What it takes to install a Silicon Valley in Europe

Europe is struggling to keep up with the global tech elite. But what is it that makes a location attractive for startups? Researchers in Vienna are investigating digital platforms to identify local networks of skilled professionals that promote the establishment of young companies. Their goal is to work out what measures cities must introduce to become more competitive in the world of software development.

Culture and Society

by Alois Pumhösel

April 13, 2026

In theory, software development does not require a permanent location. Via digital platforms, IT professionals can work together from anywhere. But software companies still tend to form clusters in certain cities. Places such as Silicon Valley in San Francisco or Toronto, London, Berlin, Tel Aviv and Singapore have seen the emergence of IT hubs with different specializations. But what triggered this development? What are the underlying labor market structures and mechanisms favoring such geographical concentrations? And how can one harness them to build new international IT hubs in the face of global competition?

These and other questions will be explored by the researchers in the netidee-SCIENCE project “Skills and the Geography of Software Startups” over the next three and a half years. This basic research is funded by the private netidee foundation via the Austrian Science Fund (FWF). Economic geographer Frank Neffke and his colleagues at the Complexity Science Hub (CSH) in Vienna focus on economic transformation processes and the “Science of Cities”. They are investigating how the distribution of coding skills interrelates with the establishment of young software companies. The team starts by analyzing a unique dataset: the wide range of open-source software created on digital platforms such as GitHub, through which coders worldwide collaborate on joint projects. The highly interdisciplinary research group at the CSH is developing new approaches to extract insights into economic developments from this data.

“Microsoft, Meta, Alphabet, Amazon, Apple: five of the ‘magnificent seven’ – the major U.S. tech companies – focus heavily on software. It’s an enormously important sector where you can also observe in great detail how products are created,” explains Neffke. “If you study the open-source sector, which accounts for a significant portion of all software development, you can see exactly how code is created – right down to who edited which line and when. You can also find out about collaborations within projects, the companies involved, and even the locations of the participating coders.”

In the race for artificial intelligence, the U.S. is in the lead, followed by Europe. A team led by transformation researcher Frank Neffke is investigating which factors contribute to the formation of regional IT hubs and how Europe can attract talent.

Discover more

From the US pioneering era to today’s high-tech world

Neffke has already explored in manifold ways how the availability of specialized labor transforms economic locations. In an earlier project, which began while he held a position at the Harvard Kennedy School, he examined the development of corporate structures in the U.S. from 1850 to 1950 – from family-run businesses that marketed inventions to large corporations that maintained first research laboratories of their own and coordinated a wide range of expertise. Back then already, the studies brought together different levels of analysis – from the micro level of individual inventions to the macro level of regional innovation clusters.

Whereas that project collated a large amount of historical data, Neffke is now focusing on the very latest developments with an eye toward global software development. “In the last 15 years, the field of software production has changed almost as much as the general labor market has for its part over the course of 60 or 70 years,” Neffke notes. “Fields such as front-end development, AI specializations, or cloud computing didn’t even exist 20 years ago.” Neffke thinks it is good news for Europe that an entire industry can be transformed in such a short time, since right now it is lagging behind when it comes to technical innovation. Catching up might very well occur quickly.

For many decades, Silicon Valley has been a hub of digital innovation with a global reach. Accordingly, IT experts settle there, frequently switching between companies, fostering knowledge transfer, and contributing to a vibrant ecosystem. “From a theoretical point of view, it’s not obvious why such strong spatial concentration exists,” says Neffke. “But if we understand how the skills of experts spread throughout the industry, we can also provide an explanation for this enormously successful regional cluster formation.”

Aerial view of Silicon Valley in California, with a massive ring (Apple Park) in the foreground — Silicon Valley in California is the world's leading high-tech hub. It is home to the largest American IT companies, particularly those in the software industry. © Zetong Li, Pexels

Northern Europe as a cloud paradise

So far, Neffke’s research team has used the Stack Overflow platform to learn more about the geographic distribution and dynamics of coding skills. The platform serves users who want to ask a specialist community questions about coding problems. “We could see here, what skills the platform’s users have and where in the world they are located,” reports Neffke. “Cloud computing, for example, is strong in northwestern Europe – in Denmark, Sweden, or the Netherlands. AI hubs, on the other hand, can be found in German cities like Munich and Karlsruhe that have renowned universities of technology.” One of the things the researchers use this data for is predicting which new skills are emerging in what cities.

Researchers are breaking new ground

Recently, however, the use of Stack Overflow has declined sharply. This is due to AI services like ChatGPT, Gemini, or Claude, which are taking over the role of coding assistants. In the new project, Neffke and his colleagues are now working on transferring their approach to GitHub. Launched in 2008 by American IT experts, GitHub is now used by more than 100 million developers. “In this case we’re not dealing with questions and answers about coding, but with actual lines of code within projects,” explains Neffke. “Our goal is to derive the underlying programming skills from these software snippets in an automated way. In other words, we analyze whether a software project is difficult to implement and what skills are required for it.” This type of evaluation is absolutely uncharted territory. New methodological approaches must be developed and tested for this purpose.

The findings will then be combined with data on young companies. “The study will show what capabilities startups have at their disposal – and why that is the case,” says Neffke. “We know far too little about new company projects and would like to get to know them better. In most cases, information about their exact activities is only available once they are no longer startups.” Since many of the companies – or at least their employees in their spare time – develop open-source programs, clear statements can be made about the skills available.

AI helps improve expertise

Another question the research deals with is how increased use of AI in coding work will influence this field in the long term. In a recent publication in the renowned Science journal, Neffke and his team demonstrated that AI-assisted coding is spreading extremely rapidly. They found that by December 2024 30 percent of lines of code in the U.S. were already generated by language models, and the figures in Europe were not far behind. Overall, the researchers have observed a 3.6 percent increase in productivity so far thanks to these new tools. One finding of the study is that AI support makes it easier for coders to expand their skills. “Experienced professionals in particular can benefit from it to venture into new fields of software development more easily,” Neffke summarizes. “This factor could also influence the development of regional clusters.”

Munich and Zurich Ahead of Vienna

In an international comparison, Vienna ranks in the middle of the field when it comes to location pull for highly qualified coders – which is better than Budapest or Prague, but behind cities like Munich or Zurich. “I believe Vienna has good prospects for development,” Neffke notes. “The technological academic programs are of high quality here, and the research landscape is diverse and interesting.” In addition, the high quality of life is a bonus that may gain in importance in the future.

The analyses at the CSH in Vienna are intended to provide information about the specific steps cities can take to attract startups and experts and grow as IT hubs. “Instead of bemoaning the brain drain from Europe, we should be asking ourselves how we can concentrate talent here in one place,” Neffke points out. “After all, skilled workers face many obstacles in the EU, such as switching between pension and healthcare systems or enrolling children in school.” Neffke is convinced that there is a need for adapting regulations, image, and what is on offer: “In addition to its high quality of life, one of the things Vienna stands out for is its excellent selection of multilingual schools, which facilitate family mobility.”

About the researcher

Frank Neffke has been at the helm of the “Science of Cities/Transforming Economies” research program at the Complexity Science Hub (CSH) Vienna since 2021 and has been a professor of Economic Transformation and Complexity at the Interdisciplinary Transformation University Austria (IT:U) since 2025. Until 2022, he was research director at the Growth Lab at the Harvard Kennedy School in Cambridge, USA. In 2025, Neffke was awarded the “netidee SCIENCE” grant, Austria’s most highly endowed privately funded research award for excellent online research. This grant from the netidee internet foundation, administered by the Austrian Science Fund (FWF), provides the project “Skills and the Geography of Software Startups” with a total of EUR 435,000.