Industrial VR adoption: what 750+ projects show
Most of what gets published about industrial virtual reality is promises. This report brings together nine years of measured projects from San Luis Potosi, Mexico, 750+ projects for 130+ clients across 16 countries, to answer with first-party data what the market asks and almost nobody documents.
Methodology
This report aggregates Fyware internal records from September 2017 to October 2026: 750+ virtual and augmented reality projects for 130+ clients, with deployments in 16 countries. Activity series are published as an index (2019 = 100) over 1,400 record units. The index starts in 2019, the first year of full-scale operation: using 2017 or 2018 as the base, with the company just starting, would inflate the multiple. Capability figures are self-reported (the share who answered they could perform their task) and come from client-run evaluations; GIRI platform data comes from the manufacturer official figures (ar-giri.com). No data point identifies a client.
Demand, year by year
With 2019 activity as base 100, 2024 closed at 544: demand for industrial immersive reality multiplied by 5.4 in five years.
Activity index, 2019 = 100. *2026: solid = actual through October (460); light gray = projection at current pace (~553), above the 2024 peak (dotted line). The 2025 adjustment reflects two simultaneous factors: the broad pause in industrial investment amid tariff uncertainty in North America, and an internal commercial reorganization at Fyware. Through October, 2026 is running at an annualized pace above the 2024 all-time high. Source: Fyware internal records.
Exported from Mexico
In 2019, all of Fyware's activity was domestic. In 2024, 45% was international, and since 2017 Fyware has had clients in 16 countries. Germany is the second foreign market, after the United States and ahead of the United Kingdom and Brazil.
Percentage of international activity per year. Source: Fyware internal records.
What the pilot measured on the floor
A global auto parts manufacturer tested digital work instructions against the paper manual across two plants in two countries, in November 2024, with 44 operators in total. At the Kentucky plant, 100% of the group using digital instructions reported feeling capable of performing their activity, against 11% of the group using the paper binder (n=20). At the Nuevo León plant the gap was 87% against 33% (n=24).
The same pilot timed how long the process engineer takes to create ten work instructions: it went from 125 to 45 hours at one plant and from 170 to 60 at the other, 63% and 65% less time.
As an external reference, the meta-analysis by Howard, Gutworth and Jacobs (Computers in Human Behavior, 2021) finds a moderate overall effect (d = 0.541) that fades or soars depending on how well the task fits the technology: d = 0.207 when the fit is poor, d = 0.820 when it is strong.
What an execution error costs
In a 250-person plant, the GIRI digital work instructions platform reports 62.3% fewer execution errors, 37% savings versus traditional methods and US$1.5 million recovered per year out of US$4.1 million in detected inefficiencies (source: ar-giri.com; Fyware is the official GIRI distributor in the Americas). In training, savings measured by Fyware reach up to 60% versus traditional methods.
Questions this report answers
How fast is virtual reality growing in industry?
Measured in Fyware operations, demand multiplied by 5.4 between 2019 and 2024, and international activity went from 0% to 45% in the same period.
Does VR training work better than traditional training?
The cleanest comparison in our own records is on digital work instructions, not on a simulator: in a November 2024 A/B test with a global auto parts manufacturer, 100% of the team reported feeling capable of performing their activity versus 11% with the paper manual, at the better of the two plants (n=20); at the other, 87% against 33% (n=24). On virtual reality specifically, the meta-analysis by Howard, Gutworth and Jacobs (2021) finds a moderate overall effect (d = 0.541) that fades or soars depending on how well the task fits the technology: d = 0.207 when the fit is poor, d = 0.820 when it is strong.
How much money does a plant lose to execution errors?
The measured reference from the GIRI platform is US$4.1 million per year in detected inefficiencies at a 250-person plant, of which US$1.5 million is recovered with digital work instructions.
About Fyware
Fyware is a Mexican virtual and augmented reality company founded in 2017 in San Luis Potosi. It builds industrial training simulators, safety simulators, digital work instructions and immersive experiences for clients in 16 countries. Winner of the FedEx National Award 2020 and creator of the interactive Auggie Awards filter at AWE USA 2025. To evaluate what is viable for your operation, start a conversation or read the virtual reality training guide for companies.
You are welcome to cite this report data with attribution to Fyware and a link to fyware.com.