Good morning. Earlier this week I wrote a piece in The Diff that compares the supply chains of bits and atoms. It was mostly exploratory—lots of loose ends, but if you’re familiar with the tech industry you might get a kick out of it.

I mention it here because it sparked an idea that I wanted to share, despite it not being full fledged out: regardless of your stance on the inherent inefficiency of direct air capture (DAC), it’s fun to imagine a world where energy is free, and the chemical engineer’s thermodynamic argument falls short. In this approximately zero opex environment (both raw materials and energy are free), the only thing that matters is capital costs, which means the most relevant inputs to success are 1) how easy it is to finance fixed assets, and 2) how much we can minimize shared spending on fixed assets (i.e. infrastructure).

What I’m suggesting here isn’t that we lean into DAC and burn all of our capital on what is a mostly incompatible with our current system—it’s that we recognize that it’s okay to invest in DAC a little bit, because there is this imaginary future world where it does make sense (I think). In short: maybe I should get off my thermo high horse and think about what could be true in 50 years for just a second?

If we get better at catalysis, in the way AI seems likely to accelerate, then the electrochemical reduction of CO2 looks a lot like the holy grail. We’re talking about literal alchemy; pulling valuable matter out of thin air. That being said, there probably are concepts that I’m probably overlooking here. Reaction rates and selectivity are generally good in electrochemical systems, but maybe I’m underestimating how complex such systems are to build, and maybe I’m overestimating how quickly AI will improve cathode/catalyst development. Maybe we simply can’t move enough mass through these DAC units to make it useful—I don’t know.

Unfortunately, this is the only decentralized future for chemicals and materials that I’ve been able to conceive. And it’s reliant on 1) free energy, 2) using the air as the free infrastructure for transporting carbon, and 3) a long tail of technical assumptions… all of which feel quite far out.

In any case, a decentralized future reliant on bio-based feedstocks will always be bottlenecked by a critical requirement: we need to grow crops and process them in a centralized area, so we’ll centralize downstream processing too. But if you’re transporting carbon through the air and using DAC you’ve basically got an internet equivalent for materials. And that’s a future I’m interested in.

What do you think? (Also, I plan on writing more like this and less about news, so feedback on that is encouraged.)

In the news:

• LyondellBasell is going to shutdown one of its Spherizone polypropylene (PP) plant in Italy due to poor market conditions.

• DuPont just bought back some 46.8 million of its outstanding shares for $2B, which was roughly 10% of its outstanding shares. (I think my napkin math is right, but someone should check.)

• Mitsubishi Heavy Industries is handling the front end engineering design of a new fluorspar-based hydrogen fluoride plant headed to Japan.

• BASF and TotalEnergies’ joint venture in Port Arthur had an incident: a fire in the plant’s pygas unit cause a tower to collapse, but nobody seems to have bene injured.

• Ascend Elements secured a $542 million Series D round to produce precursor cathode active materials and cathode active materials in the US.