Part 1: Climate tech is ⅔ Hardware
Chapter 3 of The Climate Startup Playbook from Third Sphere
For climate investors, understanding hardware startups, is just as important as understanding climate science. Over the last 25 years, climate scientists have developed a clear roadmap for decarbonization. But during the same period, the vast majority of startup founders and investors have been focused only on software. And that’s a problem because most climate action requires a combination of hardware and software.
The goal of this chapter is to explain why and how founders and investors should think about hardware. And then to explore some approaches to make hardware less hard, while enabling real climate action and rapid scale and adoption, comparable to what we’ve seen for software. We’re exploring themes related to key partners, customers as well as investors that can work alongside venture capitalists.
We looked at publicly available data for portfolios from USV, YC and some other notable generalist investors. And we compared their climate portfolio with their general portfolios. The results are strikingly clear. The majority of climate startups that have attracted investment from leading VCs, we’d classify as not just software. That is, there are some physical components to delivering their product or service.
We know what it takes to build $1b software companies. One common path is getting to $100m in recurring revenue for SaaS. Others are usage or transaction based, where the percentage of online transactions continues to expand. There is real magic in low marginal cost to deliver a product to one more customer. But what does it take to rapidly build a large hardware company?
Some assembly required
Apple sold over $200b of iPhones in 2022. That’s over 150m units. CATL sold $43b of batteries in 2022. Amazon did over $1t in revenue in 2022. We know massive scale is possible when hardware is involved.
When we think about VC-scale businesses, we try to think about a path to $1b revenue within a decade. . Future Motion makes one of today’s most popular micromobility products, Onewheel. To reach a $1b run rate, they need to ship about 500,000 units per year. This was never going to be easy but they’re already approaching this scale. And they’ll likely have other products that can be successful at this level.
Top: 2015 Onewheel image on Kickstarter. Bottom: Onewheel racing on ESPN in 2023.
Resonant Link makes low cost, high efficiency wireless chargers that are increasingly replacing physical connectors which allows for “opportunistic” charging for vehicles along with a bunch of benefits in medical and consumer devices. Ebikes reached 100m units and smartphones over 1b smartphones were sold in 2022. So if Resonant Link can be one of the preferred charging methods, they can achieve the right scale to get to $1b in revenue.
And connected devices like Rachio’s smart water controller, need to reach millions of customers to approach $1b run rate. Like Future Motion and Resonant Link, it’s unlikely there is just one product that gets to this scale. Rachio already has multiple offerings that work together in landscapes to manage water and weather but the point is the scale of units to be shipped, which far from easy, but also not especially scary when modeling addressable markets.
The advantage of these examples is that they’re “some assembly required”. That is, they can be built with existing manufacturing and production processes. So startups can focus on designing, sourcing and assembling versus building out their own manufacturing facilities. But that’s not always possible.
Things that make things
This is often what people have in mind when they say “hardware is capital intensive” or “very risky”. This type of company takes years to develop a product without much revenue and then still requires a $50 million production facility before you can generate revenue.
Within hardware, this remains the most unusual category because it’s by far the largest fundraising challenge — it’s unclear what parts of the capital market can fund large “First of a Kind” production. Even notable multi-time founders like Elon Musk didn’t try this. For example, Tesla’s first $50m plant came only after their first successful product. And it was really a partnership with Toyota who invested $50m and then “sold” Tesla their 1960s era plant for about the same amount.
In our portfolio, Bowery Farming began with smaller farms along with early commitments from customers like Wholefoods. Today they have 7 farms ($1b run rate requires a small multiple of this). If you are making modular concrete, which we invest in for resilience and construction productivity reasons, you need about 2,000 work cells producing rebar cages.
And you probably need 10,000 modular concrete production plants. None of these are terrifying numbers when one considers there might be 1–10 plants per major city. What is most scary about this category is how to fund the first or maybe first few plants. But once you have customers and known unit economics, the range of potential investors expands rapidly.
Hardware as a Service
This is perhaps already the most common hardware model because it describes how most customers “buy” everything from homes and cars and phones and servers. The benefit of leasing or paying as you use something, especially in climate, is that you can pull forward demand. Customers don’t have to understand the technical risk or worry about warranties. Benefit now. Pay later.
Residential solar companies offer a great example of figuring out non-dilutive fundraising as a central way to unlock scale. Far into Sunrun’s annual shareholder presentation you’ll find slide 34 with this comment “Industry leading financial execution. We have a strong track record of attracting low-cost capital from diverse sources.”
This financial execution is what allows Sunrun to install solar and batteries for a monthly fee that is likely lower than buying from the grid. And this leads to a flywheel of declining costs. Shifting to a monthly fee, allows Sunrun to commit to larger orders with their suppliers and therefore drive down the prices they pay, even as cost curves also bring down the cost of solar and batteries. And as they demonstrate the financial performance of their asset (i.e. homeowners paying monthly fees) more lenders want to lend to them so that drives down the cost of capital.
In the Third Sphere portfolio, Cycle turns eBikes into a cloud service. As they did with servers, businesses first purchased their own ebikes, but many are realizing they don’t want to deal with ongoing servicing or maintenance or supply chains that might not have the particular part right now, sorry but you can’t use your bike. Instead, like cloud services, they would prefer to have guaranteed availability of their fleet. If a bike fails for any reason, they don’t need to worry. Within X minutes a new bike is available so operations can continue. Increasingly, when businesses do the math, they realize they should just subscribe to Cycle versus try to build their own ebike fleet.
Mill is applying XaaS to an old kitchen device, the bin. For about $1 per day anyone in the United States can have a convenient, stink free way to cut waste and get uneaten food back into the food system. Like Cycle, this isn’t just about leasing versus buying. Mill takes back the processed food waste. But importantly at $1/day Mill already can compete with traditional food waste waste management services and so local governments will likely prefer to fund Mill’s service to more quickly drive down their waste management costs, while also hitting their GHG targets. Circuit has already proved out the usage based model for local governments with their electric transit services.
We spend a lot of time trying to figure out how rapidly we can help founders turn things into services. It’s certainly possible to fund a fleet of batteries, bikes or bins using traditional VC, but assuming you’d like to have fleet sizes into the millions and you might have an appliance that costs anywhere from $100 to $10000, we’re talking about $100m to $10b in equity fundraising which tends to break most cap tables however you try the math. So it’s essential to figure out a path to non-dilutive funding.
A framework for hardware + software climate startups
We know that most climate startups will combine hardware and software. This requires relearning some of the things that pre-internet VC was good at, especially working alongside other capital providers.
In the next 9 parts we’ll discuss why many VC investors fear hardware and then approaches we’ve seen to reduce hardware risks. We’ll revisit learning curves which are central to how all costs come down from bill of materials to finance. Then we’ll look at customer and partner signals as well as an idea we’ve come to call “Design for Finance” or how design decisions can help unlock and reduce the cost of funding. Finally we’ll get to our list of open questions we hope to get some help with.
