The keynote by Karina Sotnik, Founder and CEO of WorldUpstart and a founding member and president of the board for WorldUpstart Impacts, a non-profit. Having worked with more than 150 deep-tech startups worldwide, Karina offered a clear diagnosis of why scientifically strong companies often stall — and what founders can do before failure becomes inevitable.

Great science doesn’t fail because it’s hard.
It fails because the venture around it is still a fragile social system.

The myth that science guarantees success

One of the most persistent assumptions in deep tech is that once the science works, success will follow. Karina challenged this directly. While strong research is necessary, it is only one ingredient in a much larger system that determines whether a company survives long enough for timing to work in its favour.

Drawing on long-term data from venture studios and innovation programmes, she highlighted five classic ingredients of company success: idea, team, execution, funding, and timing. Yet the data shows a surprising result — timing accounts for the largest share of success, and it is the one factor founders cannot engineer.

This raises a critical question: if timing is out of your control, what allows a deep-tech company to survive until timing aligns?

How deep-tech companies really fail

Deep-tech startups rarely fail loudly. There is no single dramatic moment. Instead, momentum slowly leaks away.

Karina described a familiar pattern:

• pilots fail to convert into revenue
• decisions take longer and longer
• communication with investors slows down
• meetings increase, progress decreases

By the time failure becomes visible externally, the internal breakdown has often been happening quietly for months — sometimes years.

The four invisible killers of deep-tech ventures

Based on recurring patterns across ecosystems and sectors, Karina identified four hidden failure modes that repeatedly undermine strong science.

1. The narrative gap
Founders struggle to translate complex science into a story that investors can repeat in under a minute. Without a clear narrative, companies become trapped in grant cycles and fail to unlock venture-scale capital.

2. Decision paralysis
Scientific training rewards being right. Markets reward speed. Deep-tech teams often fail not because decisions are wrong, but because they are made too slowly.

3. Role–stage mismatch
This is the most damaging pattern Karina sees. Inventors create value by being right. CEOs create value by making decisions under uncertainty. When inventors try to remain scaling CEOs, companies stall — not due to ego, but identity and fear of letting go.

4. The scale illusion
Early interest is mistaken for traction. Pilots are mistaken for revenue. Revenue is mistaken for scale. These misreads lead to premature hiring, expansion, and resource burn.

Why leadership matters more than technology

As deep-tech companies grow, investor focus shifts. After early validation, investors stop investing in technology — they invest in decision-making systems.

They look for teams that can:

• make fast decisions with incomplete data
• challenge each other without breaking trust
• adapt roles as the company evolves
• confront uncomfortable leadership transitions early

Strong science cannot compensate for weak leadership dynamics. In fact, leadership risk is often the primary reason investors walk away from otherwise promising ventures.

Diagnosing problems before it’s too late

Karina emphasised the importance of early diagnostics — especially around team structure and decision-making. By the time a startup fails publicly, the team has often been failing quietly for a long time.

Founders were encouraged to reflect on four key questions:

• Who truly owns market reality inside the company?
• Do we have the right CEO for this stage of growth?
• How fast can we make decisions when data is incomplete?
• Is there real permission to challenge decisions regardless of title?

Addressing these questions early can prevent slow, invisible breakdowns that are difficult to reverse later.

Building science ventures that scale

Winning deep-tech companies do not eliminate uncertainty — they manage it better. They listen faster, decide faster, and adapt faster. Over time, this allows timing to shift from a risk into an advantage.

The strongest ventures consistently:

• place the right people in the right roles for the current stage
• focus on solving real problems, not showcasing technology
• avoid premature scaling by reading the right signals
• treat leadership as a system, not a title

Karina closed with a reminder that while founders cannot control timing, they can engineer resilience. Science fails not because it lacks brilliance, but because the venture around it is not built to carry it forward.

For deep-tech founders, the path to impact does not start with better science — it starts with better teams, faster decisions, and leadership that evolves as the company grows.