Deep Tech Investing 101

Deep Tech startups have been getting more and more attention, but most investors remain cautious due to perceived technical risk, and some notable failures.

Our firm SOSV has been very active in the sector (with hundreds of investments in deep tech hardware and life sciences), and organized a series of ‘Investing in Deep Tech’ seminars in 2019 in SF, NYC, Boston, London and Paris to share experience and viewpoints with other VCs. Here are some of the key ideas.

5 cities, 19 expert speakers, over 700 participant

What Is Deep Tech

Call it Deep Tech, Hard Tech, Emerging Tech, Frontier Tech, Science Tech, Physical Tech, Future Tech, or even… Dirty Tech (?), they have a few things in common:

1. Some technical risk
2. The intersection of two or more disciplines
3. Software combined with hardware, biology, medicine, material science, physics, etc.

Why Invest In Deep Tech

a. It’s now

Tech sectors are converging giving birth to more deep tech startups, supported by tech enablers (cloud, 4/5G, etc.).

b. It touches all industries

Slow-moving sectors have great needs. That also means potential exits to non-tech players (e.g. John Deere bought Blue River Technology, an A.I. and computer vision tech startup applied to weeding, for $305 million)

c. High rewards

And scarcity adds to the acquisition value.

d. It’s a blue ocean

On the opposite, consumer and SaaS are crowded (‘Every SaaS startup has 18 competitors’), and it’s hard for VCs and startups to stand out.

Five Misconceptions About Deep Tech

Not all deep tech startups are high-risk, slow and expensive.

1. Technical risk is high

If space tech, quantum computing or nuclear fusion is what comes to mind, yes.

However, there is more: clean meat, neurotech, digital therapeutics, robotics, and many other categories with a safer path. Nowadays, most early stage VCs want to see the core technology already working ‘in the lab’ (TLR 5-6) unless founders have a solid track record. Once the core tech is proven, technical risk is much lower.

2. It’s expensive

Magic Leap took over $2 billion in investment before it shipped, but not all deep tech startups need that. Some startups can get to market with less than $1 million thanks to efficient resources (e.g. access to expertise, labs or a supply chain), non-dilutive capital (grants, contests, pilots, co-development) and … frugality (especially outside Silicon Valley).

3. It’s slow

Most investors are not comfortable investing in products taking more than 1 or 2 years to ship. Fortunately, many deep tech startups can match that. The take-off phase can take over a year, but might outpace even software companies.

4. Deep Tech startups fail more

While there has been some notable failures, it’s not clear at all if deep tech startups they fail more than average.

In fact, a disproportionate number of deep tech companies cleared the $1B valuation bar: among them SpaceX, DJI, Bitmain, GM Cruise, Magic Leap, Auris Health, Zoox, View, Nuro, Zume, Carbon, Ninebot, Desktop Metal, Proteus, Butterfly Network, FormLabs, Cruise, Rocket Lab, Geek+, TuSimple, OrCam, Cyberdyne.

SoftBank’s Vision fund alone has over 10% of its portfolio in deep tech (including GM Cruise, View, Nuro, Zume, Plenty, Light, Nauto, Brain Corp).

5. You need a PhD to invest

While you might want an expert opinion to validate the science, or have proof the technology works, having an in-house PhD specialized in each field is not possible and many investors rely on external experts.

Five Biases When Investing In Deep Tech

After misconceptions, here are some biases:

1. The SciFi Bias
2. The Ugly Duckling
3. The Pretty Box
4. The Theranos Bias
5. The Darwin Paradox

Let’s get started:

1. The Sci-Fi bias

Funding cool tech (Sci-Fi or else) with no clear demand. Sci-Fi is a poor guide as it’s almost always wrong. Metropolis featured a humanoid robot back in 1927… Most investors will favor startups with clearly identified markets and customers.

SOSV went through its fair share of Sci-Fi investments to learn this. One of them was a 3D printer for fabric using a novel electro-spinning technology (using liquid polymer turned into fiber in a 20,000 volts chamber). It could not find its market before it ran out of cash.

2. The Ugly Duckling

Some startups with great technology do not package it well. Investors who can see past the brown feathers of the young swan can identify opportunities.

As pre-seed investors in hundreds of startups (especially in applied hardware and life sciences), we gained a fair bit of experience there. We also help our startups go from cygnet to young swan thanks to our labs.

Even the first product, however polished, is only a stepping stone to the second product, where all hard-earned lessons are applied.

One example for SOSV is the startup Nura, who created headphones for its personalized audio technology, and is releasing earbuds this year. They raised $21 million in an oversubscribed round in 2018.

Prototype (2016); Stevie Wonder with the first product; Second product (2019)

3. The Pretty Box Bias

This is the opposite of the Ugly Duckling. Some companies do a great job with polishing their prototype or a demo video. This can lead investors to think the final product is almost ready.

What to do? Understand that the level of risk is not linear with time. Figure out the level of development and which milestones are meaningful for the startup to evaluate the opportunity.

• In hardware, setting up a supply chain is a significant milestone. This set up can take months, during which the momentum might not be visible.
• Don’t overshoot either by waiting for a milestone too large: getting FDAapproval is such a large milestone for a medtech company that many get acquired in the process!

In the image below, 18 months separated the early functional prototype (left) from the final product (right). When do you think was the best time to invest and at what price?

Left: prototype (May 2016); Right: product (January 2018)

4. The Theranos Bias

No investor wants to be embarrassed by the next Theranos scam. Validate the tech with experts, including founders of portfolio companies, to get a better sense of the technical risk.

5. The Darwin Paradox

Darwin’s paper did not get any attention in 1858, when it was presented publicly for the first time. It is not always easy to figure out what is hype, what is real, what is under-hyped, especially at early stage when you’re investing ‘pre-trend’. We have to acknowledge the limitations of our predictive powers, and avoid falling into old decision patterns that don’t apply to new models.

Deep Tech Investing Risks

Once technology looks credible and functional, the risks are elsewhere:

1. Team

Among the team risks are:

• ‘Artists’ or ‘Science experiments’: founders fixated on an application that does not have a large enough market.
• ‘Maker addition’: not shipping due to never-ending R&D.

Startups rarely die because they can’t achieve a technical milestone. Highly technical teams will often lack skills on the business front for sales, marketing, management, financing and more. Several speakers at our events emphasized the need to address those areas early, and assess the ‘coachability’ of founders.

It is notable that the personal risk for deep tech teams is higher than in software. It is harder to resume a job in academia or science after a startup failure. The lack of publicized exits (including acqui-hires) also have a negative influence on the willingness to start companies.

2. Go-To-Market

Start with the right target market. Dominating a niche and expanding (the Peter Thiel method) might work better than try to serve the larger but slow-moving segments. Even within a niche, seek active feedback with your more responsive clients while exploring opportunities with other customers.

3. Fundability

Founders need to understand what is the next milestone that will unlock new resources. Sales tend to solve problems. If you’re pre-sales, try and figure out what is meaningful to the business and follow-on investors.

• For their first round(s), most startups will likely work with smaller or local VCs as many ‘brand name’ investors have moved downstream, to late seed, A and B rounds.
• Most VCs investing in deep tech after the seed stage don’t struggle with deal flow. Some have deep connections with some universities, or are well-known for some specific sectors.
• While many VCs focus on some broad sectors, few are ‘thesis-driven’, actively seeking startups solving a particular problem.
• Corporate VCs tend to start investing at series A and beyond. This allows them to leverage both their internal and external customers.
• SOSV has a fairly unique position as a ‘pre-seed’ investor, as startups are often not very visible in the market yet. Our deal flow depends on our own network and visibility.
• ‘Talent investors’ like Entrepreneur First also have a hard task as they support entrepreneurs even before a team is formed (note: SOSV partners with EF on some hardware and biotech projects).

4. Geography

Mature startup ecosystems (e.g. Silicon Valley, Boston) have advantages:

• More complete teams targeting better problems, and creating more fundable companies beyond the early stages.
• Easier access to funding
• Easier exits thanks to simpler integration

Now, some smaller places have better access to local talent, often cheaper and easier to recruit, train and retain. Also, many investors these days are fine with startups that are one direct flight away.

To Conclude

Deep Tech startups offer opportunities to those who take the time to understand their differences. We hope to connect with more like-minded investors, and help improve the physical world, and the health of humans and our planet! For more info, here is a link to our portfolio.

Thanks to the speakers and many participants! In particular:

• SF: @JohnMannes (@Basis Set Ventures), Guy Resheff (@Grove Ventures), @Sunil Nagaraj (@Ubiquity Ventures)
• NYC: @Zack Schildhorn (@Lux Capital), @Dan Robinson (OS Fund), @Matt Turck (@FirstMark Capital), Ahmet-Hamdi Cavusoglu, (Academic Venture Exchange)
• Boston: Matt Rhodes-Kropf (Tectonic Ventures), John Ho (@Anzu Partners), Joyce Sidopoulos (@MassRobotics), Frank Andrasco (@Next47), Alberto Moel (Veo Robotics)
• London: Matthew Scherba (@Breed Reply), @Olivier Huez, (@C4 Ventures), Kerry Baldwin (@IQ Capital), @Chris Barchak (@Next47)
• Paris: Pascale Ribon, (@BPI France), @Cedric Favier (Elaia), @Matthieu Repellin (@Airbus Ventures), Arnaud de la Tour (@Hello Tomorrow)