Scaling CO2 Extraction: What Comes After the Pilot Batch

September 28, 2026

Photo by Policarpo Brito via Pexels.

A one-kilogram closed-loop CO2 run and a 42,000-kilogram annual harvest are not the same business, even though they share a solvent, a pressure vessel, and a chiller. The gap between them is where most extraction startups quietly die, and it's where the equipment makers who supply this industry either grow up or get stuck selling pilot rigs to hobbyists forever. Christina Lake Cannabis (CLC) is a useful anchor here because the company put real numbers on that gap: it ran roughly 100 kg/day on mono-solvent CO2 extraction, then scaled to 300 kg/day once it added a cosolvent system, and used that added throughput to clear its entire 2022 harvest of about 42,000 kg into distillate by the end of March 2023.

That's not a story about better chemistry. CO2 as a solvent behaves the same way at one kilogram as it does at three hundred; the physics of supercritical and subcritical extraction don't change with scale. What changes is everything around the chemistry -- vessel certification, automation, clog resistance over multi-day unattended runs, solvent recovery load, and the sheer manufacturing capacity to build and certify pressure equipment fast enough to match how quickly cultivators are consolidating biomass into fewer, bigger operations. This piece takes a mid-term, three-to-seven-year view of that problem, using Vitalis Extraction Technology's own engineering and business choices as a case study in what it actually takes to build extraction hardware for an industrial-scale cannabis economy rather than a pilot-scale one.

Why Pilot-Scale Systems Hit a Wall

Photo by Yuri Shkoda via Pexels.

Root Sciences, an extraction consultancy and equipment integrator that's watched plenty of operators try to make this jump, frames the pilot-to-production wall bluntly: lab-scale closed-loop systems simply weren't designed with the throughput, automation, or safety architecture needed once a company's ambitions move from 1 kg batches to 100 kg batches and beyond. A pilot system's job is to prove a process works -- to validate that a strain, a pressure curve, and a temperature profile produce the cannabinoid and terpene yield a formulator wants. It is not built to run unattended for a 20-plus hour cycle, night after night, at industrial volume, with the same operator checking on it once a shift instead of standing over it constantly.

Vitalis's flagship PrimeEx R-200-H is a useful benchmark for what crossing that wall actually looks like in hardware terms. It runs dual extraction chambers and dual separation lines, so the system isn't waiting on a single vessel to depressurize and reload between runs. It's engineered for clog-free operation, which sounds like a minor spec until you've watched a subcritical CO2 line seize up on wax at 2am with no one there to fix it. It carries two 140-liter liquid CO2 vessels and a CO2 recovery and recycling loop, and it can push flow rates up to 8 kg/min at 2,000 psi. Every one of those numbers exists because someone hit a wall at a smaller scale first.

The important structural point is that this jump isn't linear. Doubling throughput on paper doesn't mean buying a proportionally bigger vessel and calling it done. Separation dynamics shift when more plant matter and CO2 are moving through a line simultaneously -- wax and lipid fallout behaves differently at higher flow rates, clogging risk compounds rather than scaling evenly, and the CO2 recovery system has to work harder to keep solvent loss and cycle time in check. That's the real entry point for everything else in this article: what has to change structurally, not just numerically, for a piece of equipment to go from proving a process to running a production line.

The Cosolvent Lever: Getting More Out of the Same Footprint

If the pressure vessel is the engine, the cosolvent injection system (CIS) is closer to a turbocharger -- it doesn't change what the engine is, but it changes what it can move. Vitalis reports that its CIS delivers up to 233% greater daily throughput than CO2 alone, with meaningfully faster run times per batch. That figure lines up with what CLC experienced in practice: roughly 100 kg/day on mono-solvent CO2, tripling to about 300 kg/day once a cosolvent was added to the process.

That jump is what let CLC turn a 42,000 kg harvest into finished distillate by the end of March 2023 -- a concrete downstream timeline that shows what scaled equipment actually enables, not just for the extraction step but for everything after it: refinement, formulation, and getting product to market before a harvest ages out of its shelf-stable window. A cultivator sitting on tens of thousands of kilos of biomass has a spoilage clock running; equipment that clears that inventory in months rather than a year is worth real money independent of any per-gram yield improvement.

The mechanism is straightforward chemistry, even if the engineering around it isn't: cosolvents like ethanol improve the solubility of target cannabinoids and terpenes in the CO2 stream, which shortens the time needed per batch to pull the compounds you actually want. Critically, that gain doesn't require a proportionally larger CO2 vessel -- you're getting more separation per unit of CO2 moved, not just moving more CO2. That's the appeal: more output from the same physical footprint.

None of this is free, though, and it's worth being clear-eyed about the trade. Cosolvent systems reintroduce a layer of solvent-recovery infrastructure and residual-solvent testing that pure CO2 systems were specifically designed to avoid -- CO2's appeal to a lot of operators in the first place was that it left no ethanol residue to test for and eliminated the flammability handling protocols that come with liquid ethanol on-site. Adding a cosolvent loop back in means added capital cost, added regulatory testing burden, and a return of some fire-code and storage considerations that a mono-solvent CO2 shop gets to skip. The throughput gain is real, documented in CLC's own operating numbers, but it comes with an added compliance and safety layer that has to be engineered and staffed for, not just bolted on.

Certification Is the Real Bottleneck, Not Chemistry

Photo by PublicDomainArchive via Pixabay.

Here's the part that trips up equipment makers who think in terms of flow rates and PSI ratings alone: scaling past pilot batches means satisfying certification regimes that don't talk to each other. A production line intended to sell into multiple markets has to clear euGMP compliance for pharmaceutical-grade output, ASME code for pressure vessel design in the US, National Board (NB) registration, CSA or CRN approval for Canadian jurisdictions, and CE marking under the EU's Pressure Equipment Directive (PED) for European sale -- often simultaneously, on the same vessel design. Pilot-scale shops building one-off rigs for a single customer in a single jurisdiction never have to touch most of this. A vessel rated up to 4,800 psi, the kind of rating industrial continuous-flow systems are pushing toward, sits at the intersection of all of them.

Vitalis's response to that bottleneck has been to stop treating it as someone else's problem. Co-founder and CEO Joel Sherlock has described the company getting greenlit to manufacture its own certified pressure vessels in-house, rather than sourcing fabrication from third-party ASME-certified shops and then integrating those vessels into a finished system. That's a meaningful shift in what kind of company Vitalis is -- less an assembler of purchased components, more a vertically integrated manufacturer that owns the regulatory relationship from raw steel to finished, certified unit.

The reasoned extrapolation here is that this certification burden is becoming a moat rather than just a cost center. As more markets -- the EU, Australia, and various Latin American jurisdictions -- adopt pharmaceutical-grade GMP requirements for cannabis extracts destined for medical use, the equipment makers who already hold multi-jurisdiction certification get first crack at those supply contracts, because a cultivator or processor entering a new regulated market doesn't want to be the one absorbing months of vessel recertification delay. Being pre-certified across ASME, NB, CSA, and CE/PED simultaneously is a real head start.

The grain of salt worth keeping here: certification regimes can also converge rather than multiply. Pharmaceutical manufacturing went through exactly this with the International Council for Harmonisation (ICH) guidelines, which reduced the need for drug makers to run separate approval processes in every major market. If cannabis-extraction equipment certification followed a similar path toward mutual recognition agreements between regulators over the next several years, the compliance moat Vitalis and similar companies are building today could narrow considerably -- a real risk to weigh against the current advantage.

Vertical Integration as a Business Strategy

Photo by Daniel Dan via Pexels.

Vitalis's move into in-house certified pressure vessel manufacturing is, at bottom, a bet about where the actual constraint sits. If the rate-limiting step in scaling is certified fabrication capacity rather than raw engineering talent, then owning that fabrication reduces lead times and insulates the company from a third-party ASME shop's backlog becoming its own backlog. That's a supply-chain-risk argument as much as a manufacturing one.

It's also not a novel playbook -- it echoes how other regulation-heavy manufacturing sectors matured. Semiconductor equipment makers vertically integrated critical subsystems once lithography tolerances and fab certification requirements became harder to clear than the underlying engineering problem. Pharmaceutical bioreactor manufacturers did something similar as GMP validation requirements tightened: owning more of the production chain in-house meant fewer handoffs where a regulatory gap could stall a shipment. In both cases, the companies that controlled certified fabrication end-to-end outcompeted those still assembling from purchased, separately-certified components.

Vitalis's current footprint gives some sense of what building that kind of position looks like in practice: reported figures put the company's installations at over 160 units across five continents, with a client list that includes Fume, Heritage Cannabis, MediPharm Labs, Curaleaf, Redecan, and Eagle Hemp. Vitalis equipment was reportedly part of the production fleet at Redecan when that company was sold for CA$925 million -- a data point that at least suggests the equipment underpinning a large, successfully-exited Canadian producer came from a maker pursuing this vertically-integrated, multi-jurisdiction-certified strategy. It's worth flagging plainly that these installation counts and client rosters come from profile pieces published between 2023 and 2025, so current figures may well be higher or the roster may have shifted since.

It's also worth resisting the idea that continuous-flow, vertically-integrated manufacturing is the only viable path to scale. Isolate Extraction Systems (IES) has been building closed-loop subcritical and supercritical CO2 systems since 2012 and represents a genuinely different architecture -- batch-based closed-loop rather than continuous-flow -- that has its own customer base and its own scaling logic. Scaling extraction capacity isn't a single-path engineering problem with one correct answer; it's a design-philosophy choice with trade-offs on both sides, and the market has room for more than one winning approach.

What This Opens Up for Equipment Makers Over the Next 3-7 Years

The clearest projection to draw from CLC's trajectory is that as more cultivators reach that kind of scale -- tens of thousands of kilos harvested annually rather than a few thousand -- the equipment business shifts with them. Selling a single pilot unit to a startup extractor is a one-time transaction; selling a full production line, with dual chambers, cosolvent injection, CO2 recovery, and a service contract to keep 20-hour unattended cycles running reliably, is a recurring, higher-margin relationship. That's a materially better business to be in, assuming the demand is actually there.

And that assumption needs to be stated plainly rather than taken for granted: this entire projection depends on continued consolidation of cultivation into fewer, larger operators. The Redecan sale at CA$925 million is one visible data point in that consolidation trend, and it's a pattern that's shown up repeatedly in maturing commodity-agriculture sectors -- fewer, larger growers who can justify capital-intensive processing infrastructure that a small operation never could. If cultivation instead stays fragmented among many small and mid-sized growers, the addressable market for continuous-flow production lines shrinks correspondingly, since only large-scale biomass volume actually justifies the capital outlay.

Assuming that consolidation trend holds, service and retrofit revenue looks like the bigger long-term opportunity relative to new-unit sales. CO2 recovery and recycling upgrades, and cosolvent retrofits bolted onto existing mono-solvent installations, let a customer who already owns a system scale their throughput without buying an entirely new production line -- which is a lower-friction sale for the equipment maker and a lower-capital path for the customer. That's the CLC story again, in miniature: the win wasn't a whole new plant, it was an add-on system that tripled output.

Two risk factors deserve honest airtime rather than a footnote. First, if US federal legalization arrives within this three-to-seven-year window, it could flood the market with cheaper equipment manufactured domestically in the US or imported more freely, potentially eroding the lead that Canadian and British Columbia-based makers like Vitalis currently hold -- a shift in trade dynamics that's happened before in other regulated-agriculture sectors once federal barriers dropped. Second, and more favorably for companies already holding euGMP compliance, the international angle: as Germany and potentially other EU member states expand domestic cultivation and import programs for medical cannabis, euGMP-certified equipment makers are positioned to supply that market directly, turning a compliance investment made for one reason into a competitive entry point for another.

Strip away the marketing language around flow rates and PSI ratings, and the bottleneck for continuous-flow CO2 extraction isn't extraction science at all -- CO2's behavior as a solvent is well understood and hasn't changed. The bottleneck is whether equipment makers can manufacture and certify pressure vessels fast enough to keep pace with how quickly cultivation is consolidating into fewer, larger operators generating CLC-scale harvests. That's a manufacturing and regulatory-capacity problem, not a chemistry problem, and it's why Vitalis's push into in-house certified vessel fabrication matters more than any single spec sheet.

Vertical integration and multi-jurisdiction certification -- euGMP, ASME, NB, CSA, CE/PED, held simultaneously -- are turning into the real competitive moat in this equipment category, ahead of raw throughput numbers. That moat could narrow if certification regimes harmonize the way pharmaceutical approvals did under ICH guidelines, so it's not a permanent advantage by default; it has to be actively maintained and extended market by market.

Over the next three to seven years, the two indicators worth actually watching are cultivator consolidation -- more Redecan-scale exits and fewer small independent grows -- and international GMP harmonization, particularly how Germany's and the EU's medical cannabis import and cultivation programs evolve. If both trends continue on their current trajectory, this becomes a genuine production-line business with service contracts and recurring revenue. If either stalls, continuous-flow CO2 extraction equipment stays a niche of pilot-scale sales to a market that never quite grew into the industrial scale its early boosters predicted.