Deep Tech startups will shape tomorrow’s industries. But who shapes their impact?

Deep Tech startups develop new batteries, alternative proteins, low-carbon materials, bio-based processes, robotics, quantum technologies, carbon capture systems, or new forms of energy production. Because they emerge from scientific and technological breakthroughs, they are increasingly expected to help address major environmental challenges.

Their impact however, is considered evident as long as they mention these challenges: it is rarely properly and seriously evaluated during the early stages of technological development, precisely when many of the most important choices are still open.

And this matters because impact is not something that appears afterward, once a technology reaches the market. Environmental considerations shape research directions themselves. They influence which technical options are pursued, which materials are selected, which applications are prioritized, which business models become possible, and ultimately which industrial trajectories emerge.

Through work conducted by SoScience with Bpifrance in 2024, several recurring challenges appeared across the Deep Tech ecosystem regarding environmental impact assessment and strategic decision-making.

The problem with impact measurement

One of the first difficulties is surprisingly basic: most Deep Tech startups cannot measure their environmental impact yet.

Unlike more mature companies, they are often still developing prototypes, validating scientific feasibility, or exploring industrial applications. There is no large-scale deployment, no stabilized production chain, and therefore very little measurable data.

This creates an uncomfortable situation. Startups are increasingly asked to demonstrate their environmental impact very early on, while the technologies themselves are still evolving.

As a result, many founders fear accusations of greenwashing if they communicate projected impact rather than measured outcomes.

But this confusion is problematic. Impact evaluation is not limited to measurement. It also includes projection.

At early stages, the relevant question is often not “What impact does this technology already have?” but rather “What kinds of impacts could this technological trajectory generate under different deployment scenarios?”

This is precisely why approaches such as impact pathways become important. They allow startups to explore how technological choices, industrial partnerships, market selection, or deployment conditions may shape future environmental outcomes long before large-scale commercialization.

A green technology is not automatically a positive technology

Another recurring issue is what is commonly called the rebound effect. A technology designed to reduce environmental harm can end up increasing overall consumption or extending unsustainable industrial systems.

One example discussed during the work conducted with Bpifrance was Sublime Energie, a startup developing liquefied biogas solutions. The environmental relevance of such technologies depends heavily on deployment conditions. The company must actively ensure that its solution replaces existing fossil fuel practices rather than simply adding another energy source on top of existing infrastructures. If biogas merely complements diesel-based transport systems instead of substituting them, the overall environmental benefit becomes far less clear.

This illustrates something essential: environmental impact is not only a property of technologies. It also depends on strategic choices regarding deployment, substitution, scale, and integration into existing systems.

But rebound effects are only part of the problem. Several startups interviewed during the study also highlighted what could be called “closet effects”: situations where solving one environmental issue generates new forms of ecological pressure elsewhere.

Highly efficient energy technologies, for instance, may depend on rare materials, complex extraction chains, or industrial processes carrying significant environmental costs of their own. This is one of the reasons why “green” narratives around technology can become misleading. A Deep Tech startup working on sustainability is not automatically producing environmentally positive outcomes across its full lifecycle.

The challenge is therefore not simply to optimize one indicator, such as carbon emissions, but to adopt a broader systemic perspective.

Technologies do not have one single future

One of the most fascinating aspects of Deep Tech is that many startups do not develop products. They develop technological platforms.

The same sensor technology, AI system, material innovation, or biological process may eventually serve radically different sectors.

And those sectors do not carry the same societal consequences.

A sensing technology may help monitor pollution and biodiversity degradation — or support new fossil fuel extraction systems. A robotics platform may improve accessibility for vulnerable populations — or reinforce labor automation dynamics under highly extractive business models.

In other words, impact is not only embedded in technology itself. It also emerges from strategic choices:

• which markets are prioritized,
• which investors are involved,
• which industrial partnerships are pursued,
• and which applications are considered legitimate.

This means that Deep Tech founders are not simply developing technologies. They are actively shaping future socio-technical trajectories, often long before these questions become publicly visible.

An ecosystem still poorly equipped

Despite growing awareness around sustainability, many Deep Tech startups still struggle to find adequate support structures regarding environmental impact.

Investors and public institutions increasingly ask startups to demonstrate positive impact. Yet few provide concrete operational tools allowing founders to actually explore, anticipate, and structure these dimensions during early R&D phases. Most support mechanisms remain centered on awareness and communication, while startups need practical frameworks integrated directly into technological and strategic decision-making.

The difficulty is not that methodologies do not exist. Over the past decade, numerous frameworks have been developed, particularly within European research and innovation policies, to help organizations anticipate societal and environmental impacts, explore alternative trajectories, and integrate impact considerations earlier in R&D processes.

The real issue is that these approaches remain poorly embedded within Deep Tech support ecosystems themselves.

Incubators, investors, technology transfer offices, and innovation support structures remain insufficiently equipped to guide founders through the concrete methodological implications of these questions. As a result, environmental impact is often treated as a reporting exercise rather than as a dimension capable of shaping technological choices upstream. This is especially important because Deep Tech startups occupy a very particular position: they contribute to shaping the industrial infrastructures of the coming decades.

This is where the challenge now lies: not only in training founders, but in transforming the innovation ecosystems surrounding them so they are capable of supporting different technological trajectories from the very beginning.

This article draws on work conducted by SoScience with Bpifrance on the environmental impact challenges faced by Deep Tech startups. The associated report and roundtable discussion (in French) are available in the Resources section.