Deeptech and why it will save the world: Part 2
Defining Deeptech in terms of its risks, talent pools, time to exit and more.
In Part 1, we discussed how creative destruction and global challenges are creating the perfect storm for the rise of deeptech innovation. In Part 2 here, we dive a little deeper into what really makes a startup or technology “deeptech” and how it might not be so easy to define.
Consider the challenges we explored in the previous section. Here’s some ways in which deeptech innovations are already flipping the script on these problems.
Source: The Deep Tea analysis
Defining Deeptech
All this talk around deeptech hasn’t come without confusion. There’s still a lot of debate around what exactly cuts it as a “deeptech” startup. To help cut the clutter, we’re bringing together research from leaders in this space to give you a clear understanding.
Let’s go over some common criteria and features of deeptech startups thrown around in thought leadership on the web and examine why they may (or may not) hold water -
What makes a startup deeptech
Deeptech solutions are multi-disciplinary and combine approaches and technologies
Complex challenges require complex solutions, and deeptech does exactly that. Deeptech solutions bring together different areas like green chemistry, bioengineering, and deep learning to tackle some pretty tough problems. This BCG report, for example, highlights how important it is for deeptech to blend design, science, and engineering. It all starts with pinpointing the problem, brainstorming ideas, leaning on science for the theory behind it all, and then using engineering to check if the idea is technically and commercially sound
Source: BCG
Take the startup, Lanzatech, as an example. They're doing some groundbreaking work by turning waste carbon into useful stuff, like chemicals for everyday products and fuel for planes. They’ve engineered Direct Air Capture techniques that capture carbon dioxide (CO2) and carbon monoxide (CO) produced from industrial processes. This carbon waste is then fermented with a special bioengineered microbe that eats it up to produce Lanzanol, a proprietary blend of ethanol, or other products like Sustainable Aviation Fuels (SAF). This is just one example of how Deeptech startups combine science, engineering, and design to achieve innovative outcomes.
Require IP protection to build a moat
The key to deeptech is owning your own cutting-edge technology or unique processes, all protected by patents or exclusive rights. In fact, 70% of deeptech ventures own patents related to their products or services. Investors really look for this strong intellectual property (IP) protection because it's a sign of resilience. Even when the going gets tough, a company with solid IP has a safety net, ensuring it can weather any storm and maintain stability.
We analyzed Pitchbook data for 100 growth stage startups (50 climate deeptech and 50 B2B SaaS) - we saw that the numerical average number of patents filed by deeptech is 2.8 (so, typically 3) while for the latter, the number is 0.7 (typically 0 or 1). Although a relatively small sample size, this indicates that deeptech startups have a much higher intensity of IP per startup when compared against traditional B2B SaaS industries.
Source: The Deep Tea analysis*
But, this IP comes at a high risk and high reward - once it’s built, the IP can give the company a solid moat or at least a lead time of a few years upon which to build their business. But, it’s extremely capital intensive, making it riskier for investors and other stakeholders to get involved.
High barriers to entry with deeply embedded incumbent
Deeptech startups build in traditional industries, or create their own categories in which there exists a strongly embedded substitute. For example, you can’t go far in biotech without running into bureaucratic hurdles with FDA approvals. Anyone providing clean fuels still faces stiff competition from the likes of Chevron and Shell. Even if you can deliver cleaner alternatives at cheaper costs and convenience, it’s difficult to get demand due to other uncertainties or just good old inertia, as is being reported in the case of green hydrogen. Deeptech startups face a trilemma here: regulatory and intellectual property concerns, competition from large incumbents, and the need for customer education and confidence building.
What's unique about deeptech startups?
Risk
The Valley of Death is a daunting stage in the life cycle of a startup, usually happening after a product has been launched but before any revenue has been generated when it mostly relies on funding to survive. Deeptech startups spend much more time in the valley of death, making them risky ventures (see graph below). This is driven by significant tech risk and long R&D times since these technologies can take a long time to test out and prove.
But the graph also points out what makes deeptech attractive - deeptech investing comes with a bigger payoff, and strong market demand can lead to fast value gains and unique opportunities post-R&D. Plus, disruptive tech often ensures solid competitive edges.
Source: Innovation Industries VC Fund
So why is the Valley of Death so much longer for deeptech cos? Georges L. Romme throws a spotlight on the early stages that cause this. Deeptech ventures usually pop up as spinouts from the brainy halls of university research centers and labs or get a leg up from university incubators. For example, the University of California churns out over 80 spinouts a year, with Harvard, MIT, and the University of Texas not far behind, each pushing out more than 20.
In the beginning, it's like these academic giants have got the startups' backs, offering everything from research grants to co-working spaces and even lab gear, not to mention a treasure trove of academic know-how. But then, suddenly, these ventures have to leave the nest and fly solo. That's when the safety net disappears, and they're out in the big, wide world, hustling for funding, snagging contracts, and trying to prove their tech can actually cut it in the market, all on a shoestring budget and a core team - creating the deeptech valley of death, as shown below
Source: Romme 2022
However, a Pitchbook analysis from Leo Polovets of Humba Ventures found that of the companies that have crossed USD 1 mn in funding, the percentage that has USD 250m+ exits is higher in deeptech vs. traditional sectors. Companies in deeptech areas like defense, space tech, and life sciences have a 2%-5% chance of having $250m+ exits, while companies in categories like SaaS, Fintech, and AI/ML have a 1%-1.5% chance.
Let’s remember there may be a degree of self-selection here. It’s much more difficult for deeptech firms to get off the ground - their science and engineering needs to work, they need to have a robust well-rounded team, and they need to get customers. Once they cross that hurdle, they’re more likely to succeed because of the inherent moat involved. This is shown by the fact that the number of deeptech exits is much lower, creating base effects.
Data from the UK shows something similar - deeptech firms struggle to reach Series A, but securing a second round significantly boosts their chances for further funding success. We can visualize this in the graph below.
Source: Invigorate
Essentially, deeptech firms face higher initial hurdles and risks, but once past the early stage, their unique technological moat and successful progression lead to a higher likelihood of continued success.
Talent
Deeptech firms are pretty picky about who they hire and really value folks with deep expertise in science and engineering. They hire tons of PhDs in their R&D teams, and this shows just how much they rely on advanced education and degrees to solve the knotty problems they tackle. When compared with SaaS startups, they don’t need large teams with Biz Dev and Sales folks but rather a small, core group of engineers/scientists and other specialists, especially in the early stages. According to a McKinsey analysis of deeptech startups in the UK, the average number of staff members is roughly half in deeptech startups as compared to other startups, and they have the highest (46 percent) share of employees in R&D roles while others typically max out at less than 40%. In fact, a deeptech startup's valuation is often dependent on the quality and pedigree of the science talent that they bring
However, this precise need for specialization also presents challenges. The interdisciplinary approach of deeptech, involving a range of fields, can lead to friction among team members due to differing methodologies. This variation in approaches may further amplify the already high risks associated with deeptech ventures.
Innovation Industries, a Dutch deeptech VC fund, suggests that deeptech companies have an easier time attracting candidates, which is coupled with low employee turnover. However, we think this is likely because the market is currently niche and contained. For instance, someone with a PhD in chemistry with a focus on green hydrogen stands out in a small pool of 20-30 green hydrogen startups in the US. But, as the talent landscape grows in both demand and supply, this advantage might not hold. So, in the near future, deeptech talent might become a hyper-competitive pool.
Time to market/exit
Investments in these deep technologies take longer to grow compared to other tech investments—about 25% to 40% more time from the initial funding stages
However, this fact is also contested. Leo Polovets, from Humba Ventures, looked at time to exit, and used Pitchbook data to show that deeptech and life science companies exit 10-25% faster than traditional ones, with exit times for big exits (> $500 m) being nearly 8 years for traditional (software/hardware?) startups, just over 7 for deeptech, and under 6 for life science.
So….what’s not DeepTech?
Shallow tech is a term used to refer to relatively simple technological advancement, often leveraging software and market players for distribution. Some key aspects of shallow tech are -
Shallow tech cos carry more market risk than technology risk
Often price takers in a market where competitors are doing something similar
Even if they start out with a moat that allows them to command price, their models are easy to replicate; demand side and competitive pressures often force them into steep discounts (think Uber)
This is not to say that shallow tech is easy to do, or not relevant. The 5th wave of innovations that initially required deeptech approaches (mobile phones, internet, etc.) have since relied on market-based approaches to reduce costs, distribute and scale up. In all likelihood, in the next 5-10 years, shallow tech will be required to commercialize, distribute, and scale the deeptech innovation of today. Consider the case of solar - Deeptech approaches were a prerequisite to developing initial photovoltaic panels for solar back in the mid to late 20th century. Now though, companies are leading market-based innovations to drive adoption. Husk Power Systems has innovated a unique pay-as-you-go service using a mobile-enabled smart metering system that has allowed it to deliver off-grid solar across 50k houses for lower than $2.35 per W across hard to penetrate markets like India and Nigeria. In developed markets, asset management and digitized solutions for solar are becoming increasingly popular. For example, Fuse is a startup that offers virtual solar panels through renewable energy tokens - making it easier for end customers to adopt solar energy for their homes and manage it through a mobile phone app.
The Deeptech market map - a (WIP) mental model
How can we make sense of this Deeptech universe?
Being part of the Harvard-MIT innovation ecosystem, we’ve had the chance to closely interact with many deeptech startups and founders on their 0 to 1 journey. As we discovered more such innovations each day, we started to think about how they fit into the larger landscape and what that landscape even looks like. There are few market maps of the deeptech space, and the ones that do exist mix both the technology/science that the startup is using and its industry of application. To think about deeptech effectively, it’s important to think about industry AND technology together as critical to a deeptech startup. We hope that this helps people get a 50,000 ft view of the deeptech landscape in <2 mins, and going forward, enable individuals, funds, institutions, and entrepreneurs to identify if a technology is within their area of focus or interest
We intend to further build this market map with input from all our readers - so if you think a startup deserves to be here - DM us on LinkedIn.
