--- title: "Japan leads AI-enabling exports complexity in global product space; China tops AECP" sdDatePublished: "2026-09-16T04:09:00Z" source: "https://arxiv.org/html/2609.17200v1" topics: - name: "science and technology" identifier: "medtop:13000000" - name: "economy" identifier: "medtop:20000344" - name: "economy, business and finance" identifier: "medtop:04000000" locations: - "Vietnam" - "Japan" - "Italy" - "Germany" - "China" - "United States" - "South Korea" --- Japan leads AI-enabling exports complexity in global product space; China tops AECP Mapping AI Economic Complexity Green economic complexity provides a generalizable framework for examining countries’ productive capabilities in a defined product set. We apply this framework to AI-enabling goods within the full product space, linking current specialization with adjacent diversification opportunities. Using BACI exports for 2007–2023 and 103 AI-enabling goods, we measure complexity-weighted specialization (AECI), product-level adjacent opportunities (AIAP), and average complexity-weighted relatedness of remaining candidates (AECP). In 2023, Japan leads AECI, while China leads AECP; portfolio breadth accounts for much of the variation in raw AECI. Initial raw potential is positively associated with subsequent changes in the AI-enabling export share, but its associations with changes in AECI and specialization counts are not statistically significant at the 5% level. Our contribution is a trade-based assessment of AI-enabling productive capabilities and related opportunities. The results and public dashboard provide a preliminary complement to publication and patent indicators, not a comprehensive measure of national AI performance or a validated forecast of diversification. National AI capability encompasses not only scientific and inventive output but also the productive structures supporting AI’s material infrastructure. Publication and patent counts do not, by themselves, reveal which countries competitively export the equipment, components, and materials enabling AI systems. The WTO’s AI-enabling goods classification provides a basis for examining this production dimension [ 1 ] . We ask which productive strengths countries reveal through these exports and how remaining AI-enabling activities relate to their existing specializations. Economic complexity examines country–product specialization as indirect evidence of productive capabilities [ 2 ] . Product-space relatedness connects existing specializations with potential new activities [ 3 ] . Together, these approaches distinguish the scale of exports from the breadth and complexity of specialized portfolios, and situate remaining activities relative to those portfolios. Research on green economic complexity shows how this perspective can be focused on a defined product set. Mealy and Teytelboym [4] distinguish current green capabilities, adjacent opportunities, and diversification potential, and explicitly note that their complexity index can be applied to any subset of products. This provides a generalizable framework: the selected goods define the domain of interest, while the wider product space supplies the information for assessing capabilities and related opportunities. We apply this framework to AI-enabling goods. Our contribution is to map countries’ revealed AI-enabling productive capabilities and adjacent diversification opportunities within the full product space. AECI characterizes current specialization by its breadth and complexity, AIAP retains product-level candidates, and AECP summarizes their average complexity-weighted relatedness. We first compare current portfolios and remaining opportunities, then examine preliminary conditional associations between initial raw potential and later changes in export composition and specialization. The emphasis is on productive specialization and structurally related opportunities, not a comprehensive ranking of national AI performance. 2.1 Country–product exports and AI-enabling goods We use BACI HS 2007, 2007–2023 [ 5 ] , aggregating exports over foreign destinations. Annual matrices X c ​ p ​ t X_{cpt} cover 228 economies and 5,047 HS6 products in current thousand US dollars. The fixed target set A A contains 103 goods (52 HS4 headings) from the project’s HS 2012 (H4) version of the WTO AI-enabling list [ 1 ] ; all identifiers occur in the HS 2007 matrices. The list is applied retrospectively and does not identify exclusively AI end uses. All products supply the information used to evaluate complexity and relatedness. 2.2 Product specialization and AECI For country c c , product p p , and year t t , revealed comparative advantage and the AI indices’ specialization matrix are All sums in Eq. ( 1 ) use the full annual trade matrix. Zero-denominator RCA is stored as zero; M = 0 M=0 denotes no revealed specialization, not necessarily no exports. Define the AI-enabling export share as s c ​ t AI = ∑ p ∈ A X c ​ p ​ t ∑ j X c ​ j ​ t s_{ct}^{\mathrm{AI}}=\sum_{p\in A}X_{cpt} \sum_{j}X_{cjt} . For comparison, bundle RCA divides this share by the corresponding share of world exports. It measures relative specialization in the bundle, rather than the breadth of a country’s AI-enabling portfolio. The Product Complexity Index (PCI) is taken from the full-product-space pipeline. Suppressing the year index, on positive-degree rows and columns write B c ​ p = 𝟏 { RCA c ​ p ≥ 1 } B_{cp}=\mathbf{1}\{\mathrm{RCA}_{cp}\geq 1\} and let D c , D p D_{c},D_{p} be its diagonal degree matrices. The product score uses the second right singular vector of D c − 1 2 B D p − 1 2} , rescaled by D p − 1 2} , oriented negatively with ubiquity, and standardized across products. Thus PCI uses RCA ≥ 1 \mathrm{RCA}\geq 1 , whereas Eq. ( 1 ) uses the strict rule. We retain and disclose this implementation difference. Let P t ∗ P_{t}^{*} contain all products with finite PCI and define Normalization of q q is across all finite-PCI products, not only AI goods. The AI Economic Complexity Index (AECI) uses binary membership weights and standardizes the sum of normalized complexity across specialized AI goods [ 4 ] . Its decomposition is F c ​ t = D c ​ t AI ​ q ¯ c ​ t AI F_{ct}=D_{ct}^{\mathrm{AI}}\bar{q}_{ct}^{\mathrm{AI}} , where D c ​ t AI = ∑ p ∈ A M c ​ p ​ t D_{ct}^{\mathrm{AI}}=\sum_{p\in A}M_{cpt} and q ¯ c ​ t AI \bar{q}_{ct}^{\mathrm{AI}} is mean complexity among those goods when D c ​ t AI > 0 D_{ct}^{\mathrm{AI}}>0 . Let u p ​ t = ∑ c M c ​ p ​ t u_{pt}=\sum_{c}M_{cpt} be product ubiquity. Following the product-space approach [ 3 ] , proximity and country-specific relatedness density are Self-links are excluded and zero-denominator proximities are set to zero. Density measures relatedness to the country’s entire specialized portfolio, including non-AI goods; it is not a geometric distance in a plotted network. Undefined density is left missing. The AI Adjacent Possible (AIAP) is a set of candidate products with density and complexity coordinates, not a scalar score: No density cutoff is imposed. Let V c ​ t V_{ct} index candidate products with finite density and complexity. The raw potential G G and its standardized AI Economic Complexity Potential (AECP) are AECP summarizes the average complexity-weighted relatedness of remaining opportunities, rather than their total [ 4 ] . It is undefined when V c ​ t V_{ct} is empty. In the reported results, all AI products have finite complexity and all candidate densities are finite, so | V c ​ t | = 103 − D c ​ t AI |V_{ct}|=103-D_{ct}^{\mathrm{AI}} . 2.4 Reporting population and interpretation For AECI and AECP, annual means and sample standard deviations ( n − 1 n-1 denominator) use countries with at least US$1 billion of total exports and at least one specialized product. This reporting and normalization rule does not screen the matrices used for RCA or relatedness. The valid sample contains 140–163 economies annually, including 162 in 2023. Scores are relative to each year’s reference population and complexity scale; they are not absolute units of AI capability. Numerical consistency checks reconstruct listed-product RCA, q q , F F , G G , and their standardizations from the derived tables. They do not independently verify the underlying BACI ingestion, full-space PCI or proximity estimation, or the substantive validity of the measures. 2.5 Descriptive comparisons and longitudinal associations Whole-space ECI is the country-level complexity score from the full-product-space pipeline; it is a comparator and control, not a component of AECI. We compare AECI with its components, ECI, bundle RCA, and AECP. Country-level correlations use paired observations within the eligible annual sample. For a separate descriptive HS6 network, proximity is computed using B c ​ p = 𝟏 { RCA c ​ p ≥ 1 } B_{cp}=\mathbf{1}\{\mathrm{RCA}_{cp}\geq 1\} and averaged over 2007–2023. The fixed backbone combines a maximum spanning tree with the union of up to four strongest links selected per node whose mean proximity is at least 0.60. Louvain communities summarize this backbone. Neither the pooled backbone nor its layout or communities enters the annual AECP calculation, which uses the full annual proximity matrix based on the strict specialization rule in Eq. ( 1 ). For the longitudinal comparison, the outcomes Y c ​ t Y_{ct} are the AI-enabling export share s c ​ t AI s_{ct}^{\mathrm{AI}} , standardized AECI, and specialization count D c ​ t AI D_{ct}^{\mathrm{AI}} . We compare baseline (2007–2011) and follow-up (2019–2023) means using where Δ ​ Y ¯ c = Y ¯ c ​ 1 − Y ¯ c ​ 0 \Delta\bar{Y}_{c}=\bar{Y}_{c1}-\bar{Y}_{c0} and subscripts 0 and 1 denote the two periods. The regressor G ¯ c ​ 0 \bar{G}_{c0} is mean baseline raw potential, not standardized AECP. Annual real GDP per capita y c ​ t y_{ct} is output-side real GDP divided by population ( rgdpo pop ) in Penn World Table 11.0 [ 6 ] ; log ⁡ y ¯ c ​ 0 \overline{\log y}_{c0} is the mean of annual log income, not the log of mean income. Baseline means use eligible years with jointly observed outcomes and regressors; follow-up means use eligible years with observed outcomes. Each regression requires at least three usable years per period and uses ordinary least squares with HC1 heteroskedasticity-robust standard errors. These are exploratory, in-sample associations rather than causal or out-of-sample tests. 3.1 Scale, complexity, and country portfolios AI-enabling goods account for 11.31% of world exports in 2007 and 12.44% in 2023, with nominal exports rising from US$1.52 trillion to US$2.82 trillion. In 2023, the five largest exporters account for 55.2% of these exports. Among the 5,026 products with finite PCI in 2023, the 103 AI-enabling goods have a mean PCI of 0.837, compared with approximately zero for the full product set; 58.3% lie in the highest complexity quartile. These values describe trade in the listed categories, not exports used exclusively for AI. • Notes: The first five rows are the five highest-AECI economies in 2023; South Korea and Italy are included as selected comparators. D AI D^{\mathrm{AI}} is the number of AI-enabling goods in which an economy has revealed specialization, and q ¯ AI \bar{q}^{\mathrm{AI}} is their mean normalized product complexity. | V | |V| is the number of remaining AIAP candidates. AECI and AECP are standardized annually across the eligible country sample. Notes: The first five rows are the five highest-AECI economies in 2023; South Korea and Italy are included as selected comparators. D AI D^{\mathrm{AI}} is the number of AI-enabling goods in which an economy has revealed specialization, and q ¯ AI \bar{q}^{\mathrm{AI}} is their mean normalized product complexity. | V | |V| is the number of remaining AIAP candidates. AECI and AECP are standardized annually across the eligible country sample. Japan leads AECI despite specializing in fewer listed goods than China, reflecting the higher average complexity of its specialized portfolio (Table 1 ). AECI also differs from bundle-level trade specialization: the United States has a bundle RCA of 0.858 in 2023 but ranks fourth in AECI. Across the 162 valid economies, AECI correlates with whole-space ECI at Pearson r = 0.744 r=0.744 and Spearman r s = 0.822 r_{s}=0.822 . Raw AECI, F F , correlates with the number of AI-enabling specializations at r = 0.997 r=0.997 . Thus, although complexity weights can affect individual rankings, portfolio breadt