Why Breakthrough Discoveries Rarely Become Companies

1 day ago
4 min read

Scientific discovery has never been more productive.

Each year, millions of publications expand the frontier of knowledge across neuroscience, biotechnology, artificial intelligence, and advanced materials. Global R&D investment now exceeds trillions of dollars annually, reflecting a sustained commitment to innovation at scale .

And yet, only a small fraction of these discoveries ever become companies.

This gap is not marginal. It is structural.

The paradox of modern science: abundance without translation

At first glance, the relationship between science and innovation appears straightforward: more discoveries should lead to more technologies, and ultimately more companies.

In practice, the opposite pattern emerges.

As illustrated in the report (see Figure 1), scientific output has increased dramatically over the past decades, while the number of research-based startups has remained comparatively limited .

The majority of discoveries:

remain confined to publications
stop at early experimental validation
fail to evolve into scalable technological systems

This does not reflect a lack of scientific value. It reflects the difficulty of translation.

Beyond the “valley of death”: a structural misinterpretation

The transition from research to commercialization is often described as a funding gap—the so-called “valley of death.”

This framing is incomplete.

The report highlights that the core issue is not primarily financial. It is architectural.

A discovery does not automatically become a technology.A technology does not automatically become a product.A product does not automatically become a company.

Each transition requires different:

capabilities
institutional environments
development logics
forms of capital

When these transitions are treated as linear, projects fail not because the science is weak, but because the structure required to sustain development is missing.

The maturation gap: where most projects stall

One of the most critical—and underestimated—bottlenecks lies between discovery and technology.

A scientific discovery demonstrates a mechanism.A technology must function as a system.

This distinction is central.

As detailed in the report (Section III), technological maturation requires:

reproducibility across experiments
robustness under varying conditions
engineering stabilization
system integration

The Technology Readiness Level (TRL) framework illustrated on page 12 makes this explicit: most discoveries remain at very early stages (TRL 1–3), while startups typically emerge only once technologies reach intermediate maturity (TRL 4–6) .

This gap is not trivial. It is often where projects stop.

Not because they lack potential—but because they lack structure for maturation.

Why startups do not emerge automatically

Even when a technology becomes viable, venture creation is not guaranteed.

Startups emerge only when multiple conditions align simultaneously:

a sufficiently mature technology
a structured development pathway
a capable founding team
access to appropriate capital

The report’s framework (Figure 3) shows that these elements must converge—not sequentially, but in coordination .

Timing becomes critical.

Too early → the company is built on unstable scientific foundations
Too late → the opportunity remains trapped in academia

In both cases, failure is structural, not accidental.

Deep technology changes the rules entirely

Scientific startups—particularly in deeptech—operate under fundamentally different constraints than traditional startups.

As detailed in Section V:

development cycles often extend over 10–15 years
capital requirements are substantial
technical uncertainty persists long into development

The chart on page 19 illustrates this clearly: deeptech ventures require prolonged capital investment before reaching commercialization, unlike software startups which generate earlier revenue signals.

This has several implications:

progress is discontinuous (milestone-driven)
validation takes time
capital must be staged carefully

Most importantly:

Capital does not reduce uncertainty—it amplifies it when deployed too early.

The organizational challenge: beyond scientific excellence

Scientific expertise alone is insufficient to build a company.

As technologies mature, organizations must integrate multiple domains:

science
engineering
regulatory strategy
product development
capital planning

Section VI highlights that successful ventures are those capable of interdisciplinary coordination, not just scientific innovation .

This coordination is rarely natural.

It requires:

new governance structures
shared language between disciplines
progressive organizational evolution

Without this, development remains fragmented—even when the underlying science is strong.

Ecosystems as hidden infrastructure

Scientific startups do not emerge in isolation.

They depend heavily on their surrounding ecosystems.

As described in Section VII, successful environments are not defined only by:

talent
capital
institutions

They are defined by their ability to reduce coordination friction .

Key factors include:

proximity between researchers and investors
availability of specialized capital
density of experienced founders and advisors

In such ecosystems, venture formation becomes faster—not because discoveries are better, but because coordination is easier.

A different way to understand innovation

The central contribution of the report is to reframe innovation as an architectural problem.

Scientific startups emerge when four systems align:

Scientific architecture – knowledge production
Technological architecture – maturation and stabilization
Venture architecture – organizational structuring
Capital architecture – long-term financing

This framework is summarized in Figure 5, where startup creation appears as the result of system alignment—not a single transition .

Most failures occur when these systems evolve in isolation.

Successful ventures emerge when they converge.

Conclusion

The rarity of scientific startups is often misunderstood.

It is not due to a lack of ideas. It is not due to a lack of funding.

It is due to the difficulty of coordinating multiple systems over time.

Understanding this changes the way we approach innovation:

from pipeline → to architecture
from discovery → to coordination
from opportunity → to structured development

In this perspective, scientific entrepreneurship is not simply startup creation.