On the chemistry side: In step 1, you say using adsorbents, so i'm guessing some combo of iron oxide to desiccant to activated carbon for the siloxanes, and then maybe ZnO (based on your likely catalyst chemistry). At typical anaerobic biogas H2S concentrations, that's quite a bit of OPEX for media I imagine, and the risk of some breakthrough poisoning your reaction catalyst. Are you only working with partners that have desulfurization in place already? And how much attention do these skids need day to day, are you expecting full-time operator presence? Remote monitoring? How are you catching breakthrough before it takes out a catalyst charge?
I'm guessing bi-reforming is partly how you tune your way out of the carbon deposition problem, but in my experience real biogas composition drifts around depending on what's going into the digester, so i'm curious how much margin you actually have on the H2O/CO2/CH4 ratio before you're back in the coking regime. Are you trimming steam in real-time based on gas composition or running fixed excess (further trading economics)? Was the bench-scale test run on a simulated dynamic biogas stream? Was there much activity decline over the run?
On the economics side: DME into cosmetics seems like a great high-margin entry point. As you point out, most methanol is produced from large centralized plants, but they have real fixed-cost advantage that a modular approach structurally doesn't (along with storage and distribution headaches from many smaller production sites). Are you assuming some customers will pay a premium for the resilience of a distributed network? How much of the methanol case is cost reduction at your expected scale versus 45Z-type credits? And you probably can't share, but I'm curious the most you can pay a producer for their biogas and still pencil relative to RNG buyers who may be bidding for the same stream?