Who Benefits from Urban AI Industry Density? Firm Value Creation and the Absorptive Capacity Threshold
This study demonstrates that urban AI industry density enhances firm value creation through technological evolution and integration, but primarily benefits moderately endowed firms by compensating for scale limitations while still requiring a baseline level of absorptive capacity to access these regional knowledge spillovers.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
In the modern economy, artificial intelligence is often described as a general-purpose technology, a tool so powerful and versatile that it can reshape almost every industry, from manufacturing to healthcare. However, a persistent puzzle has emerged: while the technology itself is advancing rapidly, the economic benefits are not spreading evenly. Instead, they seem to be concentrating in the hands of a few dominant companies that have the massive resources to build their own AI systems from scratch. This leaves smaller, less wealthy businesses at a disadvantage, unable to access the knowledge needed to use these tools effectively. The central question for economists and policymakers is whether geography can change this outcome. Can simply being located in a city with a high concentration of AI companies help a regular business thrive, or is the gap between the haves and have-nots too wide to bridge?
A new study by researchers at Soochow University in China investigates this exact question by looking at how the density of AI firms in a city affects the value created by other companies. The researchers analyzed data from nearly 52,000 listed companies across 288 cities in China, spanning from 2007 to 2024. They focused on a concept they call "urban AI industry density," which measures how many active AI companies exist in a specific city relative to the size of that city's economy and its overall business landscape. Rather than just counting how many AI firms are present, the study treats this density as a form of regional infrastructure. Just as a road network allows goods to move from factories to stores, a dense AI ecosystem allows complex technical knowledge to flow from specialized AI producers to the diverse companies that need to use it.
The study finds that this regional density does indeed help companies create more value, but the benefit is not distributed equally. The researchers discovered a specific "threshold" of capability that a company must possess to take advantage of this local AI environment. Companies that are very small or lack basic innovation capabilities often fail to benefit, no matter how many AI firms are nearby. They simply do not have the internal skills or resources to understand, absorb, and apply the advanced knowledge flowing through their city. On the other end of the spectrum, very large and highly innovative companies also benefit less from the local density because they already have the resources to build their own AI capabilities internally; they do not need to rely on the neighborhood as much.
The true winners of this dynamic are the moderately sized companies. These are firms that have enough basic knowledge and flexibility to understand new technologies but lack the massive scale to build their own AI systems from the ground up. For these companies, the dense AI ecosystem acts as a substitute for their own missing resources. The study shows that these mid-sized firms use the local environment in two specific ways. First, they deepen their own understanding of AI, learning more about the technology and expanding their internal knowledge base. Second, they successfully combine this new AI knowledge with their existing, non-AI technologies to create new products and services. This process of mixing and matching knowledge allows them to generate significantly more economic value than they would in a city with fewer AI firms.
The research also clarifies how this knowledge actually moves. It is not just a passive "spillover" where ideas float in the air and are accidentally picked up. Instead, the dense concentration of AI firms creates a structured system where knowledge is concentrated, translated into usable forms, and transmitted to other businesses. In cities with high economic density, AI knowledge is converted into modular, practical applications that can be easily plugged into different industries. This translation is crucial because raw AI algorithms are often too abstract for a standard manufacturer to use directly. The local ecosystem provides the necessary intermediaries and standards that make the technology accessible.
However, the study also reveals a limit to what this regional infrastructure can do. While the local environment helps companies broaden their use of AI and connect it to different fields, it cannot fully replace the deep, internal work required to embed AI into the core of a company's operations. The researchers found that while companies in dense areas were good at expanding the scope of their AI connections, the depth of that integration—how deeply AI is woven into the company's fundamental structure—still depended on the company's own internal commitment and management. The region provides the tools and the access, but the company must still do the heavy lifting to reorganize itself to use them effectively.
These findings suggest that the solution to the uneven spread of AI benefits is not just about building more AI companies or giving subsidies to individual firms to buy technology. Instead, it requires building the "transfer infrastructure" that connects AI producers to the rest of the economy. Cities need to foster environments where AI knowledge is not just present, but is actively translated and made available to companies that are ready to use it. For policymakers, this means focusing on the middle ground: helping moderately capable firms cross the threshold of absorptive capacity so they can tap into the regional wealth of knowledge. For business leaders, it highlights that being in the right place is powerful, but only if the company has the basic internal strength to learn from its surroundings. The study concludes that while artificial intelligence has the potential to transform the economy, realizing that potential depends on the specific organizational structures of the regions where companies operate, turning geography into a decisive factor in who gets to participate in the future of work.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.