accelerator" OR "neuromorphic computing" OR "neuromorphic hardware" OR "in-memory computing" OR "compute-in-memory" OR "processing-in-memory"

and its main topic in OpenAlex is one of: Advanced Memory and Neural Computing; Ferroelectric and Negative Capacitance Devices; Parallel Computing and Optimization Techniques; Neural Networks and Reservoir Computing; Low-power high-performance VLSI design; Embedded Systems Design Techniques; Phase-change materials and chalcogenides; 3D IC and TSV technologies. The topic check removes papers that only mention the technology.

Papers matched, 2021–2025
6,866
In the top 10%
689
In the top 1%
72
Countries with credit
46
Credit with no known country
1.1%
Every one of the 895 most cited papers was checked for relevance; 29 were left out as not about ai chips

The most cited of those left out:

  • Kokkos 3: Programming Model Extensions for the Exascale Era486 citations
  • Ferroelectric field-effect transistors based on HfO 2 : a review335 citations
  • Two-Dimensional van der Waals Ferroelectrics: Scientific and Technological Opportunities207 citations
  • Revival of Ferroelectric Memories Based on Emerging Fluorite‐Structured Ferroelectrics181 citations
  • Advancing Spiking Neural Networks Toward Deep Residual Learning161 citations
  • Roadmap for phase change materials in photonics and beyond131 citations
  • Ferroelectric transistors based on shear-transformation-mediated rhombohedral-stacked molybdenum disulfide125 citations
  • Analyzing and mitigating data stalls in DNN training121 citations
  • Binary metal oxide-based resistive switching memory devices: A status review119 citations
  • Van der Waals Ferroelectrics: Theories, Materials, and Device Applications117 citations
  • Temporal dendritic heterogeneity incorporated with spiking neural networks for learning multi-timescale dynamics117 citations
  • Self-Supervised High-Order Information Bottleneck Learning of Spiking Neural Network for Robust Event-Based Optical Flow Estimation113 citations

Definition version 0.5, pilot. It is a plain text file: copy it, change it, rerun it.

Read the method