Machinepower

Machinepower Index / Dossiers / Mexico

Rank 39 of 50 · Tier 3

Rank 37–46 across the scoring rules tested. How this is tested

Mexico

Diffusion Advantage — For nations that cannot build the frontier but can use it faster than anyone: import the models and win on deployment, across the state and the economy, with public consent.

Score3.92
Potential4.83
Bottleneck cost−0.91
Weakest cell: Sovereign Compute at 2.0 Binding layer: Watts Direction: Stagnating Edition: Q3 2026 · updated October 2026

How the Score, Potential and bottleneck cost are worked out: methodology.

The three layers

WattsMaterial power · holds the weakest cell
4.2 / 10

Energy, grids, data centres and silicon — the physical capacity to run models at scale.

Epoch AI records no large AI clusters or data centres here yet.

WeightsIntelligence power
4.6 / 10

Models, alignment capability, standards influence and the talent that produces them.

WillPolitical power
5.7 / 10

The capacity to wire AI into the state and the economy while holding public consent.

Where the score comes from

LayerSub-dimensionBasisScore
WattsPlanning & PermittingAssessed4.8
Firm Power CapacityDataset6.0
Sovereign ComputeAssessed2.0
Hardware & Silicon AccessDataset4.0
WeightsFrontier Model OutputDataset3.0
Alignment CapabilityAssessed3.8
Regulatory StandingDataset6.8
Talent DensityAssessed4.9
WillPolicy CapacityDataset3.5
Public Sector AdoptionDataset6.6
Workforce TransitionDataset6.4
Public TrustDataset Sources differ6.4

Assessed marks a measurement we judged rather than took from a dataset: 4 of the 12 here.

Sources differ marks a cell where two sources disagree by more than a point. Public Trust: Ipsos and Melbourne/KPMG differ by more than a point (7.6 against 6.5); the cell is their average. Trust in AI (7.1) and support for building data centres (5.6) differ by more than a point; the cell is their average.

Sources

CFE; CENACE; Querétaro state government (via DCD, market guide)2025–2026
Planning & permitting (assessed)
4.8/10, assessed

Querétaro's data centres draw about 750 MW, but the state utility's transmission sets the pace and 2025 reforms narrow private supply.

View source →
Stanford AI Index 2026 fig 3.5.2 (international AI commitments)2025–2026
Alignment (assessed)
3.8/10, assessed

Signed the OECD principles, GPAI and the Seoul statement; no AI law or safety body.

View source →
World BankDecember 2025
Business Ready 2025: business location and utility services
Business location 46.5/100 · utility services 75.8/100

A cross-check for planning and permitting, which stays assessed: it scores permits and electricity connections for ordinary firms, not grid-scale connections for data centres.

View source →
Ember2025
Electricity Generation
390 TWh (2025)

Feeds firm power.

View source →
TOP500.org; Epoch AIJune 2026
Sovereign compute (assessed)
2.0/10, assessed

No system on the June 2026 TOP500 and no compute controlled by the state, universities or firms in Epoch AI; set at the floor.

View source →
GitHub Innovation Graph; Oxford Insights; Stanford AI Index (LinkedIn)2025–2026
Talent density (assessed)
4.9/10, assessed

Stanford does not count Mexico. Estimated from 2.9 million GitHub developers and Oxford Insights' diffusion pillar, averaged with an estimate from LinkedIn's AI talent share (0.31% of members), calibrated on the nations Stanford counts.

View source →
Oxford InsightsMay 2026
Government AI Readiness Index 2025
#77 of 195 · 48.3/100

Feeds governance and public-sector adoption; policy capacity is listed separately.

View source →
Ipsos AI Monitor 2026April 2026
Public Trust in AI
66% say AI products have more benefits than drawbacks

The Index scores public trust as 1 + share ÷ 10. Survey of 23,532 adults across 32 countries, March to April 2026.

View source →
IMF2023 data, published June 2024
AI Preparedness Index: human capital and labour market policies
0.136 of a possible 0.25 · 6.4/10 in the Index

Feeds workforce, rescaled across 173 economies.

View source →
Ember2025
Firm generating capacity
108 GW (2025)

Fossil, nuclear, hydro and other dispatchable plant; averaged with generation for firm power.

View source →
University of Melbourne; KPMGJanuary 2025
Trust in AI at work
53% trust AI at work

Second source for public trust, averaged with Ipsos.

View source →
Oxford InsightsMay 2026
Policy capacity
35/100

Feeds policy capacity.

View source →
Public First, AI Global 20262026
Consent to build data centres
46% want more built, 20% fewer

Half of public trust: support for building more data centres, scored as 1 + share ÷ 10 and averaged with trust in AI.

View source →

Bottlenecks

These hold Mexico below its Potential. Its two lowest scores cost it the most: sovereign compute at 2.0 (part of Watts) and frontier model output at 3.0 (part of Weights).

Watts (Material)Sovereign Compute

Assessment

Mexico has a growing data-centre cluster in Querétaro and assembles AI servers for the US market, but the state utility controls the grid and how fast anything connects. Research and models are thin.

The art of the possible

Foxconn's plant in Guadalajara builds Nvidia's GB200 servers, tying Mexico into the American stack. Opening transmission to private investment would let the data centres follow the factories.

Strategic play — Diffusion Advantage

Use the factories to bargain: trade server assembly for access to chips and cloud, and open the grid so compute can be built beside them.

Leverage — the nearest position. No chokepoint. The nearest key is assembly: a share of America's AI servers is built in Mexico, under trade rules Washington can rewrite.

Open Mexico in the live Index → — every source behind every number, and the comparison tool.