Do More Robots Mean Fewer Jobs and Smaller Factories in Mexico?

The scale of that shift is difficult to observe directly. Mexico installs thousands of industrial robots each year. Still, available statistics do not provide a complete annual series for the number of robots that remain in operation specifically in the automotive industry. To approximate it, REsource reconstructed the stock using installation data from the International Federation of Robotics (IFR) and an estimated 12-year service life⁴.

The result points in the same direction. Between 2020 and 2025, the stock of industrial robots in the automotive industry is estimated to have increased about 42%, from roughly 30,800 to 43,900 units. Because employment did not grow over that period, the difference is even greater when measured against the workforce: estimated robot density rose from about 384 to 565 robots per 10,000 workers, a 47% increase.

That does not mean robots directly replaced the workers who disappeared from the employment figures. The data do not establish that causal relationship. They do show, however, that as the relationship among production, space and labor diverged, the relative presence of robots within the industry increased sharply.

The divergence becomes clearer when looking at how the relationship between space and the people working within it has changed. In 2020, there were about 15 square meters of occupied industrial space per worker in the markets analyzed. Five years later, that rose to just over 20. This is a 34% increase⁵.

Space, however, did not diverge from production in the same way. Between 2020 and 2025, both grew at virtually the same pace, while the relationship between national production and occupied space in the markets analyzed barely changed⁶, showing that the more pronounced divergence occurred not between production and space, but between space and labor.

That difference matters for the industrial real estate market because the data do not show an industry concentrating substantially more production within a smaller physical footprint. Instead, they show an industry that has expanded its space at roughly the same pace as production while requiring relatively fewer workers to operate.

If that trend continues, a more automated industry will not necessarily require fewer square meters. What may change instead is what those square meters contain and the role they play, as machinery, automated lines, storage, circulation space and electrical infrastructure gain relative importance compared with space directly associated with human labor.

That transformation is already visible at some automotive plants in the country, where automation has not meant empty factories or processes performed entirely by machines. Instead, it has advanced in stages and across specific tasks, with varying degrees of automation even among major manufacturing complexes. A study of robotization and employment in Mexico’s automotive industry documented, for example, that Volkswagen had automated about half of its processes in Puebla; Mazda reported roughly 40% in Salamanca; Ford used 270 robots in Cuautitlán and 172 in Hermosillo, while Audi operated about 1,440 machines—including 600 Kuka robots—in the body shop alone at its Puebla plant⁷.

The available evidence for Mexico also does not point to a linear substitution of robots for workers. The same study notes that automotive employment began declining before the pandemic, after peaking at just over one million workers in October 2018. For the 2013–2022 period, its model finds a positive and statistically significant relationship between automation, employment and wages, pointing to a transformation of job profiles rather than an overall decline in employment⁸, as occurred at some of the plants studied, where line operators moved into maintenance technician roles for automated equipment while other workers took on new functions. Automation, then, changes not only how many people a factory needs, but also what they do inside it.

For the industrial real estate market, that changes how future demand should be assessed. If a factory can produce with a different relationship among workers, machines and square meters, employment alone is no longer a sufficient measure of demand for space. Mexico’s automotive industry is already offering a sign of that transformation: its factories can contain relatively less human labor without containing less industry.

Want to learn more about the trends reshaping Mexico’s industrial real estate market? Visit SiiLA Market Analytics or email us at contacto@siila.com.mx.

¹ Sources and coverage: light- and heavy-vehicle production, INEGI’s Administrative Registry of the Automotive Industry of Light Vehicles (RAIAVL) and Administrative Registry of the Automotive Industry of Heavy Vehicles (RAIAVP); employment and hours worked, INEGI’s Monthly Survey of the Manufacturing Industry (EMIM), covering industries 3361 (motor vehicles), 3362 (motor vehicle bodies and trailers) and 3363 (motor vehicle parts). Real estate data come from SiiLA Market Analytics and cover Aguascalientes, Mexico City, Ciudad Juárez, Guadalajara, Monterrey, Querétaro, San Luis Potosí and Tijuana. These eight markets are used because they have comparable, consistent historical series for the entire period analyzed; therefore, the industrial-space figures represent this sample rather than the entire national automotive inventory.

² Between 2020 and 2025, vehicle production increased 28.8% and industrial space occupied by vehicle and auto-parts companies grew 29.6%, while employment declined 3.3%. Over the same period, production per worker rose 33.2% and industrial space per worker rose 34%.

³ Using 2022 as the base year, production increased 16.6% through 2025 and industrial space grew 20.9%, while employment fell 4.8%. Production per worker increased 22.5% and square meters per worker rose 27%.

⁴ REsource estimate based on International Federation of Robotics (IFR) data. The stock is approximated using a 12-year rolling window, consistent with the average service life IFR uses to estimate operational stock. Because an automotive-sector breakdown is not directly available for every year, low, central and high scenarios were constructed for the missing periods. For 2025, the central scenario estimates approximately 43,900 robots in operation; the low and high scenarios place the stock between roughly 41,100 and 46,500 units. These figures are therefore REsource estimates and not an automotive robot stock published directly by IFR.

⁵ Across the eight comparable markets analyzed, square meters of occupied industrial space per worker increased 34% between 2020 and 2025. Using 2022 as the base year to reduce the pandemic period's effect, the increase through 2025 was 27%.

⁶ Between 2020 and 2025, vehicle production per 1,000 square meters of occupied industrial space declined by about 0.6%. The indicator relates national light- and heavy-vehicle production to occupied space tracked by SiiLA across the eight markets analyzed and should therefore be interpreted as a comparative measure of change over time rather than as a direct measure of the physical productivity of the properties in the sample.

⁷ Márquez Stone, V. and Sandoval Cabrera, S. V. (2024), Effects of Automation on Mexican Automotive Employment: 2013–2022, The Indian Journal of Labour Economics, 67, 661–680. The authors compiled the plant-level automation information primarily from Carrillo et al. (2023) and other sources cited in their Table 1.

⁸ The study uses a random-effects model for 14 Mexican states and adapts the robot-exposure methodology of Acemoglu and Restrepo (2020). Its results find a positive and statistically significant relationship between automation, employment and wages during 2013–2022, and the authors point to a transformation of job profiles rather than an overall decline in employment. They also describe shifts in roles between operators and technicians associated with automated processes. The authors caution that limited information on monthly changes in the number of automotive robots complicates the estimation and that their coefficients should be considered illustrative and explanatory rather than definitive.