What is Deep Tech? A rigorous definition and the fault lines

Appearing in investment reports around 2013-2014, the term Deep Tech became, in less than ten years, one of the most widely applied labels in global venture capital. And probably one of the worst-defined. Quantum, biotech, nuclear fusion, AI, blockchain, robotics, batteries, hydrogen, neuromorphic: all these categories were Deep Tech at some point under one definition or another. This article restores the canonical definitions and offers a grid to settle the question.

The origin: BCG 2014

The term appears in its modern sense in the report The Dawn of the Deep Tech Ecosystem published by the Boston Consulting Group in 2014. For BCG, Deep Tech designates startups that build on scientific breakthroughs or disruptive engineering innovations, as opposed to startups that combine mature technological building blocks to solve a market problem.

The opposition is clear. A SaaS startup that assembles Stripe, Auth0 and a commercial LLM to offer a service to SMEs is not Deep Tech: it innovates on the model, the packaging, perhaps the distribution, but not on the technology. A startup inventing a new type of qubit, a new shape-memory alloy, or a new neuromorphic architecture, is Deep Tech: its existence rests on a scientific advance it alone exploits.

Hello Tomorrow: the five criteria

Hello Tomorrow, a French organization founded in 2011 and now a European reference on Deep Tech, offers an operational definition in five criteria a company must combine to deserve the label.

• Engineering challenge. The product cannot be built by assembling available bricks. You must solve a complex engineering problem nobody has yet solved.
• Science fundament. The technology rests on cutting-edge scientific research — papers in peer-reviewed journals, foundational patents, doctoral theses.
• Capital intensive. Development requires substantial investment, often tens of millions of euros over several years, before first revenue.
• Long timelines. Technological maturation takes time, five to fifteen years between initial R&D and commercial scale.
• Transformative impact. If the company succeeds, the impact far exceeds the business case: it changes an entire sector, or even opens a new category.

Hello Tomorrow insists that the five criteria must be combined. A company can have a solid science fundament without being capital intensive (rare software case), in which case it is more an R&D startup than a Deep Tech in the strict sense. A capital-intensive company without a science fundament is infrastructure or a commodity, not Deep Tech.

Bpifrance: the institutional French definition

Bpifrance, the French public investment bank, structured its Deep Tech program around a slightly different definition, more oriented towards eligibility criteria for public aid. For Bpifrance, a company is Deep Tech if it satisfies four conditions: a strong link to research, technological barriers difficult to reproduce, significant investment intensity, and a long time-to-market.

The most operational criterion is the second: technological barriers difficult to reproduce. Concretely, this translates into patents on fundamental elements, or engineering know-how such that no competitor could copy the product in less than two to three years, even with equivalent means. That is what differentiates a Deep Tech from a SaaS well defended by a network effect or a commercial moat.

Three distinct categories: SaaS, Tech, Deep Tech

To clarify the boundary, you can distinguish three categories of technology companies.

Classic SaaS

Assembling mature bricks to solve a commercial problem. Salesforce, HubSpot, Notion, Slack. Competition plays out on product experience, go-to-market, pricing, network effect. No major technological risk — the tech is known, usage and scale provide the advantage. Time-to-market: a few months to two years.

Tech

New product with proven components but original architecture or integration. Stripe, Datadog, Vercel. There is technical innovation, but it stays within what the industry knows how to do. Time-to-market two to five years, moderate capital intensity, limited science fundament.

Deep Tech

New science or new engineering that didn't exist before the company. DeepMind before its acquisition by Google, Pasqal on neutral-atom quantum computing, Verkor on European lithium-ion batteries, Mistral on European LLMs, Aledia on microLEDs. These companies could not have existed five years before they were founded — the underlying science didn't yet exist.

The grey-zone trap

Many companies position themselves as Deep Tech when they actually fall in the Tech category. The trap is easy: boundaries aren't always sharp, and the marketing benefit of declaring oneself Deep Tech is real — access to specialized funds, public aid, a positive narrative.

Three questions help avoid the trap. First, would the company be viable if you removed the innovative component? If yes, the innovation is a nice-to-have, not a foundation. Second, can you identify the scientific publications underpinning the product? If not, the science fundament is hollow. Third, how long would it take to replicate the product with equivalent means? If less than two years, it is probably not Deep Tech.

Why rigour matters

Why insist on this definitional rigour? Because Deep Tech has become an investment category in its own right, with its dedicated funds, eligibility criteria, public programs. Diluting the term harms every company in the category: it attracts entrepreneurs without a solid science fundament, creates unrealistic expectations among investors, and ends up undermining the Deep Tech promise with LPs and public decision-makers.

Maintaining a demanding definition is not academic snobbery. It is what allows the category to deliver on its promise: turn scientific breakthroughs into products, create lasting technology champions, give an economy structural competitive advantages. When a company truly deserves the Deep Tech label, it brings value others cannot replicate — that is the whole point.

Deep Tech is above all a scientific commitment with a long horizon and a transformative promise. Without these three elements, you are talking about something else.

— Inspired by the BCG 2014 report