DexCom, Inc. (DXCM) Earnings Call Transcript & Summary

All right. Good afternoon, everyone. Can you hear me? Yes? Thank you for coming. We are pleased to have Steve Pacelli and Sean Christensen with us from DexCom. We are going to do a fireside chat, and then there is the breakout in Salon K after next door. Do you guys want to make any prepared remarks? Should I dive right in?

No. I think we just dive right in. I think most of the audience is familiar with the company. So let's just do it.

Awesome. All right. I'm going to try and cover a lot of topics that we won't get to, I'm sure. But we'll start off with 2020 and some of your expectations. So you talked about on the call, 2020 guidance, assuming new patient adds are about flat year-over-year. You're capacity constrained over the course of 2019, penetration in all of our core markets, both the U.S. and internationally, remain very low. More DTC later this year. How do you think about -- how do you get to that new patient addition assumption? Is there the chance that you add even more patients than that? And if that was going to occur, are there any areas where you see the most potential for upside?

Yes. So I'll take that and then I'll let Sean chime in, as needed. So we were thoughtful in our guidance. Our guidance implies 15% and 20% or, I think, 17% to 21% something like that, taking into account a number of factors, right? There's continued pricing pressure being the obvious key headwind, if you will, in the guidance going forward. And we've experienced that pricing pressure, really, self-imposed pricing pressure, is the way we describe it, over the past couple of years. And I say self-imposed, it's largely as we've made the affirmative decision to move our business for a number of reasons from what we were historically in the U.S., durable medical equipment, into the pharmacy channel. And the real benefit there is, we believe it opens up additional access for the patients to a much easier path to obtaining the product and to using the product on a recurring basis, but also opens up from a doctor's perspective today. In the DME world, the doctor is burdened with a significant amount of paperwork and other factors that go into just getting the patient on that first product. And if we alleviate that burden, doctors tell us, they tell our sales force all the time, "If I didn't have to go through those hoops to get the patient on the product, I would write lots more prescriptions." And so...

You needed to show a [ survey ] on that.

And we have. And we know it to be the case, right? If you make it easier for the patient to get and you make it easier, frankly, for the payer to open up the access by giving a little on price, we know we can drive volumes. And we saw that last year as we really started to push into the pharmacy last year, and we continue to do so this year.

Yes. I think he's framed exactly correct. I mean, we're incredibly bullish about the business. As we think about 2020 and we start the year with guiding to about $250 million to $300 million of absolute revenue growth. And that's net of what we've identified, some of this pricing channel mix initiatives where we're netting out about $150 million there. So if you factor those together, you're getting to about $400 million to $450 million of, kind of, gross revenue growth before the pricing impact. So we think this is a -- which is about what we did in 2019. So we think this is about the right place to start as we get to 2020. Obviously, as we go throughout the year and we try to drive incremental business, we'll be excited to update you as it progresses.

Okay. Going back to late 2018, you provided some Analyst Day targets that were focused entirely on the intensively managed or the core intensive patient populations looking out to 2023. Things moved a lot faster since then, I would say, especially on the intensively managed type 2 side. You started talking about starting to get some revenue contributions from the non-intensive business in the 2021 to 2023 time frame. But excluding that, how much longer do you think your core intensive markets can support a 20%-plus growth rate? And when do you start to need those non-intensive contributions to sustain that type of growth?

Yes. So just to frame it up for everybody. Back in late '18, we did our first ever Analyst Investor Day. And then what we framed up was over a 5-year time horizon, that we expected to achieve somewhere between $2 billion to $2.5 billion in, really, in our core markets. So that contribution would come from markets where we were already competing, where we already had approved products, maybe still working on some reimbursement, but in markets where we're really focused on type 1s and intensively managed type 2 patients. So we're obviously, over the past 18 months or so, we've kind of jumped ahead on that. We're going to probably wait until the end of this year to really reframe up a, kind of, 3- to 5-year look going forward. But we're still -- even in spite of that, even in spite of the phenomenal patient growth we've had over the past year, 1.5 years, we're still really under-penetrated. I mean from a type 1 perspective in the U.S., we think, we're, as a category, somewhere between 35% to 40% penetrated; intensive type 2 -- so type 2s, who take multiple daily injections or wear an insulin pump, we're probably still less than 15% penetrated in the U.S. market. So that would encompass about, we think, about 3 million patients in total. What we framed at our Investor Day was an addressable market in that 5-year time horizon of about 7 million to 8 million patients in those markets that we currently address. And so outside of the U.S., we're probably -- in a few markets in Europe, we're maybe 20% penetrated. Outside of that, probably much less than 20%, and on the intensive type 2 space, even less so. So looking out, even 5 years from today, we have a significant runway just in that core market opportunity in intensively managed patients. That being said, what we framed up at a conference earlier this year, for the first time, we actually presented some meaningful data, not kind of RCT-type data, but really real-world pilot -- real-world utilization data in type 2s using our products and type 2s who don't take insulin. And it was remarkable. I mean, just to give you an example, one of the studies we saw patients -- these are non-insulin using patients who wore our product, half of the population wore our product continually for 6 months; the other half was on, kind of, standard of care using ordinary fingersticks. And at the end of that 6-month period, we're working with a payer, Intermountain Healthcare. They estimated that the cost savings for the group on CGM was roughly $5,000 per year per patient. And this is not extrapolated for future A1c reduction and costs associated with long-term outcomes. These were hard costs associated with reduced doctor visits, reduced hospitalizations, reduced labs, just across the board. And so we're looking to expand that trial and, frankly, other pilot work we're doing with UnitedHealthcare, with Onduo, who is a close partner of ours, and some others, to really start to build real-world evidence of the utility of this product and CGM across a much broader spectrum. So what I don't want to do is give you an exact time frame when that contribution will happen. We don't have anything meaningful baked into our guidance this year in terms of contribution for non-intensive insulin-using patients. But going out perhaps next year in the out years, next 3 to 5 years, you're certainly going to start to see some meaningful contribution from markets outside of our core intensive business.

And in the meantime, you should still be generating pretty healthy growth over that time frame just in the core market.

We would expect to. Yes.

I want to come back to the non-intensive opportunity in a second. But let's talk a little bit about intensively managed type 2s. So you talked about penetration going up to about 15% or so in 2019. We think that was up from 5% or 6% in 2018 and only 2% in 2017. So that's really inflected over the past 2 years, driven by expanded access via Medicare and product innovation, particularly factory calibration, maybe the pharmacy channel as well. How do you think about penetration in that market over the next couple of years? And how important are catalysts are expanding commercial coverage as well as the new products you're launching with G7 and now G6 in Medicare?

Yes. So I think you almost answered part of the question in your question. But I think it's -- the expanded access has been key, right? I mean the growth that you've seen over the past couple of years was largely driven first, by Medicare adopting both type 1 and intensive type 2 coverage. The German government, out of the gate, we're starting to see a number of private payers in April of this year, United Healthcare will include intensively managed type 2s as part of the coverage universe. So what we're seeing is really that coverage universe expand, and that will enable us, particularly now that we're not capacity constrained, to really start to open our direct-to-consumer marketing efforts in that we really haven't marketed to those people. And in fact, we just launched G6, again, due to our capacity constraints last year, we just launched G6 into the Medicare population in Q4 of last year. But we did so without really any meaningful marketing or any meaningful promotion to those patients. So now that we're out of that capacity constraint, we're in a position to really start to dial up in the first half of this year, really start to dial up DTC, kind of across our channels, but really focused not only on the type 1, but on all of these intensive type 2s. And as we see coverage evolve across the commercial payer landscape, we continue to dial up that DTC and really target those patients.

So I think most people when we think about type 1 penetration over time and, "standard of care," whatever you mean that to be 70% to 80%, I would argue, even 90% plus over the very long term. I think most investors we talk to still think about the intensively managed type 2 patient population as inherently having a lower ceiling penetration than whatever type 1 is. Do you think, a, that the trends we've seen over the past couple of years are going to continue? Is penetration going to continue increasing rapidly? And then over the next couple of years, do you think that investor mindset will change as to, essentially, do you think it will become apparent that the peak penetration for this patient population is also very high over the long term?

Yes. Honestly, internally, we really look at our business now as intensive versus non-intensive. We don't actually bifurcate the business in terms of type 1, intensive type 2 and then non-insulin using patients. So we really are focused on that intensive piece versus the non-insulin, non-intensive piece. I see no reason why, again, predicated on coverage, if there's coverage for these people, there's no reason why adoption in the intensive insulin type 2 space should not be the same or even as good as -- certainly, as type 1, and ultimately, evolve to standard of care in that patient population as well.

Okay. And then how do you think about international for this patient population? So you talked about Germany with coverage. I'm not sure if there's really much coverage outside of Germany. Are you -- do you think there's any chance for that to change over the next 1, 2, 3 years? Is G7 a catalyst for that change or other lower cost avenues for those patients that you could talk about? How do you think about that long term?

I don't know that G7, in and of itself, has really an inflection there. But certainly, we're pushing, certainly, as CGM becomes more of a standard for type 1s, where we find benefits. So the German example is a great example, right? Ultimately, the German government and, frankly, why we believe Medicare out of the gate, has adopted both type 1 and intensive type 2, is that -- from a type 2 perspective, they really -- and frankly, they shouldn't even both be called diabetes, I say that all the time. But type 1 is an autoimmune disorder, type 2 is more of a metabolic, it's a progressive disease. But by the time those folks have progressed to the point where they're on mealtime insulin in addition to a long-acting insulin, they're probably in pretty bad shape. And they're probably going to -- if they haven't already developed the comorbidities associated with diabetes, they're probably pretty close to, and they're going to. So Medicare, they're kind of stuck with them for the rest of their lives. German government's basically stuck with them for the balance. So if you look at these -- kind of, these single-payer systems outside of the U.S., my belief is it's more likely that those systems would be willing to adopt a broader coverage universe because they're much more incentivized than maybe a private commercial payer in the U.S. who may or may not have this person on their plan 2 or 3 years from now, right? Once they're on, either with the government-sponsored plan, they're on it for good, whether it's Medicare or outside the U.S. So I think we will see -- the coverage outside the U.S. has historically been slow. So it's going to be slower, but it will evolve.

That's interesting. They care about long-term cost and outcomes. Maybe shifting to the product pipeline, unless anyone has any intensively managed type 2 questions? All right. Looking at G7, I think, if you look at our model and consensus model, I think most are assuming that market share for DexCom remains relatively stable in -- over the next, really, 3 to 5 years, somewhere in the 40%-plus range. Do you think that we're underappreciating the potential for G7 to even improve upon DexCom's already strong competitive position today and drive more share gains?

Yes. That's a great question. I mean, again, not fully appreciating what might be in the pipeline from 1 or more of our competitors. G7 is a category killer, right? It has all of the attributes and the performance of G6. So no fingersticks. Hopefully, at launch or at least some period very shortly after launch, it'll have a wear period longer than G6. It's a form factor that will be about 30% smaller in total than the Libre. Again, no calibration. I can't underestimate how important no calibration has been to our market adoption over the past few years. So I don't see why G7 isn't, from a share perspective. But again, it's important to remember that this is not a zero-sum game today that we are still -- we talked about penetration at the outset of the presentation. This is not a zero-sum game. There's plenty of opportunity for multiple competitors, whether it be Abbott or if -- Medtronic, if they improve their sensor quality or even if there's another market entrant that we might not be aware of today. There's so much room for growth just in the intensive space in those core markets that, for the foreseeable future, this isn't about shifting share from us versus Abbott, for example. It's really about growing the category.

And do you think, from a hardware perspective, upon the launch of G7, that the product is really reaching a level of maturity? And if so, how should we think about what comes next for DexCom's innovation? How much focus on software will there be? And when might we get updates on that as investors?

Yes, we'll update you guys at the various -- either from the stage at a diabetes conference, for example, or during a quarterly call. What I would point to as I would point to the iCGM standard, and you've seen as Abbott has really fought to get that iCGM designation and others will try, it's really hard. The standard that the FDA established with our G6 approval, even internally, I mean, we're really taking our time to make sure, as we kick off the G7 pivotal, that we really have our ducks in a row. Because you've got to hit -- you can't have even 2 or 3 or 4 outlier patients in a multi-hundred patient study, or else, you run the risk of tripping the iCGM standard. So what we do is really hard. So when you think about performance, I don't -- if I tell you G7 is going to be an iCGM sensor, that should tell you it's going to be good enough, right, because there are some things we can do on the margin. All sensors have varying degrees of performance on day 1 that might be improved. Again, reducing some outlier points over the course of the full wear period, things like that. And then it does become -- our innovation, going forward, really does turn to software. Predictive analytics, much more robust tools. I would tell you, for the intensively managed patient, probably a different software experience entirely, as we look to either the hospital or the non-intensive type 2 patient population. We're not really even sure what that user interface and what that software experience needs to be. That's part of the work that Matt Dolan and his team are doing. Frankly, that's what they've been tasked with. But I think it will be different. And I think -- I don't think we've reached a place where G7 is the end for the intensive business. But I think, certainly, from a form factor perspective, I think we expect to live with the on-body experience of G7 for a long time, particularly with the amount of capital we're going to spend in terms of scaling, manufacturing and the automation we're putting into G7. But then you take that, you can certainly change membrane technologies on that same product platform. You can change algorithm technology. That's just software. And you can change user interface. And I think that's why you see the technology evolve over the next few years.

Would it be fair to say that, just looking historically, DexCom has never stood still on performance? And acknowledging it's already a very accurate product, there still are outliers, and you see it. And even -- the data I found most interesting, as you present the data on, like, manufacturing lots over the years and showing the distribution of sensors and how -- it's really almost just a couple of sensors within, however many, a lot that you would think would drive a bad user experience. What's the driver of those outliers? And how do you continue to improve upon that and eliminate them over time?

Yes. I mean, there's a host of factors. In the past, it could be something as simple as a bad calibration point. If the patient took a bad -- and you don't ever want to blame the patient, but if the patient took and entered a bad calibration point, our algorithms, even with Gen 5, got much more sophisticated, where they could -- if they -- the algorithm's detected that the patient took a bad calibration, they would actually ask for -- the algorithm would ask for another one. I mean there's a host of issues that cause sensor error. But I think you're right in the sense that what we need to do, and then some of it's even just manufacturing variability. You mentioned lots. So one of the things that used to drive us crazy, prior to the iCGM designation, was competitors would have this practice of building multiple lots of sensors and then kind of cherrypicking the lots that actually worked and putting those into a study and then comparing it head-to-head with DexCom, and saying, "Look, we're just as accurate as DexCom." That's the beauty of the iCGM standard is that isn't -- that can't happen anymore. The FDA has now required, in any iCGM trial, multiple lots of sensors. You're required to manipulate patients' glucose levels up and down. So you have to run a meaningful number of points all the way from severe low to severe high and capture data and compare it to a YSI reference. So again, back to the -- the iCGM standard is really hard. It's really hard to meet. But we're going to get there.

The manufacturing might be the most underappreciated element in this -- the whole story of developing the sensors. You have -- from the production, you have your algorithm, you have your sensor membrane. But to be able to manufacture at scale and to make millions upon millions of things with very little variability from sensor to sensor, you have to control a lot of elements in the process of manufacturing. It's expensive to do, and it's something you really have to become an expert at.

Okay. Let's shift to the non-intensive type 2 opportunity. In our opinion, the progress being made within that opportunity, by both DexCom and partners, is one of the most underappreciated aspects of the story. You've talked about progress across 3 fronts: payers with UNH with their own programs; third-party software vendors like Livongo, Onduo and others; and then DexCom directly with providers, and I guess, Intermountain is kind of a tweener as both a provider and a payer. On the Q4 call, you talked about starting to expect some material contributions at some point in the 2021 to 2023 time frame. How fast could this scale in 2021 to 2023 in that time frame and for contributions to be material enough to really notice it externally to really move the needle within your overall financials? Would that require like a massive scale up in UNH or something like that? Or could that could be driven just by continued progress that we're kind of seeing already today?

Yes, again, I got to be a little careful. We don't -- just to be clear, we don't have anything baked into guidance this year for contribution...

Long term. From...

Yes, from that non-intensive patient population. But yes, I think, over the next several years, 2, 3, 4 years, you're absolutely going to see a much greater level of contribution. The one thing I would caution, though, we can be sitting here 5 years from now and the number of patients that we have using DexCom, in the non-intensive world, the number -- aggregate number of patients might be much higher than our intensive business. But the average revenue per patient in the intensive business is still going to exceed, what we believe, we'll find in the non-intensive business. So even 5 years out, I would tell you, I believe our intensive business, from a revenue perspective, from a contribution perspective, is going to be bigger. But at some point, there's no reason, in the longer -- I think someone in an earlier one-on-one said, in the longer, longer term, there's no reason that -- because there are just so many people, whether it's in the U.S. or globally, 10 years out, the numbers, even, like, in China, they're saying 1/3 of the population is either going to be diagnosed with type 2 or be prediabetic. I mean, the numbers are just staggering. You're talking about 500 million people. So over time, I think the non-intensive -- the application of this technology into non-intensive type 2 and even migrating down to prediabetes, using our sensors, we're fully convinced that if you apply the technology to someone who is at risk of developing type 2, there's absolutely no reason you can't intervene early, potentially even keep them off of metformin, just modifying diet and exercise using our technology and keep them from ever developing type 2 diabetes. I mean the technology is remarkable in that respect.

Yes. I think, United hasn't shared much, but one of the details they shared on their website is like 77% of the patients in the trial had an altered decision made on therapeutics, which was fascinating.

That's correct. Yes.

Let's talk a little bit more about United. So the pilot that began -- I believe, began in 2018 with 10,000 patients. We've heard it scaled significantly, although you haven't told us much.

They won't let me.

How do you think about the potential of this partnership over the next 1 to 3 years? UNH representing about 15% of total covered lives in the U.S. It sounds like they have big ambitions for this program that they're developing together with you. And if that was to scale up, do you believe it will be done exclusively with DexCom CGM as we sit here today?

Yes. So our relationship with United is not exclusive. However, they are certainly using our sensors as part of this development process. I think there's a fundamental belief that our sensors -- I mean, there's -- not even a belief, our sensor is the most accurate sensor available today. And they recognize that the need for accuracy, even in that -- there's -- we've had questions before is, do type 2s who aren't taking insulin need as accurate of a sensor? And I think the belief out of United and others is absolutely, that accuracy is still paramount across all of the -- across the spectrum of these patient populations. So like you said, I mean, United has been very close to the vest on the data, they're not sharing data with us. They're not going to share data with you guys because they deem it to be very proprietary. United is just building -- they're looking to productize this, right? This is going to become something that will be part of a diabetes product or diabetes program that they will, in turn, then take to employers, take to even potentially other health plans, take to -- take to Medicare, right? So they try to gain more Medicare Advantage patients, so that kind of thing. So I think it could be huge. I'm not going to try to quantify it for you. But I think it could be huge over the next several years. And if anybody has the firepower to do it, it's obviously United. And they're super-progressive in their endeavors here. So I think it could be big.

And do you think of that domino fell, is there any reason why other commercial payers would not follow suit? I guess, maybe taking a step back, have you had any conversations with other commercial payers? We've seen small pilots, but nothing seemingly at the scale of how United is talking. Do you think that's just a matter of time?

Yes. It's really nothing to the scale of United. Obviously, if United is successful in launching, kind of, a diabetes program that they're selling to either employers or otherwise, certainly, you should expect that there will be other followers. They're going to have to build it from scratch. We're not permitted to share even the limited amount of information that we have that we've developed as part of our programs with United, would be proprietary. We're not certainly precluded from providing sensors to other payers. If they're looking to develop programs, we would certainly look to do that. But today, United is by far the biggest for us.

Okay. Let's briefly talk about some of the third-party software vendors we talked about before. So you just recently, announced the partnership with Livongo. You're obviously working closely with Onduo. Across all of these companies, we estimate they're already greater than 500,000 patients using these programs, almost all of which are fingersticks today. But as CGM continues to get cheaper, you used G7 being a great data point on that trend, why wouldn't all of these patients be using CGM over the next 5 years, longer term?

Yes. I think it's a function of cost today, right? I think -- absolutely, right? I mean, if your business model is premised on taking in data and running analytics on it and/or looking for anomalies and then providing proactive coaching or advice to the patient, having the equivalent of 288 fingersticks a day versus an average patient is going to take 4 to 6, maybe, fingersticks -- a motivated patient, probably 4 fingersticks a day, the wealth of data that we provide is amazing. And to be able to enable, not even singling out any one particular company that you just mentioned, to be able to provide them with the data, to be able to apply their technology, whether it's automated coaching or live coaching or a combination thereof, the dataset that we provide is just so much more robust. And we're certainly working on tools internally. I was asked a question in one of our one-on-ones today about Livongo specifically. Don't you deem Livongo to be a competitor? And my response is absolutely not. I mean, they could be a very good partner for us going forward. Even if DexCom -- even if we'd developed our own data management tools and we're providing these types of services directly to employers or to payers or otherwise, if we're empowering others, they're still using our sensors. So at the end of the day, if we're selling sensors to -- regardless of who we're selling them to and to which programs, we benefit. So we're not looking at any of these guys as competitors. They're good partners.

Should we expect you to launch your own solutions for this patient population over time? Or should we think of you to remain exclusively as an arms dealer to all the players?

We're working on both. I don't have any near-term time lines to give you on our own stuff. But we're certainly in development on some things.

Okay. And then briefly on Intermountain. So $5,000 in annualized savings per member, like you mentioned, all direct savings, which, I think, is really fascinating relative to a lot of the data presented in the space, which is including all the long-term model that you talked about. I guess taking it up a level, how would you characterize the body of clinical data today? It seems like it's moving very quickly. Where do you think that will be over the next 1 to 2 years? And at that point, will it be ready for prime time? Just thinking back to -- I know it's a very different patient population. But even when everyone knew that CGM was really effective and was changing lives in type 1, it took a long time for everyone to finally embrace it. And maybe this is different with payers actively pushing it. But how do you think about the clinical data today and how do you see that evolving over the next couple of years?

Yes. I think the big differentiating factor there is that although type 1s are expensive patients for any given payer, there's just not that many of them sitting in any given payer. And so the economic burden of those patients is not as significant. If you go meet -- even we meet at the highest levels with a number of payers, right. You ask them about the costs associated with type 1. They don't have readily available data. If you talk to them about their type 2 patient population, they know how much these patients cost. They know how much, as you migrate them from metformin to the next drug to the next drug, they know to a tee, because there are so many more patients, and they're so expensive for these payers. So what I would tell you, there's a significant level of additional motivation on the part of the payers to reduce cost, whether it's preventing someone from going on the next most expensive compound, whether it's an SGLT2 or GLP-1 or insulin, even for that matter, or if you can really dive in and really get the patient motivated and under control, you can actually dial back their drug regimens. The cost savings are just astronomical. And even, frankly, over time, as our business model evolves, you're aware that we don't really know what the business model is going to look like 3 to 5 years out. We would look for opportunities to share in that cost savings, right? For -- the Intermountain example, if we're saving $5,000 per year per member who's in a program using CGM, we might be better off going at risk and actually sharing in the benefit of the savings, right, rather than just selling sensors. So these are all things that we're going to explore, and we'll report out as we have more concrete data and more concrete evidence to show you guys.

Great. All right. I'm going to shift back to a couple of G7 questions. One other fear, you can say, that we get occasionally from investors is -- or when they launched G7 and they have the cost capability to compete lower, why wouldn't they lower price and drive volumes up? Or what's the risk of that occurring? I think your commentary has been pretty clear in terms of relatively linear annual revenue per patient declines and you centered a lot of your pharmacy contracts on annual revenue per patient. I guess, as you sit here today, what do you think the risk is of any proportion of payers not looking at you as annual revenue per patient? Does that make sense? And I guess, as you launched G7...

Yes, I mean, that's what we've been trying to shift between sell-side and even all of our buy siders, we have been really trying to shift the mentality from what was historically component pricing, sensor transmitter or receiver to thinking of it more holistically, like, this is what we expect to receive on a per patient on an annualized basis. And that number has come down. And today, we haven't given a -- like an average price. But you guys know Medicare is around $2,500. We've said DME -- typically, DME in the U.S. would be slightly higher than that. But on the pharmacy side and particularly OUS, where we go through the third-party distribution, that average revenue is already lower than, maybe, even you guys anticipate, lower than Medicare, for example.

Is [ pharmacy ] in the U.S. north of $2,000?

I'm not going to comment on that. But we're certainly looking forward -- we're not going to go to Abbott pricing. So if you think Abbott's around $1,500, we know we're going to get a pricing premium to Abbott. We've had discussions with the payers. It isn't going to be -- if I told you 2 or 3 years ago, our average revenue per patient on an annualized basis would have been about $3,000. It's not going to be $3,000. It's not going to get at 2x to Abbott. But we're going to maintain a pricing premium. And the payers have been pretty clear that they're supportive of that because of the performance attributes of our technology. And then that's talking specifically the intensive business. Again, the non-intensive business is still much evolving in terms of the business model.

Last one. So you've talked about being able to manufacture G7 at a $10 to $15 cost at scale. Even if you came down to Abbott pricing, that would seem to support at least a 65% gross margin, if not higher. As we look out over the next couple of years, is that an area where you see clear room for upside versus your targets if everything plays out as expected, if you maintain a premium? Or am I not able to do the math?

So what I'm -- the way I'm going to answer is to simply say that with G7, the COGS profile of G7 at scale -- and remember, when we scale up over the 2021 time frame, we're not going to be at a $10 price point -- cost point for G7. But once we scale up and we get there, we can be very flexible in what we do with G7. So I don't want to give you any specifics on gross margin over the long term. We may do that again at the end of the year when we do our next Analyst Day.

Great. And with that, we will take all other questions in the breakout, next door, Salon K. Thank you, guys. Appreciate it.

Thanks.