Funding Deep Tech Startups: Evaluating Science Risk, Team Risk And Your Own Biases

What’s the difference between deep tech startups and others? Do you need a PhD to invest? Is Deep Tech too deep?

Note: this presentation is based on a recent talk — the slides are here.

At SOSV, we fell into Deep Tech early on:

• To start, our founder put maps on computers in the 90’s (pre-Google Maps ), and coined the word ‘cloud computing’ in 1996.
• But things really took a turn when we started HAX, our hardware accelerator program — the world’s first — in 2012, quickly followed by our equally pioneering programs for biotech IndieBio and RebelBio.
• We kept forging ahead since then and have invested in >800 startups. The majority qualify as ‘deep tech’.

Now, why did we believe this category was an opportunity and not stick with SaaS like many of our peers? This post will share some of the lessons we learned from investing at scale in deep tech.

Deep Tech: A Definition

There are many names for it Science Tech, Hard Tech, Tough Tech, Physical Tech, Emerging Tech, Future Tech, Frontier Tech, Extreme Tech, and even Dirty Tech.

The common factor is the technical risk, as they require both science and engineering.

An even simpler definition, however imperfect, would be ‘anything that’s not an app’.

Deep Tech = Anything that’s not (just) an app?

Generally (but not always), deep tech startups involve longer timelines and higher costs.

Why Deep Tech Is Attractive

All startups involve risk, particularly Execution and Timing. Deep tech startups add to this: Science Risk and (sometimes) Regulatory Risk.

So why bother?

• First, because of impact. Most industries — from Energy to Healthcare or Fashion — involve the physical world. To act on the physical world you need to involve not just bits, but also atoms. The UN Sustainable Development Goals on hunger, water, clean energy, or life on land will clearly need scientific innovation at scale.
• Second, because of returns. Today, 1 in 5 of the 500 global unicorns are in deep tech. But the category is still underfunded, and relatively few investors are specialists, which means better deals should be available.

Image: The number of deep tech startups has exploded over the past 5 years

Evaluating Science Risk

Investors new to Deep Tech often ask:

• Do you need a PhD to invest?
• How to avoid the next Theranos?

At SOSV we have invested in dozens of robotics startups, med-tech devices and synthetic biology companies. While our team has years of experience with various technologies, we certainly don’t have a PhD in every specialty.

Here is what we do to evaluate the science & regulatory risks of deep tech:

1. Prototypes: Do they have a working prototype demonstrating the core function? We do not worry if it is ugly, bulky or expensive.
2. Patents: Do they have credible patents or potentially patentable technology? Often, what matters more is the software or algorithms that are better kept as trade secrets, so patents are rarely a must-have for us.
3. Experts: We often reach out to experts for opinions. Having many portfolio companies and founders in deep tech does help.
4. Customers: Do they have serious prospects who understand your technology? This early proof point will do wonders to convince us!

Evaluating Team Risk

If the science checks out and the market looks promising, your biggest risk at early stage is the team.

Here are some ways founders can kill their startup:

1. Can’t execute. Lack of organizational and management skills.
2. ‘Maker Addiction’. In love with R&D rather than shipping.
3. Artists. Artists won’t care about pivoting to a more promising market.
4. Can’t get along. This risk is higher with newly formed teams.
5. Silicon Valley Burn Rate. Don’t move there too early, or at all.
6. Give up early. Low conviction and grit.
7. Poor social skills. There is a limit to what’s productive.
8. Caught up in Startup Theatre. Another event, another week wasted.
9. Harassment. This can break a company too.
10. Death. Accidents, but also depression-induced. Mental health matters.

As the startup transitions from lab to market and starts scaling, founders need to evolve from inventors to engineers, managers and CEO. They might have or learn those skillsets, or hire new team members.

Diversity of thought matters. At SOSV, we use a psychometric test to help founders assess their thinking style, in order to improve their team dynamics and uncover gaps. It scores people on four aspects:

• ‘Left brain’ (rational) = Analytical + Practical
• ‘Right brain’ (intuitive) = Experimental + Relational

Image: A balanced team includes at least one half-Vulcan

Generally, founders score high on the Experimental and Analytical parts, and low on the Relational and Practical ones. This translates into a strong vision and technical chops, but poor management and communication. Those gaps better be noticed and remedied early to be able to build and sell a product.

The Enemy Within: Our Own Biases

We all suffer from various biases — there are too many to count. Deep Tech offers a few specific ones.

To remember them, here are some original names:

1. The Star Wars bias
2. The Ugly Duckling bias
3. The Make-up bias
4. The Darwin bias

The Star Wars Bias

If you’re working in technology, chances are you like science-fiction. We’ve grown up on a mix of old and new tropes — from humanoid robots to flying cars.

Image: Anything that looks like science fiction is suspect

When we come across technology that resembles those fond memories, we might be inclined to invest because it’s exciting, even when it’s the proverbial ‘hammer looking for nails’.

Unfortunately, as the master William Gibson himself said: ‘sci-fi writers are almost always wrong.’

With rare exceptions, the ‘wow’ is often orthogonal to profits. Practical use and costs have plagued many sci-fi ideas from reaching scale, while some more innocuous projects have had massive impact.

With rare exceptions, the ‘wow’ is often orthogonal to profits.

So be sure to validate a real need before committing too much resources into a technology. It happened to us with a 3d printer for fabric. Another famous example might be Lytro, the maker of ‘light field cameras’.

After seeing thousands of projects and investing in over 200 hardware startups, we now tend to prefer B2B projects with identified customers.

For instance, we invested in various types of cleaning robots (e.g. Avidbots, a cleaner & scrubber for commercial floors, which raised $26.6m), inspection robots (e.g. Simbe Robotics, a retail inventory robot that raised $26m), and more recently construction robots (e.g. Rebartek, using robots to assemble reinforcement cages for concrete walls). Looks boring? Great business!

Investor Tip: Try to ascertain demand even if you’re impressed by a technology.

The Ugly Duckling Bias

Early prototypes might look much worse than they are. If the underlying technology is solid, sometimes a cosmetic effort is all it takes to attract funding or customers.

As Billy Zane, an actor-turned-investor said at a recent event: ‘Every company and country should have their own production department.’

The Canadian startup Avidbots built multiple prototypes of their commercial cleaning robot before they reached the glory of their current product. Technology, design and business often advance at a different pace.

Image: Avidbots early prototypes

Often, this bias is even harder to overcome, as the first product is not the ‘real’ one, it is a first attempt full of compromises. It can thus take another product cycle before seeing the swan in full swing. Today, Avidbots has raised $26.6m and has over 100 staff!

Image: Adult Avidbots

This is also a good reminder that the minimum fundable milestone is not always about technology: it could be design, marketing, a new team member, a customer...

Investor Tip: If you can see past an underwhelming appearance will find great opportunities.

The Make-Up Bias

This is the opposite of the above: some prototypes look very slick, when in fact it’s mostly a ‘Wizard-of-Oz’ mockup and the technology is far from ready (or even possible).

Image: Issues remain under the foundation

On the consumer side, many crowdfunded projects are guilty of this, promoting slick videos and handmade concepts as if they were the real thing.

In a nutshell, if you can differentiate a 3d-printed prototype from a pre-series product, you might uncover great investment opportunities before a ‘hockey stick’.

Investor Tip: Don’t be fooled by the make-up, and understand the technologies, processes and timelines involved to get to the next step.

The Darwin Bias

Finally, we have to acknowledge that great ideas are rarely obvious. The examples abound — from PCs (who needs them?) to iPhones and more.

The name derives from the time when the joint Darwin-Wallace paper on natural selection was first presented at the Linnean Society in 1858. It had so little impact that the President of the society summarized the year this way: “The year has not been productive in contributions of interest and value.”

Later, Darwin was also mocked by various newspapers, as his ideas sounded preposterous in a society rooted in Christian beliefs.

Investor Tip: Suspend disbelief and consider unconventional ideas.

Within our portfolio, the hologram company Looking Glass Factory might have the most charming mix of ‘wow’ and practicality. Time will tell how far they will go!

The full presentation on Funding Deep Tech Startups is available below: