Factories Are Going Digital: Why Isn’t Green Innovation Keeping Pace?

A factory can become more digital every year while its green inventions remain sporadic. The question is what turns new technical capability into a sustained effort to use less energy and develop cleaner ways of producing goods. For manufacturers, that requires more than putting software beside machinery: digital and industrial knowledge must come together in the work of invention.

In "Driving Sustained Green Innovation in Manufacturing: The Role of Technology Convergence Between Digital and Real Economy Industries," published in Sustainability, researchers Chuan Sun and Guangfan Sun investigate that connection across 1,756 publicly listed Chinese manufacturers. They find that firms whose patents link digital and conventional technologies tend to maintain stronger green patent activity over time.

Financing, workforce capabilities and the capacity to combine technologies all appear to influence whether firms can keep innovating. The research points to a sharper test for industrial digitalization: whether it helps manufacturers develop cleaner technologies repeatedly, and whether those inventions eventually improve performance on the factory floor.

What separates a digital factory from an innovative one?

Manufacturers increasingly use digital technologies to monitor machinery, coordinate production and manage information. However, adopting a system and creating a new technology are different activities. The researchers focus on the point where digital knowledge enters an industrial invention, identifying a connection when a patent application for a non-digital technology cites a digital technology patent.

This gives the research a sharper focus than a broad measure of digital adoption. A company may use software throughout its operations but draw little on digital knowledge when developing cleaner processes or products. Another may combine digital and manufacturing expertise in ways that reshape its inventions. The researchers ask whether the latter kind of connection is associated with green innovation that continues, rather than appearing in a single burst of patent applications.

Their analysis covers 1,756 Chinese A-share listed manufacturing firms and 11,954 firm-year observations from 2011 to 2023. To measure sustained green innovation, they combine the scale of green patent applications with evidence of continuing activity across overlapping two-year periods. Firms with more links between digital and non-digital patents tend to score higher on that measure, an association that remains statistically significant across the authors' main models and additional checks.

Those results do not mean that every digital patent citation leads to a cleaner factory. Patents capture inventions, including ones that may never be widely used, while citations provide an indirect view of how technical knowledge moves. The finding is nevertheless useful: it directs attention from the amount a manufacturer spends on digital tools to whether it can put digital knowledge to work in a continuing environmental innovation programme.

The resources behind a continuing stream of green ideas

Green innovation is difficult to sustain when a firm cannot fund research long enough to refine an idea, lacks employees who can work across specialties or has little capacity to test new combinations of technology. The study examines all three conditions. Its analysis links greater technological convergence with fewer financing constraints, a larger share of employees holding undergraduate qualifications or above, and higher R&D intensity.

Each measure captures part of a wider organizational challenge. Access to finance can help a company keep developing an invention before commercial returns arrive. An educated workforce may give it more capacity to absorb unfamiliar knowledge. R&D spending can support the experimentation needed to connect digital techniques with established manufacturing processes. Together, these pathways offer a plausible explanation for why technological connections accompany more persistent green patenting.

The measures also require careful interpretation. A degree does not tell researchers whether an employee has the specific skills needed for digital manufacturing or environmental engineering. R&D intensity is a proxy for a firm's capacity to integrate technologies, not a direct test of how effectively its teams collaborate. The study identifies statistical pathways consistent with the authors' explanation; it cannot show precisely how a particular invention moved from an employee's idea to a production line.

The strength of the association differs across firms and locations. It is greater among manufacturers with stronger digital capabilities and more stable top management teams. It is also greater in cities covered by an intellectual property demonstration programme, which the authors use as an indicator of stronger intellectual property protection. These differences suggest that the ability to absorb technical knowledge, maintain a long-term strategy and benefit from inventions can shape the returns from cross-domain innovation.

A manufacturer considering a digital investment can draw a practical question from those findings: who will use the new knowledge, how will they develop it and what will fund the work after the initial purchase? For policymakers, the results raise a related question about whether programmes promoting digital adoption also address firms' financing and skills constraints.

Do the inventions deliver gains beyond patents?

A rise in green patent applications is not, by itself, proof of lower emissions or less energy-intensive production. Recognizing that gap, the researchers examine whether their measure of sustained green innovation is related to corporate energy productivity. They calculate this as annual operating revenue divided by standardized energy consumption, drawing on firms' use of electricity, coal, natural gas and other energy sources.

In that further analysis, covering 11,512 firm-year observations, sustained green innovation is positively associated with revenue generated per unit of energy consumed. The result gives the patent findings a connection to business operations. It is consistent with the possibility that continuing green invention helps manufacturers use energy more productively, although the analysis does not establish that the patents caused the observed difference.

Energy productivity is also distinct from an emissions measure. For instance, a firm can generate more revenue per unit of energy because of changes in prices, its product mix or other business conditions. The study does not track which patented technologies were installed, how widely they were used or how much carbon they avoided.

For businesses and investors, this distinction is consequential. Patent counts may show inventive activity, but decisions about environmental performance need evidence from production as well. Tracking energy use, technology deployment and emissions alongside inventions would make it easier to see whether a promising idea survives the demanding transition from patent application to routine factory practice.

An industrial policy lesson with limits

The study arrives as governments and manufacturers seek ways to advance digitalization and greener production together. It shows that the relationship may depend on what firms do with digital knowledge. Supporting connections among manufacturers, technology providers and researchers could help create opportunities for invention, but the findings also direct attention to finance, workforce capability and the continuity of company strategy.

The combo has implications for developing economies. Manufacturers with limited access to capital or specialist skills may struggle to undertake digital and green investment at the same time. It would be reasonable for development agencies to examine those constraints when designing support, rather than assume that providing new technology will be sufficient. The Chinese results, however, cannot establish which policies would work for firms operating in other institutional and industrial settings.

The research has further limits. Patent citations miss know-how that companies do not patent, including some process improvements developed inside factories. Patent classifications may not perfectly identify an invention's environmental relevance. Because the sample consists of listed manufacturers, its findings may not reflect the experience of smaller, unlisted businesses. The authors conduct several checks on their results, but differences between firms that are difficult to observe could still influence both technological convergence and green innovation.

Future work could pair patent records with factory-level evidence on installed technologies, energy use and emissions. Studies of smaller firms and manufacturers outside China would test whether the same relationships hold where finance, skills and intellectual property systems differ.