Big and Slow: Why Chemical Startups Are Rare

How high fixed costs and long payback periods encourage scale, inhibit value capture, and make it difficult to bring sustainable alternatives to market.

Companies love to talk about value creation, but never mention value capture. Maybe capture sounds a bit too mercenary for public-facing business speak, but the concepts of creation and capture are inseparable: companies aren’t just tasked with creating value, they’re tasked with capturing the value they create, and not everyone is dealt the same hand.

Consider semiconductors: the companies that fabricate these devices effectively enable all of the world’s modern electronics, and create multiple trillions of dollars worth of value by doing so, but capture mere billions.

It’s sort of what the game is all about: company managers are tasked with figuring out how they can collect as much upside as possible, and investors are tasked with figuring out which of those companies will end up collecting it. Then, once investors figure that out, they give the companies the capital they need to make it happen, and they do so according to how their risk appetite balances out with the rate and timing of expected returns.

Investment in chemicals/materials is generally associated with delayed returns, eroding margins, and a lot of CapEx (i.e. not exactly the type of opportunity that gets VCs excited). But alas, projects do get funded by other means (discussed later), and they typically play out like this:

1. A new chemical/material/process is invented by a company.

2. That company introduces the chemical/material/process to a pristine market, and tries to gain market share as quickly as possible.

3. Before you know it, patents are expiring and there’s a pressure to license the process, so competitors enter the market.

4. Competition between producers reduces everyone’s ability to price discriminate between buyers, so buyers get a market price and value capture shifts downstream.

5. Now, without any meaningful amount of pricing power, the chemical/material producer can only increase profits by reducing costs, so the price for the chemical/material slowly approaches the cost of production.

6. As time goes on and the cost of production drops, new applications are enabled, making it even harder to price discriminate.

This process, also known as commoditization, is inevitable in all industries, but to different extents. It happens to some extent when industries produce a similar product for many decades, to a greater extent when that product is identical for many decades, and to the greatest extent when that identical product has many different applications.

For example, even though we’ve been making semiconductors for decades, we introduce a slightly improved version every few years, where each new version enables some new applications. So semiconductors are commoditized in the sense that they can’t capture enough value to justify higher margins, but not in the purest sense, because each new version is differentiated. Now consider another product: we’ve been making polyester terephthalate (PET) for over half a century, and it’s the exact same thing that it was when it was introduced for use in soda bottles. In the early days competition in PET production led to some commoditization (stages 3-4), but it wasn’t until PET found a growing home in other applications, like clothing, that true commoditization occurred (stages 5-6).

But, like a slightly improved semiconductor, slightly improved versions of PET also exist (hello, PEF). So why hasn’t PEF become the world’s dominant polyester? One particularly problematic issue is that even if PEF achieved price parity it still might not enable more applications (unlike the slightly improved semiconductor), it would just replace existing applications, immediately dooming it to stage 5. And even if it did enable new applications, the future PEF producer would still be burdened by a tragic double whammy:

1. Building large production sites requires high fixed capital costs. So you need to find investors with deep pockets.

2. It can easily take 3-5 years to build one of these large production sites, and 5x longer if it’s a new chemical/material/process. So you need to find investors who also don’t mind waiting a decade or so to start seeing returns.

All else being equal, higher CapEx (1) implies longer payback periods, but for chemicals/materials there’s an additional delay (2) that makes the payback period even longer. But that’s not necessarily an issue. It’s not hard to find investors with deep pockets and patience—mutual funds, pension funds, endowment funds, and insurance companies all fit the bill—the issue is that those investors don’t take on technological risk.

Fortunately, processes that involve fluids are uniquely receptive to scaling laws, so scale quickly becomes a chemical engineer’s best friend. First we try to increase production rates without drastically increasing fixed capital costs, like converting from a batch process to a continuous one, and then we try to increase production rates by building bigger. Both of which reduce CapEx per ton of output and reduce the payback period in one fell swoop.

From there, the degree to which a company chooses to scale a single site is eventually limited by either a) the demand for what they plan to produce, or b) by some sort of physical or equipment-related constraint; whichever comes first.

But either way, the fate of a new chemical/material/process will ultimately be decided by what and how many new applications it enables, and the amount of value it can capture by enabling those new applications. Applications are enabled when a chemical/material meets a set of specific constraints, such as a desired strength to weight ratio, optical transparency, or gas barrier. If those constraints can be met by many different chemicals/materials, then your new chemical/material is less of an enabler and more of a replacer. This is important: you’re only going to have pricing power if you’re the only one capable of enabling a new application, and pricing power is what lets a company capture value.

So if you have a small end market, you’ll build small. Sure, you won’t reap all of those fun scaling benefits, but you’ll always be able to charge high margins because you’ll always have pricing power and the resulting ability to capture value (it’s a lot easier to price discriminate when you only have a few customers). Conversely, if you have a huge end market, you’ll build sites as big as you can. You may or may not have pricing power in the beginning, but either way, competition is coming, and it’s coming soon, so the objective effectively becomes “maintain what pricing power you have long enough to establish your position in the market”.

If you’re familiar with the chemicals/materials industry, you’ll know how rare it is to be the only one capable of enabling an application, and that it’s virtually non-existent at scale. And while application producers are still coming up with new applications (s/o to Downy Unstopables), their bottleneck isn’t some sort of chemical/material limitation. Their task is more about choosing between multiple viable options, and less about wishing for a magic chemical/material to arrive on the scene. The bottom line is that the materials we have today can satisfy pretty much all imaginable applications, so it’s going to be hard for new chemical/material producers to enjoy pricing power at scale for decades.

The good news is that you don’t need maintain pricing power for decades to establish a market position.

I think we have several options:

1. We can influence consumers or impose ESG upon application producers to make the application constraints more demanding. If we added sustainability or recyclability to an application’s constraint matrix, then the solution would be some smaller set of eigen-materials (h/t John Bissell).

2. We could work to reduce CapEx. For a CapEx reduction to be meaningful you’d need to drastically reduce the amount of required equipment (perhaps by not making by-products), or you’d need to drastically improve your production rate (which generally doesn’t bode well for bio-processes).

3. The value chain itself could be shortened—if we went from raw material to the chemical/material used in an application with fewer intermediates there would be more room to capture value in the middle.

4. If we removed some regulatory hurdles, got better at simulating processes (maybe with quantum computing), or figured out how to build demonstration scale sites faster (maybe with modular construction) we might be able to meaningfully reduce the payback period.

5. It’s hard for me to imagine a world of decentralized chemical production because centralizing always leads to better process economics, and because most sustainable feedstocks still need to be transported prior to processing. But government incentives and financial engineering can compete with physics. And there’s also a utopian future to consider where water and energy are free, and suddenly direct air capture (DAC) looks less like a hyped climate solution and more like The Alchemist’s air separation unit (ASU).

I’ve written a lot of this as if the solution to our sustainability woes is new chemicals/materials. This is somewhat true, but new chemicals/materials means that everyone downstream will have to incur switching costs, so finding sustainable pathways to the same chemicals/materials would generally be preferred. Unfortunately, finding sustainable pathways to the same chemicals/materials dooms us to the realm of application replacement, which leaves us with no pricing power.

Unless, of course, application producers are incentivized to adopt chemicals/materials made via sustainable pathways, either because of ESG, or because consumers will choose the sustainable option at price parity—both of which effectively change the constraints. And it seems like that’s where we’re headed.