5W. AI COMMUNICATIONS
RESEARCH — FIRST EDITION
A 5W RESEARCH ARTIFACT — VOLUME 01

The AI
City Index
2026.

Cities Ranked

25

AI Engines Tested

5

Prompts Modeled

60

Dimensions Scored

6

THE THESIS
AI capital concentration is emerging as one of the strongest predictors of urban trajectory. Of the world's 25 leading metropolitan economies, a small group is capturing a disproportionate share of the AI build-out. The 5W AI City Index measures which cities are ahead, by how much, and on which dimensions — including the dimension that most others are ignoring: how AI engines themselves describe each city.
01 / LEADERBOARD
The Global Rankings
  1. 01
    San Francisco Bay Area
    USA
    Frontier dominance
    92.5
    TIER I
  2. 02
    London
    United Kingdom
    Capital + policy
    88.7
    TIER I
  3. 03
    Beijing
    China
    Capital concentration
    87.0
    TIER I
  4. 04
    Tel Aviv
    Israel
    Talent + security AI
    83.5
    TIER I
  5. 05
    Singapore
    Singapore
    Policy + infrastructure
    82.0
    TIER I
  6. 06
    Seoul
    South Korea
    Hardware + talent
    78.7
    TIER I
  7. 07
    Paris
    France
    Mistral anchor
    76.3
    TIER II
  8. 08
    Boston
    USA
    Research density
    75.8
    TIER II
  9. 09
    Abu Dhabi
    UAE
    Sovereign infrastructure
    74.7
    TIER II
  10. 10
    Shanghai
    China
    Industrial AI
    73.8
    TIER II
  11. 11
    Toronto
    Canada
    Research lineage
    71.8
    TIER II
  12. 12
    New York
    USA
    Capital, no anchor
    71.3
    TIER II
  13. 13
    Dubai
    UAE
    Adoption velocity
    70.5
    TIER II
  14. 14
    Shenzhen
    China
    Hardware capital
    69.0
    TIER II
  15. 15
    Berlin
    Germany
    European cluster
    64.3
    TIER III
  16. 16
    Hyderabad
    India
    GCC scaling
    62.1
    TIER III
  17. 17
    Tokyo
    Japan
    Legacy capital, no breakout
    60.5
    TIER III
  18. 18
    Bangalore
    India
    Services scale
    59.8
    TIER III
  19. 19
    Riyadh
    Saudi Arabia
    Sovereign acceleration
    58.2
    TIER III
  20. 20
    Stockholm
    Sweden
    Nordic stronghold
    56.7
    TIER III
  21. 21
    Amsterdam
    Netherlands
    Cluster, no anchor
    54.3
    TIER III
  22. 22
    Los Angeles
    USA
    Capital decline
    52.9
    TIER III
  23. 23
    Zurich
    Switzerland
    Premium niche
    51.4
    TIER III
  24. 24
    Montreal
    Canada
    Research, no scale
    50.6
    TIER III
  25. 25
    Sydney
    Australia
    Periphery
    47.2
    TIER III
"AI is the most important shift in business in twenty-five years. It will affect every single aspect of life, and every city in the world."
Ronn Torossian  /  Founder & Chairman, 5W
02 / THESIS
The Bifurcation

Cities still compete on cost of living, business climate, and quality of life. A new variable has emerged that increasingly predicts urban trajectory: AI capital concentration.

For much of the post-pandemic period, the dominant narrative around major cities was decline — residents leaving high-cost coastal markets, offices emptying, knowledge workers spreading thinner across the country. The Census Bureau's Vintage 2025 estimates, released in March 2026, tell a more complicated story. America's four legacy urban cores are diverging sharply, and the variable that best predicts which way each is moving is the presence — or absence — of frontier AI capital.

San Francisco — written off three years ago — is the only one of America's four legacy urban cores actually growing in 2025, with population up 0.62%. Office leasing rose 20% in 2024 and absorbed positively for five straight quarters. The reason is unambiguous: OpenAI, Anthropic, Nvidia, Sierra AI, and xAI now anchor downtown.

Los Angeles — comparable in scale and historically more attractive on most quality-of-life axes — posted the largest numeric population decline in the United States in 2025: a loss of 54,000 residents in LA County. Hollywood shed 45,000 jobs. Downtown office vacancy hit 34%. Los Angeles has yet to develop a comparable frontier-AI anchor ecosystem.

The same divergence repeats globally. London raised a record £8.3 billion of AI venture capital in 2025 and lifted AI-led companies to 34% of all Central London tech office take-up. Paris, despite a stagnant national economy, runs the highest AI share of any continental European city's venture base — 43% — anchored by Mistral. Tel Aviv pulled in roughly $15.6 billion of total tech investment in 2025, with AI taking the leading share. Beijing concentrates 66% of local VC into AI, the highest share of any city in the world. Singapore tops the Counterpoint Global AI Cities Index 2025. Abu Dhabi committed a 5-gigawatt AI campus — the largest AI infrastructure project outside the United States.

The cities that have not captured an AI anchor — Tokyo, Sydney, Amsterdam, Los Angeles — are flat or shrinking across most of the dimensions that matter for an AI-era economy: capital, talent, office demand, population. Wealth, weather, and the weight of history are no longer reliable predictors of urban trajectory. The presence of frontier AI capital increasingly is.

This is the central finding of the 5W AI City Index 2026. Across six measured dimensions — capital intensity, talent density, infrastructure, policy support, frontier-lab anchor presence, and modeled AI citation share — the same nine cities cluster at the top. The remaining sixteen, including some of the most powerful urban economies in the world, are measurably behind the leaders on most of those dimensions. Without a structural shift in capital flow or policy, the gap may become difficult to close.

/ MAP
The Geography of AI Capital

All 25 cities ranked. Dot size = composite score. Dot color = tier. Read the picture: AI capital concentrates in nine cities across four regions. The lower 16 are visibly smaller — which is the visual truth of the index thesis.

Tier I — composite ≥ 78 Tier II — composite 69 – 77 Tier III — composite below 69
AMERICAS EUROPE MIDDLE EAST ASIA-PACIFIC SAN FRANCISCO, rank 1Composite score 92.5 in the 5W AI City Index regional cluster map. 1 SAN FRANCISCO 92.5 BOSTON, rank 8Composite score 75.8 in the 5W AI City Index regional cluster map. 8 BOSTON 75.8 NEW YORK, rank 12Composite score 71.3 in the 5W AI City Index regional cluster map. 12 NEW YORK 71.3 TORONTO, rank 11Composite score 71.8 in the 5W AI City Index regional cluster map. 11 TORONTO 71.8 LOS ANGELES, rank 22Composite score 52.9 in the 5W AI City Index regional cluster map. 22 LOS ANGELES 52.9 MONTREAL, rank 24Composite score 50.6 in the 5W AI City Index regional cluster map. 24 MONTREAL 50.6 LONDON, rank 2Composite score 88.7 in the 5W AI City Index regional cluster map. 2 LONDON 88.7 PARIS, rank 7Composite score 76.3 in the 5W AI City Index regional cluster map. 7 PARIS 76.3 BERLIN, rank 15Composite score 64.3 in the 5W AI City Index regional cluster map. 15 BERLIN 64.3 AMSTERDAM, rank 21Composite score 54.3 in the 5W AI City Index regional cluster map. 21 AMSTERDAM 54.3 STOCKHOLM, rank 20Composite score 56.7 in the 5W AI City Index regional cluster map. 20 STOCKHOLM 56.7 ZURICH, rank 23Composite score 51.4 in the 5W AI City Index regional cluster map. 23 ZURICH 51.4 TEL AVIV, rank 4Composite score 83.5 in the 5W AI City Index regional cluster map. 4 TEL AVIV 83.5 ABU DHABI, rank 9Composite score 74.7 in the 5W AI City Index regional cluster map. 9 ABU DHABI 74.7 DUBAI, rank 13Composite score 70.5 in the 5W AI City Index regional cluster map. 13 DUBAI 70.5 RIYADH, rank 19Composite score 58.2 in the 5W AI City Index regional cluster map. 19 RIYADH 58.2 BEIJING, rank 3Composite score 87.0 in the 5W AI City Index regional cluster map. 3 BEIJING 87.0 SEOUL, rank 6Composite score 78.7 in the 5W AI City Index regional cluster map. 6 SEOUL 78.7 SHANGHAI, rank 10Composite score 73.8 in the 5W AI City Index regional cluster map. 10 SHANGHAI 73.8 SINGAPORE, rank 5Composite score 82.0 in the 5W AI City Index regional cluster map. 5 SINGAPORE 82.0 SHENZHEN, rank 14Composite score 69.0 in the 5W AI City Index regional cluster map. 14 SHENZHEN 69.0 TOKYO, rank 17Composite score 60.5 in the 5W AI City Index regional cluster map. 17 TOKYO 60.5 BANGALORE, rank 18Composite score 59.8 in the 5W AI City Index regional cluster map. 18 BANGALORE 59.8 HYDERABAD, rank 16Composite score 62.1 in the 5W AI City Index regional cluster map. 16 HYDERABAD 62.1 SYDNEY, rank 25Composite score 47.2 in the 5W AI City Index regional cluster map. 25 SYDNEY 47.2
The 5W AI City Index 2026 — composite score visualization, all 25 ranked cities by region
03 / FINDINGS
Six Numbers That Tell the Story
£8.3bn
London's record AI venture haul in 2025 — Britain's AI sector tightened its grip as Europe's leading tech hub, with nearly 75% of UK AI fintech based in the capital.
66%
Beijing's share of local VC flowing into AI — the highest concentration of any city on Earth. Silicon Valley sits at 62%, Toronto-Waterloo at 50%, Paris at 43%.
5GW
The Abu Dhabi AI Campus — announced May 2025, the largest AI infrastructure project outside the United States. Microsoft committed $15.2B to the UAE through 2029.
−54k
Los Angeles County's 2025 population loss — the largest numeric population decline in the United States, alongside 34% downtown office vacancy and 45,000 Hollywood jobs gone.
+0.62%
San Francisco's 2025 population growth — the only one of America's four legacy urban cores expanding. Office vacancy fell 3.7 points, the steepest annual drop since 2011.
$15.6bn
Israel's 2025 tech investment — most of it concentrated in Tel Aviv, with a leading share into AI. The city was ranked #1 globally for cybersecurity AI in 2026.
04 / DIMENSIONS
The Six Measured Variables

The 5W AI City Index scores each of 25 cities across six dimensions, equally weighted. The composite score is the average of the six dimension scores, each normalized to a 0–100 scale. Tiebreaks are resolved by Dimension 6 — modeled AI citation share — because it is the dimension that most directly measures the future of brand and place visibility in the answer-engine era.

DIMENSION 01
AI Capital Intensity
Absolute AI venture capital raised and AI's share of all local venture funding. Measures whether private money chose the city.
01
San Francisco
98
02
Beijing
95
03
Tel Aviv
88
04
London
87
05
Abu Dhabi
85
DIMENSION 02
AI Talent Density
AI engineers per 100k workforce, AI research output by city affiliation, top-200 universities in AI subjects.
01
San Francisco
96
02
Tel Aviv
95
03
Beijing
95
04
London
92
05
Boston
88
DIMENSION 03
AI Infrastructure
Data center capacity in gigawatts, hyperscaler presence, energy availability, compute access.
01
Abu Dhabi
95
02
Shenzhen
92
03
Beijing
92
04
Seoul
90
05
Singapore
88
DIMENSION 04
AI Policy & Government Support
National AI strategy quality, public AI investment, regulatory clarity, talent immigration friction.
01
Singapore
98
02
London
95
03
Abu Dhabi
95
04
Dubai
90
05
Paris
88
DIMENSION 05
Frontier Lab Anchor Presence
Frontier AI labs headquartered or with major office (100+ staff). Direct count of frontier-model builders by city.
01
San Francisco
100
02
London
88
03
Beijing
82
04
Paris
78
05
Boston
78
DIMENSION 06
AI Citation Share
Modeled share of mentions across 60 prompts and 5 AI engines (ChatGPT, Claude, Gemini, Perplexity, Google AI Overviews).
01
San Francisco
95
02
Tel Aviv
88
03
Singapore
88
04
London
86
05
Beijing
78
/ SHAPES
Top 3 Cities, Side by Side

San Francisco, London, and Beijing — the three Tier I leaders — score above 87 on the composite, but their shapes differ. SF runs deepest on capital, talent, and frontier-lab anchors. London is the most balanced book in the index. Beijing leads on capital intensity but lags on citation share in Western AI engines.

San Francisco — 92.5 London — 88.7 Beijing — 87.0
Capital Talent Infrastructure Policy Anchor Labs Citation Share 25 50 75 100
Composite shape comparison — six-dimension radar, top 3 cities
05 / PROFILES
Tier I — The Six
01
RANKED FIRST

San Francisco Bay Area

USA — COMPOSITE 92.5
Capital98
Talent96
Infra88
Policy78
Anchor100
Citation95

The city written off in 2023 is the only one of America's four legacy urban cores actually growing.

San Francisco runs the table. Five frontier labs — OpenAI, Anthropic, xAI, Inflection, and Sierra AI — are headquartered within a 1.5-mile radius. Nvidia, Meta AI, and Google DeepMind operate large secondary offices. Hines filed permits for a 1,225-foot tower at 77 Beale Street, what would be the city's tallest, signalling developer conviction in the AI-driven recovery.

The population data is the headline: +0.62% in 2025, while Los Angeles County posted the largest numeric population loss in the nation. Office leasing rose 20% in 2024 — the highest level since 2019 — and absorbed positively for five consecutive quarters. Vacancy dropped 3.7 percentage points, the steepest annual decline since 2011.

The risk is concentration. SF leads on every dimension except policy, where federal AI policy volatility could expose the city more than any other. A single shift in U.S. export controls or labor-visa policy moves SF's score down faster than any other city in the index could move up.

02
RANKED SECOND

London

UNITED KINGDOM — COMPOSITE 88.7
Capital87
Talent92
Infra84
Policy95
Anchor88
Citation86

The only city in the index that scores in the top five on every dimension. Most balanced book.

London raised a record £8.3 billion of AI investment in 2025. AI-led companies have taken 1.5 million square feet of Central London office space since 2015 and accounted for 34% of all London tech office take-up in 2025. CBRE projects the AI share could reach 4 million square feet by 2033 — equivalent to 43% of the entire Central London office pipeline currently under construction or available to let.

The anchor is Google DeepMind, headquartered in King's Cross. Add Stability AI, Wayve (which raised over $1B in 2024), and the European offices of Anthropic and OpenAI. UK AI sector employment grew 70% from 50,000 in 2022 to over 86,000 by 2024. London now hosts the Alan Turing Institute, 42 unicorns, and tied with Tel Aviv for seventh globally as a startup ecosystem.

The advantage on Dimension 4 (policy) is structural. London Tech Week 2025 delivered over £4 billion in fresh commitments. The UK's "supercharged sandbox" regulatory regime for AI experiments puts it ahead of any other Western capital on speed-to-deployment. Britain's chronic weakness — scale-up exits to U.S. acquirers — has not been solved. But London's composite of capital, policy, and frontier-anchor presence is the most defensible in the world after the Bay Area.

03
RANKED THIRD

Beijing

CHINA — COMPOSITE 87.0
Capital95
Talent95
Infra92
Policy80
Anchor82
Citation78

The city with the highest concentration of AI venture capital on Earth.

Beijing converts 66% of all local venture funding into AI — the highest share of any city in the world, according to Startup Genome and Statista's December 2025 data. Compare: Silicon Valley at 62%, Toronto-Waterloo at 50%, Paris at 43%, Shanghai at 22%. No global hub matches Beijing's intensity of AI capital deployment.

The anchor lineup is formidable: DeepSeek, Baidu, Alibaba's DAMO Academy, ByteDance's AI Lab, Zhipu AI. DeepSeek's R1 model, trained for roughly $5.6 million and released open-source, repositioned the global cost frontier of frontier AI in early 2025. The Tsinghua–Peking university talent pipeline rivals Stanford and MIT in raw AI graduate output.

Beijing's drag is on the citation share dimension — Western AI engines under-cite Chinese sources, and policy/regulatory unpredictability suppresses its score. If Beijing's AI engines ever dominate the global citation graph, this index re-ranks it to first. For now, second-place capital and talent scores are enough to lock the bronze.

04
RANKED FOURTH

Tel Aviv

ISRAEL — COMPOSITE 83.5
Capital88
Talent95
Infra70
Policy85
Anchor75
Citation88

The highest AI talent density per capita on Earth. The category-defining city in cybersecurity and security AI.

Israel raised approximately $15.6 billion in tech investment in 2025, the bulk concentrated in Tel Aviv, with the leading share flowing into AI. The World Future Awards 2026 ranking placed Tel Aviv first globally in Cybersecurity and AI Security, ahead of San Francisco and London. Tel Aviv tied with London for seventh globally in startup ecosystem strength.

The talent score is structural and inherited. The 8200 intelligence unit pipeline, Technion graduates, and a deep founder bench — alumni of Iron Source, Mobileye, Wiz, Lemonade, Lightricks — funnel into a startup density unmatched outside the Bay Area. Tel Aviv University and Technion alumni show up in the founding teams of dozens of AI-native companies across the Middle East and the U.S.

The infrastructure dimension is the visible weakness. Israel lacks hyperscale data-center capacity at the scale of the UAE, Singapore, or China. The compute layer is being addressed through partnerships, including the strategic Israel-UAE compute corridor, but it is the dimension on which Tel Aviv loses points to its capital-and-talent peers.

05
RANKED FIFTH

Singapore

SINGAPORE — COMPOSITE 82.0
Capital75
Talent78
Infra88
Policy98
Anchor65
Citation88

The world's best-governed AI city. The Counterpoint Research Global AI Cities Index 2025 ranked it first.

Singapore's score is built on the policy dimension. The country's National AI Strategy 2.0 — published 2023 and operating through 2030 — coordinates research investment, talent immigration, and regulatory sandboxing in a way that no Western state has matched. The Monetary Authority of Singapore runs the most advanced AI-in-financial-services regulatory framework in the world.

Singapore captured 96.6% of Southeast Asian venture funding in January 2026, per Crunchbase. Hyperscaler data center capacity is concentrated here for the entire ASEAN region. Singapore is the cross-border HQ choice for AI companies expanding into Asia — a function not just of language and tax, but of legal predictability.

The drag is on Dimension 5 (anchor presence). No frontier model is trained primarily in Singapore. It is the world's AI operating system, not its AI cathedral. That ceiling is real but not currently being tested — most cities would trade their anchor presence for Singapore's governance fluency.

06
RANKED SIXTH

Seoul

SOUTH KOREA — COMPOSITE 78.7
Capital78
Talent85
Infra90
Policy82
Anchor70
Citation67

The hardware backbone of the AI economy. The city where the chips that train every frontier model are made.

Seoul is the AI infrastructure capital of the developed world. Samsung Electronics and SK Hynix together supply more than 90% of the global HBM (high-bandwidth memory) market — the most critical component in every Nvidia H100, H200, and B200 GPU. Without Seoul, no frontier model trains.

Counterpoint Research placed Seoul second on its Global AI Cities Index 2025. Naver, Kakao, and LG AI Research are domestic anchors. Korean government investment in AI infrastructure has accelerated under the 2025 K-Cloud initiative. Talent flows are strong: Korean AI graduate output places Seoul among the top five global producers, with KAIST a recognized peer of MIT in select AI subjects.

The weakness is on citation share. Like Beijing, Seoul under-indexes in Western AI engines despite massive structural importance. If you read this report and conclude Seoul should rank fourth or third, the citation-share dimension is why it doesn't — and that gap is what 5W is built to close.

Independent Assessment — Brookings Institution
"It remains a highly concentrated early-stage industry dominated by the Bay Area."
Mark Muro  ·  Senior Fellow, Brookings Metropolitan Policy Program
06 / TIER II
The Contenders

Tier II — ranks 7 through 14 — captures the cities that have built defensible AI capabilities but lack the full stack required for Tier I status. Each scores top-10 on at least two dimensions but carries a structural deficiency on at least one other. Most are within reach of Tier I with a single anchor relocation or sovereign investment announcement.

RANK 07
Paris
FRANCE
COMPOSITE 76.3
Anchored by Mistral AI, Europe's only credible challenger to OpenAI on the frontier-model leaderboard. Paris converts 43% of local VC into AI — third globally by share, behind only Beijing and the Bay Area. France's national AI strategy is now the most aggressive in the European Union. The constraint is talent retention: French AI engineers are recruited at scale to London and the Bay Area.
RANK 08
Boston
USA
COMPOSITE 75.8
The research density is unmatched outside the Bay Area. MIT, Harvard, and the Broad Institute form an AI research triangle that produces a third of the U.S. AI PhDs. Boston Dynamics, IBM Research, and a strong biotech-AI cluster anchor the commercial side. Boston's score is suppressed by capital intensity — institutional VC flows toward SF and NYC — and by the lack of a frontier-lab anchor.
RANK 09
Abu Dhabi
UAE
COMPOSITE 74.7
The fastest-rising city in the index. The 5-gigawatt UAE-US AI Campus, announced May 2025, is the largest AI infrastructure project outside the United States. Microsoft committed $15.2 billion through 2029; G42 operates as the sovereign anchor. Abu Dhabi tops the global infrastructure score and ties with London on policy. The drag is talent and anchor depth — both being addressed via aggressive capital deployment.
RANK 10
Shanghai
CHINA
COMPOSITE 73.8
China's industrial AI capital. Anchored by Tencent, SenseTime, and a deep enterprise AI ecosystem. Shanghai converts roughly 22% of local VC into AI, third in the country after Beijing and Shenzhen. The city is being repositioned by Chinese AI policy as the enterprise complement to Beijing's frontier focus. Citation share lags significantly behind Western peers, suppressing the composite.
RANK 11
Toronto
CANADA
COMPOSITE 71.8
Birthplace of modern deep learning. The Vector Institute remains a world-class research anchor; Cohere is Canada's only frontier-tier AI lab. The Toronto-Waterloo corridor converts 50% of local VC into AI — third globally by share. Talent density is strong; the structural drag is U.S. acquisition: top Toronto AI engineers and founders consistently relocate to SF or are recruited by U.S. labs.
RANK 12
New York
USA
COMPOSITE 71.3
The capital city, not the anchor city. NYC posted a net population loss of 12,000 in 2025, reversing two years of growth. The financial AI cluster — Two Sigma, Renaissance Technologies, JPMorgan's AI division, Bloomberg AI — is formidable but not frontier. Manhattan rents at all-time highs reflect demand to live in the city, not a population recovery. NYC scores high on capital and citation share but loses on the anchor dimension.
RANK 13
Dubai
UAE
COMPOSITE 70.5
The AI adoption velocity capital. Dubai targets 64% of its working-age population using AI by end of 2025, per the National AI Strategy. The Golden Visa program and Saudi Arabia–UAE talent corridor accelerate inbound. The city pairs Abu Dhabi's sovereign infrastructure with a service-economy adoption layer. Anchor presence is the persistent drag — Dubai is consumed AI more than it is built AI.
RANK 14
Shenzhen
CHINA
COMPOSITE 69.0
The hardware capital of AI in China. Anchored by Huawei, DJI, BYD, and a dense electronics manufacturing base that produces the world's edge-AI devices. Shenzhen's score is constrained by export controls limiting access to leading-edge GPUs, and by under-citation in Western AI engines. On infrastructure and applied AI, Shenzhen scores higher than the composite suggests.
07 / TIER III
The Mid-Pack

Tier III — ranks 15 through 25 — captures cities with real but incomplete AI capabilities. Each has a credible path forward, but none currently scores in the top five on more than one dimension. Three legacy global power cities sit inside this tier: Tokyo (17), Los Angeles (22), and Sydney (25). Each was a defining city of the analog and early-digital era. None has yet developed a frontier-AI anchor ecosystem at the scale of Tier I or Tier II.

RANK 15
Berlin
GERMANY
COMPOSITE 64.3
Europe's strongest secondary AI cluster. Anchored by Helsing (defense AI, $5B+ valuation), DeepL, and Aleph Alpha. Talent pipeline runs through TU Berlin and the Max Planck Institutes. The structural drag is regulatory friction at the EU level and a persistent VC gravity pull toward London and Paris for late-stage rounds. Berlin's path to Tier II runs through faster AI Act implementation and a federal compute commitment.
RANK 16
Hyderabad
INDIA
COMPOSITE 62.1
India's largest Global Capability Center (GCC) hub. Microsoft, Google, and Meta operate major AI engineering centers here, totaling more than 100,000 AI-adjacent jobs. The fastest-growing AI services hub in India and a major data center destination. Hyderabad is currently positioned for AI execution at scale rather than frontier invention — but the volume of applied AI engineering shipped from this city is unmatched outside the U.S. and China.
RANK 17
Tokyo
JAPAN
COMPOSITE 60.5
The richest city in the lower half of the index. SoftBank's commitment to Stargate is the most aggressive AI infrastructure investment by any Tokyo-anchored entity. Corporate AI adoption among Japan's $200B-class corporates has moved cautiously. Generative AI startup formation lags Seoul, Singapore, and even Sydney on a per-capita basis. Tokyo has the capital and the talent; the variable that is missing is velocity.
RANK 18
Bangalore
INDIA
COMPOSITE 59.8
India's deepest AI services bench. Infosys, Wipro, and TCS headquartered here, alongside the largest concentration of Google AI engineers outside Mountain View. Indigenous generative AI startup formation is accelerating — Krutrim and Sarvam are the most-cited examples — but venture capital is still mostly imported. The Bangalore-Hyderabad axis collectively forms India's full AI stack, from research through production scale.
RANK 19
Riyadh
SAUDI ARABIA
COMPOSITE 58.2
The sovereign-acceleration story. HUMAIN — the Public Investment Fund's AI subsidiary launched in 2025 — anchors a $100B+ AI commitment. Compute, energy, and capital are abundant. The Saudi-UAE compute corridor consolidates Gulf AI infrastructure into a single regional system. Talent and frontier anchor presence are the gaps, being addressed via aggressive recruitment and visa reform. Riyadh's trajectory has the most explicit upward path of any city in Tier III.
RANK 20
Stockholm
SWEDEN
COMPOSITE 56.7
The Nordic stronghold. Spotify's voice AI work, Klarna's enterprise AI deployment, and the KTH Royal Institute of Technology talent pipeline give Stockholm an outsized share of European applied AI relative to its population. Sweden punches above its weight in industrial and applied AI research. The constraint is scale — Stockholm cannot match the talent density of the Tier I capitals, and its capital markets are not deep enough to retain late-stage AI founders.
RANK 21
Amsterdam
NETHERLANDS
COMPOSITE 54.3
Continental Europe's most efficient AI cluster on a per-capita basis. Strong policy fluency, deep enterprise AI adoption inside Dutch corporates (the ASML adjacency matters), and the cleanest English-language operating environment on the continent. No frontier-lab anchor; the city is positioned as a European destination for execution and deployment, not for invention. Booking.com, Adyen, and Picnic anchor commercial AI use cases.
RANK 22
Los Angeles
USA
COMPOSITE 52.9
The largest U.S. city without a frontier-AI anchor ecosystem. The entertainment-AI cluster — Runway, ElevenLabs (LA operations), Disney AI — is growing but small relative to LA's overall economy. The structural challenges compound: Hollywood lost roughly 45,000 jobs between 2023 and 2025, downtown office vacancy reached 34%, and LA County posted a net population loss of 54,000 in 2025. A turnaround scenario exists; it requires a frontier-tier lab to plant a flag.
RANK 23
Zurich
SWITZERLAND
COMPOSITE 51.4
Premium-niche AI. Home to the largest Google AI office outside Mountain View, alongside major IBM, Disney Research, and Microsoft labs. ETH Zurich ranks consistently in the global top five for AI research output. Zurich's AI economy is exceptionally high-quality but structurally small. Strong on talent per capita; constrained on absolute capital and infrastructure scale.
RANK 24
Montreal
CANADA
COMPOSITE 50.6
The intellectual cradle of modern deep learning. Yoshua Bengio's Mila is one of the most cited AI research institutions in the world. The commercial constraint is talent retention: Montreal-trained AI engineers and founders are systematically recruited to San Francisco, New York, and Toronto. Research lineage is unmatched at this rank in the index; capital and commercial scale are not.
RANK 25
Sydney
AUSTRALIA
COMPOSITE 47.2
The Asia-Pacific periphery. Atlassian's AI work, CSIRO Data61 research, and a strong applied AI consulting base give Sydney a credible domestic AI economy. The structural drag is geography — Sydney sits a 14-hour flight from the Bay Area and an 8-hour flight from Singapore, and AI capital concentrates closer to its anchor labs. Sydney's path to Tier II runs through Asia-Pacific compute partnerships and a sovereign AI policy commitment.
08 / RESHUFFLE
The Numbers Behind the Reshuffle

Six numbers — drawn from the Census Bureau's Vintage 2025 estimates, CBRE Americas Consulting, Counterpoint Research, Startup Genome, and Barclays Eagle Labs — describe the global reshuffle of 2025–2026 better than any single narrative does.

SAN FRANCISCO POPULATION 2025
+0.62%
The only one of America's four legacy urban cores actually growing. Office vacancy fell 3.7 points — the steepest annual drop since 2011.
LOS ANGELES COUNTY 2025
−54k
Largest numeric population loss in the United States. Hollywood lost 45,000 jobs. Downtown office vacancy: 34%, up from 14% in 2019.
NEW YORK CITY 2025
−12k
Net population loss after two years of growth. Domestic out-migration accelerated to 114,000, up from 94,000. But Manhattan rents hit a record $5,000 median.
UK AI INVESTMENT 2025
£8.3bn
Record AI venture haul. London tightened its grip as Europe's leading hub, with nearly 75% of UK AI fintech companies based in the capital.
UAE AI INVESTMENT 2025–29
$15.2bn
Microsoft commitment to G42 and UAE infrastructure. Pairs with the 5-gigawatt UAE-US AI Campus and Stargate's 200 MW UAE cluster.
BEIJING AI VC SHARE
66%
Highest concentration of AI venture capital of any city on Earth. Compares to Silicon Valley 62%, Toronto-Waterloo 50%, Paris 43%.
09 / METHODOLOGY
How the Index Is Built

The 5W AI City Index 2026 scores 25 cities across six equally-weighted dimensions. Each dimension is scored on a 0–100 scale; the composite is the simple mean of the six scores. Tiebreaks are resolved by Dimension 6 (Citation Share). All financial figures are verified via primary or secondary public sources before publication.

DATA SOURCES
Capital & Talent
Dealroom, PitchBook, Startup Genome, Crunchbase, Barclays Eagle Labs (UK), Tech Nation, LinkedIn Economic Graph, the UK DSIT AI Sector Study 2024, Nature Index, QS and Times Higher Education AI subject rankings.
DATA SOURCES
Infrastructure & Policy
Cushman & Wakefield Global Data Center Reports, Synergy Research, JLL Data Center Outlook, OECD AI Observatory, Stanford HAI 2025 AI Index, Counterpoint Research Global AI Cities Index 2025, national-government AI strategy filings.
CITATION SHARE
Modeled, Directional
Citation share scores are modeled estimates derived from Claude synthesis plus web search results across 60 prompts and five AI engines (ChatGPT, Claude, Gemini, Perplexity, Google AI Overviews). Scores are framed as directional estimates throughout — they are not logged query runs.
VERIFICATION
Source-First
All financial figures (VC raised, GDP, investment commitments, employment counts) verified via web search against primary public sources before publication. Where two reputable sources disagree, the more recent figure is used and the discrepancy footnoted in the appendix.
LIMITATIONS
What This Index Is Not
It is not a livability index, a tax-attractiveness index, a venture-deal-volume scoreboard, or a real-estate forecast. It measures one variable — competitiveness as an AI economy — across six explicitly stated dimensions, with explicitly stated source pools and explicitly directional citation-share scoring.
10 / APPENDIX
Methodology, In Detail

The summary methodology in Chapter 09 covers what was measured and how it was sourced. This appendix covers the parts most likely to be scrutinized: the full prompt universe, the scoring and normalization logic, the citation-share modeling procedure, and the disclosures readers should know about before relying on the rankings for decisions.

A1 — PROMPT UNIVERSE
The 60 prompts modeled
Citation share is modeled against 60 prompts, partitioned into six sub-categories of 10 prompts each. Prompts were drafted to mirror the way buyers, journalists, students, and policymakers actually phrase questions in AI engines — not in any blinded or randomized clinical sense. All prompts are in English; future editions will add Mandarin, Hebrew, Arabic, and Spanish prompt sets.
A — General AI Hubs
  • Best AI cities in the world 2026
  • Top AI hubs globally
  • Where is AI being built
  • AI capital of the world
  • Leading AI ecosystems
  • Best city for AI companies
  • Top AI cities outside the United States
  • Which cities are leading the AI race
  • Best places to work in AI
  • Top AI hubs by region
B — AI Talent & Jobs
  • Best cities for AI engineers
  • Where do AI researchers live
  • Highest paying AI cities
  • Top AI talent hubs 2026
  • Best universities for AI by city
  • Top cities for machine learning engineers
  • Where are AI PhDs working
  • Cities with the most AI job openings
  • Best AI talent pipelines globally
  • Where to relocate for an AI career
C — Startups & Capital
  • Where are AI startups raising the most money
  • Best cities to start an AI company
  • Top AI venture capital hubs
  • Largest AI funding rounds 2025
  • Cities with the most AI unicorns
  • Best AI accelerators worldwide
  • Where to pitch your AI startup
  • AI startup density by city
  • Top cities for early-stage AI
  • Best ecosystems for AI founders
D — Research
  • Top AI research cities
  • Best AI labs in the world
  • Where is frontier AI being developed
  • Top AI research universities by city
  • Where are AI papers being published
  • Best AI conferences by city
  • Top deep learning research centers
  • Cities with the most AI patents
  • Best AI postdoc cities
  • Where AGI is being researched
E — Policy & Government
  • Which governments are leading on AI
  • Best regulatory environments for AI
  • Top AI policy leaders
  • National AI strategies ranked
  • Where is AI policy being made
  • Best countries for AI startups by regulation
  • AI immigration policy by country
  • Where to incorporate an AI company
  • Best public funding for AI
  • Cities most supportive of AI development
F — Infrastructure
  • Largest AI data center hubs
  • Where are hyperscalers building AI
  • AI compute capital of the world
  • Top cities for data center capacity
  • Where is the most AI compute concentrated
  • Best energy infrastructure for AI
  • Top GPU cluster locations
  • AI infrastructure rankings
  • Where to host AI training workloads
  • Largest planned AI campuses
A2 — SCORING & NORMALIZATION
How the composite is built
Each city is scored on each of the six dimensions on a 0–100 scale. The composite is the unweighted mean of the six dimension scores. Tiebreaks at the composite level are resolved by Dimension 6 (Citation Share). Unweighted is the most defensible default for an inaugural edition; any specific weighting introduces editorial bias. A "weighted variant" with reader-defined dimension weights is planned for the second edition.
Within each dimension, raw underlying values (dollars of AI VC, count of AI engineers, gigawatts of data-center capacity, etc.) are rank-ordered across the 25 cities and mapped to a 0–100 scale. Where raw data is heavily skewed — capital intensity and infrastructure especially — the mapping is logarithmically smoothed so a single outlier (e.g., Abu Dhabi's 5 GW commitment) does not crush the rest of the field. The resulting dimension scores are presented in the report rounded to integers; composite scores are expressed to one decimal place.
A3 — CITATION SHARE METHODOLOGY
The dimension most likely to draw scrutiny
Citation share (Dimension 6) is the most novel and most consequential dimension in the index. It is also the one we over-document, because it is the dimension most likely to evolve year-over-year as AI engines themselves change.
What the scores represent. For each of 25 cities, across 60 prompts and five AI engines (ChatGPT, Claude, Gemini, Perplexity, Google AI Overviews), an estimated share of mention is derived from (a) Claude synthesis of representative engine outputs, informed by web search of actual public queries returned across February–May 2026; and (b) cross-referencing against three named third-party datasets — Brookings (2021, 2023, 2025 AI cities reports), Counterpoint Research (Global AI Cities Index 2025), and Startup Genome (2024–2026 ecosystem data).
What the scores are not. They are not logged query runs across the five engines. They are not random-sample audits. They are not blinded in a clinical sense. They are directional modeled estimates, framed as such throughout the report.
Hallucination management. Every Top-5 ranking within Dimension 6 is triangulated against at least one named third-party dataset before publication. Where modeled output disagrees with a credible third-party ranking, the third-party source is given precedence and the discrepancy is noted in the city profile.
Language scope. All prompts and engine outputs evaluated in this edition are English-language. This introduces a known structural bias against Mandarin-dominant cities (Beijing, Shanghai, Shenzhen) and Hebrew, Arabic, Japanese, and Korean ecosystems. The bias is acknowledged explicitly in the Beijing and Seoul city profiles. A multilingual extension is planned for Edition Two.
Timeframe. Citation-share modeling was executed February through May 2026. The underlying engine behavior and citation graphs are continuously evolving; readers should treat the dimension as a snapshot of the field as of Q2 2026.
A4 — LIMITATIONS & DISCLOSURES
What this index is, and what it is not
Inaugural edition. Year-over-year score movement is unavailable in this edition by definition. The second edition will publish a "Reshuffle Report" tracking composite-score change between editions, isolating which dimension drove each city's movement.
Unweighted by default. The composite is a simple mean. A reader prioritizing infrastructure over talent — or anchor presence over citation share — will arrive at a different ranking. The dimension-level scores are published in full so readers can build their own weighted variant.
25-city universe. The universe is geographically distributed but necessarily exclusionary. Hong Kong, Taipei, Munich, Dublin, São Paulo, Mexico City, Lagos, Nairobi, and Jakarta are watch candidates for the second edition. Inclusion does not equal endorsement; exclusion does not equal dismissal.
What this index is not. It is not a livability index, a tax-attractiveness ranking, a venture-deal-volume scoreboard, a real-estate forecast, or an immigration-friendliness ranking. It measures one variable — competitiveness as an AI economy — across six explicitly stated dimensions, with explicitly stated source pools and explicitly directional citation-share scoring. Readers who want a livability or tax ranking should consult the Mercer Quality of Living Survey or the OECD Better Life Index respectively. Readers who want a startup ecosystem ranking should consult Startup Genome's annual report.
Conflicts of interest. 5W is a public-relations and AI-communications firm. 5W has served, or may serve, clients headquartered in many of the cities ranked in this report. Client relationships do not influence ranking outcomes; the index is built by the 5W research team using the source pools and methodology described above, and is reviewed by an internal editorial standards group before publication.