Deep tech has become a catch-all term. Anything sufficiently complex or technical tends to get labelled as such, which makes the category feel broader than it actually is. What gets lost is a more precise distinction about how these companies behave over time.

A more useful way to think about deep tech is not through complexity, but through whether research compounds into defensibility. That is the lens Sasha Vidiborskiy, Partner at Atomico, applies.

Before venture, Sasha trained as a quantum physicist, spending years building qubits in a lab when real-world applications still felt distant. He now invests across enterprise software and deep tech, including developer tools, AI/ML applications and blockchain/web3.

In a recent conversation with our very own Andreas Munk Holm, he laid out a framework that cuts through much of the noise.

For Sasha, deep tech is not defined by category, but by how advantage is built. In his words, it is where “investment in R&D can fundamentally make something more defensible.”

That defensibility is not about creating an unbreakable moat. As he explains, “I inherently don’t believe that there are pretty much any moat in the tech itself.” Given enough time, capital and motivation, most technologies can be rebuilt.

For investors, this reframes the core question: does R&D compound into defensibility? It is not whether something is difficult to build today, but whether it becomes harder to replicate tomorrow.

Timelines in deep tech are hard to predict, particularly in areas like quantum computing. As Sasha puts it, “they’ll give you the answer that is five years away,” regardless of when you ask. This reflects how uncertain and evolving the field still is.

Even with experience investing in quantum, Sasha notes that “we were right on many things, but we were wrong on the timeline.”

The takeaway is not simply that timelines are long, but that they are difficult to estimate with precision, even for experienced investors.

In Sasha’s view, founder quality plays a critical role in deep tech, given the complexity of the problems and the uncertainty involved.

Technical depth is required, but it is not enough on its own. What matters is whether founders are motivated to build companies, not just solve technical challenges. The expectation is to combine strong technical ability with the intent to build a commercial engine that can scale.

In his opinion, operating at the frontier means making decisions in situations where there is no clear path and many possible directions. As he describes it, “it’s very hard to know what the signal is and what is the noise,” which makes prioritisation and focus essential.

Ambition is equally important. The bar is not incremental improvement, but building something significantly better than the status quo.

Investing in deep tech requires a different approach to diligence. The key question is whether something could work. As Sasha explains, “for most of the deepest investments, we do technical due diligence before term sheet,” because that fundamental question needs to be answered upfront.

Once that conviction is reached, commitments are taken seriously. A term sheet is treated as a strong commitment, which means the work needs to be done in advance rather than after.

This also shapes how decisions are made. The focus is on understanding the broader opportunity without getting lost in unnecessary detail, while still maintaining a clear view of the risks. As Sasha puts it, investors need to “optimise for the vision and the upside.”

Deep tech companies are still expected to show progress, particularly when raising capital. As Sasha explains, “you need to show progress,” even if the definition of that progress differs from more traditional companies.

In practice, this means demonstrating progress in ways that may not follow the usual commercial metrics, but still give investors confidence.

At the same time, long feedback loops make outcomes harder to judge. Companies can raise multiple rounds and still fail, which makes it important to separate decision-making from outcomes.

A key shift in deep tech comes from the impact of AI on research and development, making things faster and cheaper. As Sasha says, this now applies beyond software to areas like materials and system design.

This changes what is possible within a venture model. Areas that were previously too slow or too capital intensive are becoming more viable, as development cycles shorten and costs decrease.

Within this context, Atomico is focusing on areas where these shifts are most pronounced, including compute supply chain, robotics, energy, space and materials. These sectors combine deep technical challenges with the potential for significant breakthroughs driven by compounding R&D.

The common thread is that these are areas where new capabilities can unlock entirely new applications, rather than incremental improvements.

At its core, deep tech is not only about solving complex problems, but about convincing others those problems can be solved.

Founders need to communicate clearly across stakeholders, aligning employees, partners and investors around a vision that may not yet exist.

As Sasha explains, this comes down to “simplifying complexity… and telling the story in the right way.”