What We’ve Learned From Investing in Deep Tech

A month ago, we published a post outlining our latest thesis update, “We back mission-driven founders who are early in new areas.” 

In it, we expanded on the types of founders and opportunities we’re looking for:

 “We’re interested in those rare opportunities where a founder is creating a brand new category. These opportunities are fringy and emerging (some might call them crazy, weird, whacky or wild). There’s nothing obvious or inevitable about it. They’re building something the market hasn’t yet imagined and doesn’t fit neatly into an existing market category. But if they’re right, then the fringe picks up momentum and ultimately becomes a wave.”

The post mentioned that deep tech / hard tech is very much on our radar, but we haven’t spent much time explaining what this means to us, and what makes it so compelling and different from our typical investment focus.  

For us, deep tech (or hard tech) is an umbrella term for tech that touches atoms, and not just bits. There’s a physical (hardware) component beyond the digital (software). Such opportunities typically involve years of R&D in science, physics and engineering. And given the fact that we’re looking for transformational ideas, there’s a slight possibility that the vision isn’t even possible! 

Over the past decade, we’ve invested in many deep tech startups across a range of categories including: 

  • Autonomy: Pickle Robot (2019) which unloads trucks; a startup (2024, not yet announced) that is building an autonomous robot that can traverse extreme terrains for industrial data collection; and Dabble (2014), which back in the day was called Perceptiv Labs and building autopilot for drones.
  • Genomics & Proteomics: Gencove (2017), a platform for low-pass sequencing; and a startup (2023, not yet announced) that is building Illumina for proteins.
  • Diagnostics/Biology: Qvin (2017) which analyzes women’s health via menstrual blood; and Scanwell (2018) which before its acquisition enabled smartphone-powered at-home medical tests.
  • Climate/Energy: Moment Energy (2020) which is building energy storage systems by repurposing retired EV batteries; and EnergyBank (2021) which is designing long-term energy storage for the offshore wind market.
  • Materials: Shellworks (2022) which is creating biodegradable materials as an alternative to plastic; and Bedrock Materials (2024) which is replacing lithium-ion with sodium-ion for batteries.

What have we learned from investing in deep tech? 

1. First, the founding team really matters

Considering the product development timeline is so long, deep tech startups need passionate mission-driven people at the helm. Beyond passion, deep tech founders should have a strong hardware/science background if that’s core to their business and domain expertise, especially if there are regulatory hurdles to overcome. And as always, a stellar founder is an incredible storyteller, who’s able to break down complex ideas for the layperson – and bonus points for founders with a strong branding instinct to position themselves as a category leader.

And, we’re particularly interested in founders who are focused on commercialization, versus a technology solution looking for a problem.

2. Defensibility is different for hardware and software

We used to say that we’d invest in companies that touch hardware as long as the main value creator is in the software or the data collected. But that’s no longer the case. We’re comfortable with the idea that “hardware doesn’t get commoditized over time.” We think about where value is being created, captured, and accrued… and how a company is on top of all that to stay defensible. In other words, it’s very likely that full vertical integration is the way to go!

The path to defensibility differs for software and hardware/deep tech companies. Hardware is hard to start up and scale; but once you scale, you have the opportunity to build a highly defensible company. This is the opposite of software where starting up is easy but defensibility is hard (unless there are true network effects). 

With all this said, it’s yet to be determined how deep tech fits into the traditional venture capital model. Scaling hardware is incredibly capital intensive with all the R&D, manufacturing supply chain, logistics, implementation, and professional services involved. 

3. We want to be early advocates
As for our role, we believe that being the earliest advocates is the right strategy but we also acknowledge that we’re probably not the right investors for startups that are at the “proving the science” stage since we don’t have the domain expertise. 

4. Hard tech is hard
Not surprising but we’ve learned that therapeutics are tough! Regulatory matters are no joke. Clinical trials take a long time, and the success rate in translating from animal models to human models is quite low (<10%).

Consumer hardware is also hard. End products need to be flawless as consumers can be unforgiving. By contrast, enterprise customers seem to be a little more willing to work through issues and partner with a vendor to get the solution they need.

5. Friction is decreasing
As we mentioned above, the development cycle for deep tech is long. But today, thanks to AI and 3D printing, for instance, there’s less friction than ever for prototyping and getting the first product off the ground.

As investors, we believe that momentum is higher than ever as capital allocators are looking for alpha in new areas.

Talk to us!

We’re not focused on specific themes within deep tech but instead are excited to hear what founders come up with. We’ve invested in a broad range of deep tech companies over the past decade and learned from the experience. We look forward to backing more amazing founders in new areas because the potential impact today is bigger than ever before.

Read Next