From the lab to the factory: the challenge of scaling semiconductor innovation

AI-generated summary

The seminar highlighted that scientific breakthroughs in semiconductors are only the beginning; the real challenge is turning lab prototypes into industrial-scale products. Deep tech startups now explore areas beyond CMOS—integrated photonics, neuromorphic and quantum technologies, new materials and architectures—creating new opportunities but also raising difficult scale-up problems. After early research and prototype funding, many startups hit a “valley of death” where industrial validation, repeatability, specialized equipment, customer relationships and much larger investments are needed. Investors often favor design and software firms, leaving hardware and manufacturing innovations underfunded and creating value-chain imbalances.

Speakers stressed the critical role of intermediate industrial infrastructure—pilot lines such as Barcelona’s InnoFab—to test technologies under near-factory conditions without prohibitive upfront costs. Success also depends on building trust through industrial testing and customer validation, and on startups shifting from academic mindsets toward priorities of reliability, automation and mass production. Finally, scaling requires not just capital and equipment but specialized talent and training for technicians and engineers; without these resources, industrial expansion and commercialization of next‑generation chip technologies will be constrained.

Transforming a scientific breakthrough into a scalable semiconductor product requires overcoming the complex process known as lab-to-fab. Industrial infrastructure, financing, market trust and innovation ecosystems are essential for deep tech startups to cross this valley of death.

The development of new technologies in semiconductors rarely ends in the laboratory. The real challenge begins when a scientific idea must be turned into an industrial product capable of being manufactured on a large scale.

This process – known in the industry as lab-to-fab – was one of the central topics of the Bankinter Innovation Foundation’s Future Trends Forum on Semiconductors . Participating in this session were Antonio Mesquida, CEO of Tech-Diligence; Francisco J. Díaz, founder and CEO of SPARC Foundry; and Jekaterina Viktorova, CEO and co-founder of Syenta.

The three experts discuss the obstacles that deep tech startups face in scaling their technologies, the role of intermediate industrial infrastructures, and the importance of building ecosystems that enable advanced research to be transformed into viable technology companies.

The conversation reveals a key reality of the sector: scientific innovation is only the first step on a much longer road to industrialization.

The rise of deep tech startups in semiconductors

The semiconductor industry is experiencing a unique moment. For decades, technological progress was mainly focused on the miniaturization of the transistor within CMOS technology.

Today the panorama has expanded considerably. Along with the evolution of CMOS, new technological areas are emerging, such as:

  • integrated photonics,
  • neuromorphic computing,
  • quantum technologies,
  • new materials and chip architectures.

Numerous technology startups are emerging in these fields, especially in Europe. Many of them work on chip design, new manufacturing processes or specialized equipment.

This technological diversification opens up unprecedented opportunities for new industrial players. Leadership no longer depends exclusively on reaching the most advanced node, but also on developing differentiating technologies in emerging areas.

However, turning these innovations into sustainable companies remains extremely difficult.

The scale-up “valley of death”

One of the biggest challenges for deep tech startups is overcoming the well-known “valley of death” of technological scaling.

The early stages of innovation are usually relatively well covered: university research allows prototypes to be developed and early investors can support the creation of technology demonstrators.

The problem appears in the later phase. To move from a functional prototype to an industrial product, it is necessary to:

  • validate technology in industrial environments,
  • demonstrate reliability and repeatability,
  • access specialized equipment,
  • build relationships with customers and industrial partners.

In the case of hardware technologies, these requirements imply much larger investments and longer development cycles than in software. For this reason, many investors prefer to finance design or software companies rather than companies focused on equipment or manufacturing processes.

The result can generate imbalances within the value chain, where some areas receive more funding than others equally essential to the ecosystem.

Industrial infrastructure: the role of pilot lines

To bridge the gap between laboratory and industry, experts highlight the importance of creating intermediate infrastructures for technological scaling, known as pilot lines.

These facilities allow startups and research centers to test their technologies in an industrial environment before making the leap to commercial production. They can work with real manufacturing equipment and validate processes in conditions close to those of a factory.

An example of this approach is InnoFab, a project under development in Barcelona aimed at creating a pilot line for beyond CMOS technologies. The facility will include:

  • a laboratory environment for early-stage technologies,
  • an infrastructure for more mature industrial processes,
  • specific spaces for startups,
  • access to advanced equipment.

The goal is to facilitate the transition from low Technology Readiness Levels (TRLs) to near-commercialization technologies.

For many hardware startups, access to this type of infrastructure is critical. It allows you to demonstrate the technology’s performance under industrial conditions without making initial investments of tens of millions of euros.

Trust, data and customers: the keys to entering the industry

The semiconductor industry is characterized by a strong risk aversion. Every technological change can affect extremely complex manufacturing processes that require close to 100% returns in hundreds of production stages.

In this context, trust becomes a fundamental asset for technology startups.

Building that trust requires demonstrating the viability of the technology through experimental data, industrial testing, and customer validations. Collaborating with a first industrial partner can completely change the trajectory of a start-up, as it provides credibility with the rest of the market.

In addition, entrepreneurial teams must face a major cultural change. Many startups are born as academic spin-offs focused on research. Becoming an industrial company means adopting new priorities: performance, reliability, repeatability, and production at scale.

From prototype to product: translating science into industrial processes

One of the most complex obstacles in the lab-to-fab process is translating laboratory processes into processes that are compatible with industrial factories.

In academia, researchers often employ experimental techniques tailored to research: manual processes, laboratory tools, or manufacturing methods that prioritize flexibility.

Semiconductor factories operate very differently. Processes must be fully automated, compatible with industrial tools, and ensure extreme repeatability.

This requires many aspects of the technology process to be redesigned before production can be scaled up. In many cases, the challenge is not to prove that a technology works, but to adapt it to an industrial environment capable of producing thousands or millions of devices.

This intermediate step requires process expertise, access to advanced equipment, and close collaboration between scientists, engineers, and manufacturers.

Talent and infrastructure: the critical resources of the scale-up

In addition to capital and technology, the growth of the sector will increasingly depend on two strategic resources: industrial infrastructure and specialized talent.

The global expansion of new semiconductor factories will require thousands of technicians, engineers, and operators capable of working with highly sophisticated equipment.

Training this workforce is one of the great challenges for the sector in the coming years. Factories need not only scientists or researchers, but also specialized technicians capable of operating and maintaining complex production systems.

Without a clear strategy for training and talent development, the sector’s industrial growth could face significant bottlenecks.

See the presentations

To delve into the challenges of the lab-to-fab process, the role of deep tech startups and strategies to scale new technologies in semiconductors, you can watch the presentation by Antonio Mesquida (Tech-Diligence), and the subsequent round table with Francisco J. Díaz (SPARC Foundry) and Jekaterina Viktorova (Syenta) at the Future Trends Forum of Semiconductors of the Bankinter Innovation Foundation:

Antonio Mesquida: “Crossing the Second Valley of Death in Deep Tech” #semiconductors

Antonio Mesquida/Francisco J.Diaz/Jekaterina Viktorova: Pilot Lines-Bridging the Gap #semiconductors

In this conversation, experts discuss how to transform advanced research into real tech companies, what obstacles hardware startups face, and why the development of scaling infrastructures will be key to the future of innovation in the chip industry.