The Next AI Wave Could Fit in the Industrial Space Mexico Already Has

The difference in those figures suggests that real estate occupancy does not necessarily follow the sector’s manufacturing pulse, something particularly relevant to expectations for 2027, since if production rebounds as Mexico’s Secretary of Finance and Public Credit expects, companies could absorb part of that growth within the space they already occupy and only later require more. As a result, a manufacturing recovery would not necessarily translate immediately into further real estate expansion, so the absence of such expansion would not, by itself, be enough to conclude that the expected boost had failed to materialize.

Nor is that divergence limited to the beginning and end of the period. A quarter-by-quarter review of the data found no systematic pattern in which increases in production or exports coincided with increases in occupied space, or declines coincided with reductions⁵. This reinforces the idea that manufacturing and real estate cycles can move at different speeds.

However, real estate growth was not uniform either. Among companies present since late 2020, five manufacturers—Foxconn, Molex, Inventec, Pegatron and TE Connectivity—accounted for nearly 60% of expansions, virtually the same proportion as the five largest among those that appeared later in the series⁶. The growth was also geographically concentrated, with Ciudad Juárez, Guadalajara and Monterrey accounting for nearly three-quarters of the net increase in space.

If this pattern continues, the boost associated with artificial intelligence could drive significant real estate expansion, but concentrated among certain companies and markets rather than translating into a broad-based expansion of the industrial space these activities occupy in Mexico.

For 2027, therefore, the size of the real estate impact will depend on how much of the new associated demand reaches Mexican manufacturing and, of that, how much ultimately translates into new space requirements. To follow the evolution of these activities in Mexico’s industrial market, visit SiiLA Market Analytics or email us at contacto@siila.com.mx.

¹ Source: Mexico’s Secretary of Finance and Public Credit, “General Economic Policy Criteria for the Revenue Law Initiative and the Federal Expenditure Budget Proposal for Fiscal Year 2027.” The Finance Secretary notes that investment associated with the development of artificial intelligence contributed to the momentum of economic activity and trade in high-tech goods in 2026; that U.S. growth is being supported, among other factors, by investment in semiconductors, data centers and computer equipment; and that Mexico is among the economies benefiting from increased exports of capital goods linked to AI. For 2027, it also identifies the recovery of U.S. industrial production and momentum in computer and electronics equipment among the factors expected to support Mexican exports.

² The cohort was constructed from industrial occupiers registered by SiiLA in the Technology and Electronics categories. From that universe, companies engaged in the manufacturing of computer equipment, semiconductors, electronics, and related components were selected, excluding those without an identifiable link to these activities. The analysis covers Aguascalientes, Mexico City, Ciudad Juárez, Guadalajara, Monterrey, Querétaro, San Luis Potosí, and Tijuana. It begins in the fourth quarter of 2020 to maintain a consistent comparison period across the eight markets. Occupied GLA was calculated at the company-building-quarter level, without double-counting the same occupancy, and the cohort’s share was estimated against total occupied GLA in Technology and Electronics across the same markets. The cohort was defined using companies identified in the second quarter of 2026 and reconstructed retrospectively through the fourth quarter of 2020. The results therefore describe the trajectory of that group rather than the universe of companies operating in these activities in each quarter.

³ Sources: OECD, “Promoting the Development of the Semiconductor Ecosystem in Mexico,” Feb. 27, 2026; Mexico's Secretary of Economy, Data México. The OECD notes that semiconductor companies operating in Mexico focus primarily on assembly, testing, packaging, and design. According to Data México, Baja California, Jalisco and Chihuahua were among the five Mexican states with the highest exports of diodes, transistors and similar semiconductors in 2024.

⁴ Sources: INEGI, “Monthly Survey of Manufacturing Industries” (EMIM), 2018 Series; Mexico's Secretary of Economy, Data México; SiiLA REsource. For production, industries 3341, Computer and Peripheral Equipment Manufacturing, and 3344, Semiconductor and Other Electronic Component Manufacturing, were used. Their monthly values were expressed in second-quarter 2026 prices using Mexico’s National Consumer Price Index (INPC) and subsequently aggregated by quarter. For trade, quarterly exports of diodes, transistors, and similar semiconductors recorded by the Secretary of Economy were used; the values were adjusted to second-quarter 2026 prices using the U.S. Bureau of Labor Statistics’ IR21320 Semiconductor Import Price Index, used as a sector-level approximation rather than an index specific to Mexican exports. The missing October 2025 observation for IR21320 was estimated through linear interpolation between the adjacent months. Industrial space corresponds to the occupied GLA of the cohort defined in note ². All three series cover the fourth quarter of 2020 through the second quarter of 2026 and measure different dimensions—production, exports and real estate occupancy—so their comparison describes the relative evolution of each indicator and does not imply a causal relationship among them.

⁵ To assess whether changes in real estate occupancy were associated with changes in manufacturing activity, quarterly changes in the series were compared rather than their levels, to avoid common or divergent trends over time generating spurious statistical relationships. Spearman correlation was used, a nonparametric measure that identifies whether two variables tend to move systematically in the same or opposite direction without assuming a linear relationship between them. Using 22 usable quarterly observations, no statistically significant association was found between changes in the cohort’s occupied GLA and changes in real production of computers, peripheral equipment and electronic components (ρ = −0.232; p = 0.299), or between occupied GLA and real exports of diodes, transistors and similar semiconductors (ρ = −0.169; p = 0.453). As a sensitivity test, production was expanded to also include communications equipment manufacturing, without changing the conclusion (ρ = −0.264; p = 0.236). These results do not demonstrate that the variables are independent or rule out lagged relationships; they indicate only that no systematic contemporaneous association was detected among their quarterly changes during the period analyzed.

⁶ Among companies present in the fourth quarter of 2020, the five with the largest expansions accounted for 59.7% of the increase in space among those that grew. Among the 32 companies that appeared later in the series, the five with the largest amount of space accounted for 59.3% of the total associated with those additions. A later appearance in SiiLA does not necessarily mean that the company began operating in Mexico at that time.