Deal Memo: Terran Robotics
Investor’s Note: If you are an investor with Rolling Fun in the Q1 fund (before Feb 28th), you are now a proud shareholder of Terran Robotics, the subject of this post! Congratulations on being pre-rich*! (*this is in no way an actual prediction, promise, or guarantee of a good outcome in any way. It’s just a dumb fun thing to say when we make investments. I swiped it from crypto-master Jason Hitchcock.)
Sometimes insights align in surprising ways. Disparate experiences snap together and information transforms into insight, which enables action.
That’s what happened with our investment in Terran Robotics.
Writing a book, watching Star Wars, and going down Youtube rabbit holes all accidentally prepared me for this investment.
In early-stage investing, it pays to spend half your time studying the most future-thinking technologists in the world. Since I’m currently working on The Almanack of Balaji Srinivasan, I have been reading his patterns and predictions for new technologies and companies.
One of his predictions is more robots executing digital instructions autonomously. This includes delivery robots, 3D printers, and autonomous food prep. Imagine fewer and fewer human interventions between a digital direction and an event in the physical world.
A quote from the manuscript of the Almanack of Balaji:
I think in the future most value will be created online, and you'll print it out by invoking robotics. In the 2020’s we're going to see more actual robots in the field. It's going to come pretty quickly, because once a robot can do something you've turned labor into capital. All labor can become capital.
He even calls out a future ideal of architecture as “The ultimate goal is to hit ‘enter’ and have a building built with drones.” From CAD to Castle.
I saw a version of this in The Book of Boba Fett (great show, btw). Skip to 1:22 for the drone part.
The scene stuck with me because it showed the power of a swam of drones (droids in this case) to collaborate and execute on a blueprint.
But is that realistic? How close is drone technology to construction today?
Advancements in Drone Technology
This is where the youtube rabbithole comes in. Quadrotor drones are awesome. They can do some incredible things. I remembered seeing some researchers who had been training drones to pilot themselves through impressive feats:
Part I
Flying through windows narrower than the drone
Flying between other drones
Perching on inverted surfaces
Part II
Drone being chucked into the air sideways
Flying through hula hoops as they’re tossed into the air
Part III
Landing a small drone on a larger drone
Coordinated flight in formations between drones without collisions
And the swarms get increasingly complex and stay coordinated.
…AND THIS WAS ALL TEN YEARS AGO
Currently in China there are formations with 5,000+ coordinated drones following very precise patterns.
So, we have autonomous, nimble drones that can coordinate and act in formation – but can we build with them?
Well, yes.
Here is some footage of researchers using small quadcopter drones to assemble a 18-foot (6m) tower out of 1500 blocks in a lab setting.
Given the size of those drones, those blocks were super light and obviously not building a sound structure. You can’t build a house out of yoga blocks, but it’s a proof of concept from 2016.
Full-stack architecture, Star Wars, and Youtube were all dormant seeds planted in me when we got an email from Bo’s brother Luke that a former employee of his was starting a company and raising an early round.
That company was Terran, which was using “AI and robotics to turn dirt into extraordinary homes.”
Suddenly all of these disparate pieces snapped into a cohesive picture of the future.
Terran is working with huge drones – like 8-10 feet across with six rotors – which will enable them to move larger, heavier materials and manipulate tools. This is the technology we actually need to build strong, human-scale structures.
Rather than trying to replicate our current (mystifying) standard wall structure of timber frame, insulation, and drywall they use drones to execute an age-old building method: earthen walls.
This started small, training the drone and ensuring it can lift the payloads and tools required:
Now, they are able to deliver bricks to specific targets, and are beginning training on using tools (like the impact hammer.)
One of the classic mental models of early-stage investing is “Invest in Lines, Not Dots.” When you see progress over weeks or months, you know the team is executing and progressing, and get a sense of the pace of improvement.
Terran is still early, but watching the technology in the AI Drone space improve continuously over the past 10 years gives me confidence that the rate of improvement will mostly come at AI-and-software pace, not wait-for-another-breakthrough pace.
Stability, agility, and coordination improvements are not questions of hardware, but piloting software. That also means that precision and speed can increase well beyond human limits with AI pilots. It’s conceivable that a well-coordinated swarm of drones could be buzzing around working at speeds we could barely track with our eyes. Like giant hummingbirds darting around doing home construction.
We’re a ways from that, as you can see in the demos. The drones are still moving slowly and carefully as they train. But – critically – as investors in technology we have to imagine how good things can get. What happens if everything goes right?
Speed continues to increase.
Cost continues to decrease.
Construction of structures can happen 24-hours a day.
Construction drones can do more and more tasks (flooring, roofs).
In 20 years we can imagine humans designing structures (or re-using existing designs), then dispatching drone delivery and drone construction robots to complete all structural work with hardly any human oversight or involvement — for a fraction of the current cost.
Why we think Terran is a good investment
Unique, bold vision for the future. We love to see founders and teams who are a few years ahead. Drone construction still sounds a little crazy, which is a good thing. There’s real technology behind this company, technology in the sense of “something that just barely works.”
Accumulating Advantages. As my friend Alex Rubalcava says, we “like to invest in future monopolies.” If all goes well, Terran’s combination of unique software, hardware, data, and relationships should be quite defensible and provide a healthy profit.
Market Size + Expansion. The market for wall construction of residential homes is big ($100B+ yearly). And that’s just where they’re starting. The same technology could do (some) commercial buildings. The technology could evolve to include other trades like flooring, roofing, fencing, etc.
As Balaji noted, “turning labor into capital” is a compelling value proposition. Terran will provide General Contractors (their customers) with cost savings, performance predictability, and the simplicity of working with fewer subcontractors.
The Terran team are new to us but come through one of the most trusted sources in our life, Bo’s brother Luke, a brilliant founder and technologist who used to work closely with the founder of Terran. We feel confident we have a smart, tenacious founding team here.
All startups are risky, and looking for pieces to be de-risked is always a part of our process. In this case, two of the big concerns are “will it sell?” and “will we be allowed to sell it?” Revenue and regulation.
Revenue is already inbound from regional real estate developers who have signed LOIs. For a company this young to have strong customer relationships and signs of interest is a great sign.
Regulation has been addressed by including Anthony Dente to the team. He has literally written some of the building codes that govern adobe structures, and is a huge asset to staying compliant and passing permitting processes.
Attractive Valuation of $5m for their pre-seed round. Terran received some non-dilutive grant funding, which helps them grow and protect our ROI. (Hope there’s more where that came from!)
Verdict: We invested $26,000 from our Q1 fund.
Caveat and Context: Startups are hard. Investing in startups is risky. There are many unknowns. Most companies won’t work out, and this is no exception. It’s a delicate dance to develop conviction in the potential a company has while remembering the base rates of success are low, and our goal is to build a portfolio of bets with big potential which will perform well in concert.
If you are an investor with Rolling Fun in the Q1 fund (before Feb 28th), then you are now a proud shareholder of Terran Robotics! To support the company, please share their website with a general contractor you know or help fill some of the open positions.
To join the fun and be an investor in the next batch of companies we invest in, become an investor in Rolling Fun. The deadline to join the Q2 Fund is May 31. Minimum participation is $4k/quarter, and we are limited to work with accredited investors. Read more and apply on AngelList.