The Hidden Weakness in Most Cannabis Operations — What We Heard at Benzinga

Listen Now

Kellen Finney: I ran an extra 200 gallons of solvent over it to make sure I got the last one gram, right? Well, did you factor in that taking that 200 gallons, turning it into a vapor using heat, and then recovering it — it doesn't seem like a lot, but when you're fighting for pennies now, A, it's time, B, there's electrical cost. Maybe if you're using liquid CO2, that's another cost to recondense it, right? And so that cost adds up. Yes, you're not losing the solvent, but you are having to recover it, right? These are just some of the simple process optimization techniques that Exxon and these large companies implement. Bryan Fields: What's up, guys? Welcome back to another episode of The Dime. I'm Bryan Fields, and with me as always is Kellen Finney. This week we've got a very special episode — just Kellen and I. Kellen, how are you doing? Kellen Finney: I'm doing good. I'm excited to just chat with you for the next hour. I think we have a lot of interesting things to talk about. How are you, bro? Bryan Fields: Thinking about it before we did this — it's kind of weird, we're going full circle. When we first started it was just you and me, then we went through a 250-episode stretch where we had guests, and now we're kind of going back to the roots, just the two of us. Kellen Finney: Yeah, it's interesting because at first, our first guest, I was so nervous for, right? And then after like a hundred you just get used to it, and now, again going back to the two of us, I'm actually a little more nervous about this episode than having a stranger on. Bryan Fields: No, I felt the same exact way. I showed up and wondered what we were going to talk about. Then I realized we came back from Benzinga, we got a ton of information, and at a certain point we've had so many private conversations with individuals and leaders of the industry that some of the information we heard made me wonder — should we share it, should we not, especially given what's going to transform in this industry. So high-level summary from Benzinga — did you enjoy it? What was the number one takeaway you had? Kellen Finney: I enjoyed it. I think the high-level takeaway is that it continues to be the conference that has the right people in the right place. You know what I mean? Bryan Fields: Yeah. I don't think there's a single conference that can compete with it. I've been pretty loud about my opinions on expos and conferences — from a marketing standpoint I don't think a lot of them are worth anything, I think they're just lighting cash on fire. Every year we go to MJBiz and I just wonder what people are spending and if it's translating to sales or awareness or whatever they're trying to accomplish. Benzinga for me continues to be the number one show — the decision-makers are there, the ability to connect with individuals is there, the information shared is paramount, and you get a good perspective on people's headspace. Some years there's massive optimism, other years it's a little downtrodden, and this year I felt like it was a little bit in between. Did you feel that way too? So I'll give a quick example — one of the conversations we had was that there are a lot of warning signs going forward, that right now, June 17th, things are bad and they're likely about to get way worse. And unfortunately that's probably the reality. And I think you mentioned the shakeout, right? This is not just small companies, but big companies that are in challenging situations and are going to have to figure out how to flip the script pretty quick. Kellen Finney: Yeah, it's kind of like a calling-of-the-herd event that's going on. Also, I rode the elevator with two unique consultants — and we've been going to Benzinga for a while and I've never met a consultant that specializes in receiverships. I got two business cards from receivership consultants, which kind of tells you the tale of the tape of what's going on in the industry right now. Bryan Fields: Yeah, that's a horrible sign. One conversation that really sparked me was that there's a large-scale company that's super healthy, one of the premier ones everyone knows, and they're taking a kind of de-risking or diversification approach with the customers they work with. I found that strategy so interesting — instead of overleveraging their relationships with certain companies, they're tightening up and saying, let's do business with smaller companies, more of them, instead of leaning all in on one. Do you think that's a domino-effect fear — if two other companies go out of business and someone gets caught, they're all caught up — or more that right now nobody's sure what's going on so they're splitting their bets across the table? Kellen Finney: I think there's a lot of that going on, but I think a lot of people right now are trying to do the 80/20 rule — 80% of revenue from 20% of relationships — and they're leaning into their really strong relationships and figuring out which other relationships they need to get rid of. There are companies doing well that have good relationships, haven't over-expanded, and focused on their core competencies — there just aren't a lot of them. I'd say the majority of the industry has placed a lot of different bets that haven't panned out as intended, partly because of political headwinds. The hemp thing was a huge topic, if you want to get into that. Bryan Fields: Before we get into that, I want to bring up one of my favorite quotes you've said ad nauseam — vertical integration, done the right way, is a massive amount of margin, really beneficial, but done incorrectly or poorly, it's highly challenging. You're running six or seven different businesses stacked on top of each other, all needing to be interwoven and deeply understood. And it's more complicated because there are no real economies of scale — vertical integration is only critical in a given state and doesn't transfer to other states. So even if you have specialization, which is nearly impossible, you don't get rollover effects elsewhere. What we're starting to see is that stacking all those businesses on top of each other built the house really fast without good layers — nobody's exactly sure what they do best, and they've got challenges across the whole supply chain without a strong enough organization to handle it. Kellen Finney: And again, MSOs are also feeling the brunt of this. You may have 14 different facilities across 14 different states, but they're not rowing together, because state A and state B have completely different regulations. When they get together to communicate, one says, "we have to go this way because of XYZ," and the other says, "well, we don't have XYZ, so we go this way." So now you have even more isolated entities that aren't growing together — and that fragmentation, plus growth-at-all-costs, has really hurt the industry. Bryan Fields: So how is this going to shake out? At a certain point teams are going to say, "we don't make money in processing, we should offload it," or cultivation, whatever it is. Do you think it becomes a shakeout fest where teams look inward and specialize — figure out what they do better than everyone else and outsource the rest — or something else? Kellen Finney: I think specialization is probably the only way forward for a lot of these organizations. There are a lot of good teams that are good at certain aspects but spread too thin. The more sophisticated organizations will look at the data, do some self-reflection, analyze which pieces of the business are and aren't making money, and clean it up — but it'll be painful. Cutting jobs and departments isn't easy, but the good ones will do it and survive. Bryan Fields: Yeah, and those become the bad PR headlines — everyone runs with it like it's doomsday. But things have to get worse before they get better, and those are really hard decisions involving jobs, personnel, assets, expenses. I wonder where that ends up, because one thing we heard was that teams are now meeting weekly instead of monthly or quarterly to review how to tighten things up. I think that's a warning bell — an internal signal that if they don't figure things out soon, things will get worse and they don't know how much longer they can operate like this. Kellen Finney: It's a sign of a market maturing. Initially when a market opens, a license itself can be worth a million dollars. Then you watch the industry mature — now in Michigan you can't even give away a license, a piece of property with a license attached is still hard to offload. It shows how lucrative things are at first, and then how quickly things tighten up in these markets. Bryan Fields: Faster than people realize. Some of these trends started as just a California trend, and what was a California trend in January now, if you don't have it by June or by the end of the year, you probably don't have a chance in certain parts of the supply chain. Everyone's leveling up their game, figuring out where they're weakest and patching holes, because if they don't fix them now it could kill them when things get even more competitive in early 2026. So the one area we should quickly touch on is the situation in Texas — do you want to set that up? Kellen Finney: Yeah, so for those who don't know, a bill was passed in Texas that bans intoxicating hemp. It's now up to the governor to veto it or not — I think he has a couple days left. Bryan Fields: Five days left, June 22nd. Kellen Finney: Yeah, June 22nd. If he vetoes it, the hemp industry can continue making intoxicating cannabinoids — this includes a lot of the D9 beverages. Texas is responsible for a lot of the conversions and infusions for beverages across a much larger industry in the United States, because they don't have the same regulations as the cannabis industry — a whole other conversation. There's a lot of political headwind because the governor is between a rock and a hard place. Bryan Fields: And it has potential implications at a federal level too — we're hearing there are Trump representatives in Austin trying to influence the outcome in a major Republican state, being influenced by the potential leader of the free world. Kellen Finney: Right, this is where things get super wild. If things take a turn, what could potentially happen fast? Rescheduling. You think that's the move Trump could potentially take? Bryan Fields: I don't know if that makes the playing ground even, but I'd think it could be a curveball — they veto the bill on the 22nd of June, and Trump comes over the top a week or two later, hypothetically, and just reschedules. Then we'd have pure chaos for a month or two. Kellen Finney: I also read a random headline that they could close the loophole federally even if they veto it. There are a lot of potential different outcomes. This is the part of the industry that's so hard to keep track of — you take three days off following everything and come back with no idea what's going on. I don't know how someone running one of these businesses can forecast with all these variables in the pipeline and massive implications. Maybe there's a cleaner approach, but this is the kind of thing that would keep me up at night — are we prepared for this? Jesse did a great job talking about optimism, and maybe we're sometimes too optimistic, but there will be a day when this optimism becomes reality, and hopefully that's sooner rather than later — before the end of 2026. Bryan Fields: It'll be really interesting to see how it plays out. You have two very different scenario-planning outcomes for the structure of your business. At a macro level it's an economy-of-scale issue — if the hemp industry can keep manufacturing these derivative products, they can out-compete any cannabis company limited by state lines and canopy limits. You can't grow 100,000 acres in cannabis — what are we doing? Kellen Finney: And the distribution aspect too — same as the smoke-shop model. Bryan Fields: Like Coca-Cola. Kellen Finney: Turns out that's not the first play — we still can't order cannabis on the internet. That's the crazy thing. Bryan Fields: Edibles.com — shout out Thomas. Kellen Finney: Let's go. Send him a bill for that one. Bryan Fields: Looking forward to having him on the pod. That's a bill — he's getting an invoice. So one of the other areas of that Texas thing — we'll have some good episodes coming up on that one. Kellen Finney: Yeah, that's some good gossip. I didn't even fully understand what he was saying when we got told that — this is wild. Bryan Fields: We'll have some good episodes coming up at the end of the month trying to make sense of it, because I need a better understanding — it's way too confusing, and every day there's a new change, a new main character telling us one thing or the other. One other area that really caught my eye at Benzinga — over the last six to eight months we've seen people like Tyler working on inventory, really understanding forecasting needs and working backwards. And on the cultivation side, everyone's got sensors, understanding best practices and really dialing in their numbers. Finally, people are realizing there's one key part of the supply chain that's leaking pretty bad, and now it's time to get visibility into it. Did you hear those same things, and did it surprise you it's taken this long? Kellen Finney: I did hear those same things. I think another driving force is the continued increase in the total market share of derivative products — infused pre-rolls, vape pens, edibles, beverages — now making up about 60% of the total addressable market. Sophisticated groups are looking at the data, seeing these trends, and realizing that box with a bunch of explosive solvents is actually a really important part of their business now, responsible for a massive and growing product feed in an industry that's otherwise been clamped down from a growth perspective. Bryan Fields: It's such a critical part of the puzzle, and one of the widest disparities between teams. I think about it in three or four buckets: the current artisanal approach with a flashlight and handwritten notes, all the way to a fully automated extraction system with real-time sensors and decisions removed from the operator's hands, and everybody in between. It no longer makes sense that teams can look at themselves and say a manual approach that worked the last five or six years will keep working in the next six months, when competitors are using automated systems with real-time sensors. Kellen Finney: I couldn't agree more. There's a full spectrum of processors in the industry. Honestly, it's straightforward — just because you're good at science doesn't mean you have a business, and a lot of companies doing extraction now are good at business. They have good sales, they've managed expenses, they're still around, they've got licensing — they did a lot of the other aspects of running a business well. That's part of why you have this huge spectrum — a lot of companies do everything else well and treat extraction like a black box managed from 50,000 feet: things go in, things come out, and they pull other levers to try to make more money, like getting cheaper inputs. But from a cost-aggregation standpoint, using handwritten notes just isn't that helpful. These are growing pains — Exxon didn't start with a hyper-focus on every fraction of distillation and how it relates to commodity pricing either, but that technology exists now, so we don't need to reinvent it. Bryan Fields: My favorite question to ask is how long this technology has existed — it's been mandated by the FDA in pharmaceutical manufacturing since roughly 2001, so over 25 years. The idea that someone can look through a tiny sight glass with a flashlight to make a key decision blows my mind, and the fact that a solution exists to replace that with a sensor that manages components in real time is no longer a game-changer — it's a necessary standard to compete. Early this year it was really only California facing this pressure, with tight margins driving the need for real-time visibility. Even in February a lot of teams still felt their numbers and SOPs were fine, but Benzinga was eye-opening because teams are now aggressively pushing this forward, realizing their competitors are doing it and if they don't have a solution to compete, they don't have a chance at all. Kellen Finney: With margins like that, it's a race to the bottom — you've seen this on the West Coast as markets mature, and every penny counts. If you're wasting even 20 or 30 minutes in some of these manufacturing facilities, that can change your margin significantly. And it's even harder because the incoming material is a plant — you're likely not logging enough operational data parameters to track which key parameters influence yield, and everything gets mapped through a color-coded Google Sheet without the ability to actually read why one batch was successful and another wasn't. I think one reason for that is the track-and-trace program — a negative side effect with good intentions, meant to prevent diversion, and it provided a framework for managing data, but it wasn't delivered properly and became a massive chore. Companies are forced to record all this information and no one uses it for anything, because it's compliance data, not operational data. Bryan Fields: Is it the same critical data — temperature, pressure in the extraction system? Kellen Finney: No, it doesn't track those things. It tracks yield — material in, oil out, and waste. The metric goal is compliance, so people got distracted by compliance and now they're not using data for optimization — it's just a chore they're over at this point. Bryan Fields: Why don't you give a quick synopsis of the critical components in extraction, and how if you don't understand the thermodynamics involved you're pretty much guessing throughout the process? Kellen Finney: There are a lot of different extraction methods, each with their own critical parameters. Hydrocarbon is simple in the sense that there aren't a lot of levers — temperature and pressure are critical, related by the ideal gas law, PV equals nRT. Record pressure, temperature, and volume of solvent used and that dictates everything. With CO2, temperature and pressure are also hugely important — they dictate what phase the CO2 is in, like supercritical CO2 extraction, where at a certain temperature and pressure CO2 changes density and phase. With ethanol, most of the cannabis industry practices cryogenic ethanol extraction, because at colder temperatures you don't extract as many sugars and gums, so when you remove the solvent later you don't clog up your distillation equipment. A warm extraction requires dealing with gumming, which is a practice pulled from the palm oil industry — a lot of processes used in cannabis at industrial or pilot scale are directly borrowed from palm oil, like winterization, a term literally derived from winterizing palm oil. I think what is unique to cannabis is hydrocarbon extraction, especially from a terpene-capture standpoint — that's not really done at industrial scale anywhere else. But the core components — temperature, pressure, understanding system conditions — are still critical, even if the specific application is unique to cannabis. Bryan Fields: And that's the point of it, and I think now's the perfect time to introduce what we've been working on, because we've hinted at it for a few years. Podcasting has just been the front, and now the front is exposed, so we can come clean. For us, we've been working on what's called Newton Insights — the ability to monitor THC in real time and meet teams where they are. We've introduced a sensor that plumbs into your hydrocarbon system and lets you digitally track all those parameters, mirrored through our AI-generated system that lets you make real-time decisions and control costs. The beauty is the simplicity of using it — we recognized we can't force operators to change their whole workflow, so the best way to start is to integrate the sensor into the background and let teams keep doing what they normally do while the data logs digitally. Kellen Finney: And this ties into what we were talking about with compliance data — a lot of organizations are overwhelmed with data, record a wild amount of information, and just put it in a filing cabinet thinking maybe one day they'll hire a data scientist. It's a really tall task to go in, parse, and clean all that data and feed it into a traditional PowerBI-type system — it would take an expert team months. This is the benefit of LLMs and AI now — they can condense that work into seconds. Looking back over a decade of information in maybe ten Excel documents, one per month, even best case and hyper-organized, it's really challenging to pull correlations out and compare, say, an OG batch from June 2018 to June 2023. The task is too great for the benefit you'd get from it. Bryan Fields: Why don't you give a quick explanation of the benefits of the project we're working on and why it grows with you, since the more you use the tools and aggregate data on the platform, the more real-time visibility you get to make better decisions? Kellen Finney: The platform has our own LLM built in, fine-tuned on an insane amount of data I've curated personally over the last ten years being a curious scientist in the industry. That gives you the ability to instantly upload your data to the platform and query the LLM to parse it and compare it to what you're currently doing. This is really powerful for getting to the root decision-making information, because a lot of times people get bogged down just trying to figure out what the data means when really what's needed is the meaning that helps you make a decision. Bryan Fields: Given the speed everyone's operating at and everyone looking inward to cut costs, this lets you do that more effectively instead of just guessing — should we change this, should we change that, or my favorite, let's just go find cheaper biomass. That lever can only be pulled so many times. This is where the wide disparity between teams shows up — the artisanal approach worked for five or six years. One of my favorite quotes was from Rob, shout out Rob, talking about how before the car, people said their horse works great, why upgrade? But in a race, that's the competition now, and you don't have the ability to compete at that speed with a horse. As the industry moves at an insane pace, the adoption gap for technology is shrinking fast, and without tools and best practices, you don't have a chance. Kellen Finney: To piggyback on that — think about the assembly line, a perfect analogy, shout out to Ford, and the horse-and-car comparison. The assembly line went from a bunch of humans putting a car together, optimized over fifty years, to now having almost no human beings in car manufacturing assembly lines. If you're forward-looking as an organization, you realize automated manufacturing will strengthen your bottom line over the business's lifetime — you'll get better margins, and if it's a race to the bottom, margins are much better with automation. The first step in automation is getting the same information a human is gathering into a digital format — that's some of the lowest-hanging fruit right now. There have been attempts at fully automating hydrocarbon extraction, but one of the biggest obstacles is that they're time-and-recipe based while the plant material is very variable — different THCA content with the same solvent gives different results. If an automated system can't be perfect because of material variation, how is an individual with a flashlight supposed to be perfect with a sensor? The sensor lets the system adapt in real time instead of only trusting your best operator to drive it manually. Bryan Fields: That's the biggest takeaway — the sensor lets everyone perform like the best operator all the time. It raises the floor of the team and lets your newest employee deliver maybe 80% of your top employee's performance. It stabilizes the floor and reduces low yields — some teams get low yields for a variety of reasons and aren't sure exactly where the problem is, sometimes catching it before it hits the bottom line, sometimes catching it two weeks too late, which is way too late when everyone's getting tight on their numbers. You need real-time visibility and cost control, not answers two weeks into the future. Kellen Finney: Besides optimizing the chemistry of the interaction, there's optimization around traditional manufacturing and equipment use time — if equipment is on, it needs to be making something. A lot of simple optimization can come from just monitoring use cases and downtime. Bryan Fields: Why don't you explain how the downtime aspect works? I think it's one of the most expensive parts of the process — taking liquid, turning it into vapor, and recondensing it. Kellen Finney: Right, when we talk about waste, it's not the trash people picture — it's running an extra 200 gallons of solvent to make sure you got the last gram, without factoring in the time, electrical cost, and if you're using liquid CO2, the cost to recondense it. That cost adds up. You're not losing the solvent, but you're having to recover it, and these are the kind of simple process-optimization techniques that Exxon and other large companies implement at scale. I don't blame extraction teams — I understand how hard their job is, being pulled from both sides — but it's impossible for them to fully understand what it costs or how damaging it is to the bottom line without tools, because they're doing their job based on the KPIs set for them. Most teams aren't considering all these elements together in real time, and that's why 'waste' isn't even the best word — it's more like multiple leaks from a pipe that you're duct-taping left and right just to keep things moving. Bryan Fields: There could be a lot more digital simulation and running these metrics from a traditional engineering perspective. Kellen Finney: There are a ton of free calculators out there. Bryan Fields: And as we said, this started as a very California-focused tool, and we got a lot of early traction there, and now the East Coast has come on fast. It's not about how quickly we get one facility anymore — it's how quickly we get it into multiple facilities, because we realize how damaging this is but teams aren't sure where the problem is or how bad it is. One of the things that's surprised me is not just the hidden waste from electricity, but that someone can plumb this into their system today and within 15 minutes start getting information they never had access to before, giving real insight into what's happening instead of just hoping everything's working out. Kellen Finney: A lot of optimization is about reducing friction points. Right now a lot of organizations are capturing proper data and hitting their KPIs, but there's a friction point of busy work — writing something down on paper, uploading it into an Excel document, maybe losing a Saturday compiling a monthly report — time that could go toward improving the business or creating new products, when a platform can now do all of that instantaneously and give people that time back, streamlining the organization from an employee standpoint too. Bryan Fields: That's a great point. Some of what's damaging is that things move so fast that people stay in the same roles even as tools advance, especially with AI, and change is hard, especially at the speed the industry moves. But if you don't change and keep heading toward the wall, you're going to hit it. Once people see the benefits, they wonder how they'd ever go back to the flashlight, the handwritten notes, the guesswork that's killing margins while competitors in the same state are using best practices from chemical manufacturing. Kellen Finney: Every single pharmaceutical pill you've ever ingested has passed by a type of sensor that measured the API concentration in it. Bryan Fields: To recap — specialization is going to be a core theme, with a lot of generalist teams needing to figure out fast what they do better than everyone else. And doubling down on sensors and technology from an automation standpoint — I expect that to be the fundamental talking point for the second half of 2025, and if you don't have it by 2026, you don't have a chance to compete. The car-and-horse visualization sums it up: the old manual approach worked for a while, but once competitors show up with cars, how can you possibly compete? Any other takeaways for the back half of '25? Kellen Finney: No, I think that's a good note to end on. Bryan Fields: Cool. So if you're looking to learn more about that sensor, you can go to Newton Insights — that's n-e-w-t-o-n-insights.com — or reach out to Kellen and I. Thanks so much for listening, and if you like this format or want to see something different, let us know — we're out here freestyling and figuring out how to fill the open space. Kellen Finney: Yeah, and just remember to make good choices, everyone. Bryan Fields: Thanks, Kellen.