From PacBio, and yeah, we'll kick it off here. Let's actually just start off with let's talk about the importance of long reads, and actually let's add to that the importance of accuracy. While we're on the topic of accuracy, can you talk about you know some of the advantages of your really low non-random error rate? Because you know I think it's pretty funny 'cause you know 10 years ago people used to talk about PacBio as being so inaccurate, but there was a mischaracterization of the accuracy, so.
Yeah, well, thanks, David, for having me here today. So, for those of you that don't know, PacBio really has been built around long-read technology, and it's a single-molecule long-read technology, which both are important for the types of ways you look at the genomes or the transcriptomes and for the non-randomness of our errors. So a long-read, you know, we're looking at stretches of DNA that are 15,000- 20,000 kilobases in length, where short-reads look at 150 bases- 300 bases. So significant difference in terms of the length of the DNA that you're sequencing, when you do this, which makes the assembly of the genome much easier.
You know, if you've got 20,000 bases of DNA to overlap and to align, you can start to assemble a complete chromosome and a complete genome, which you can't do as easily with short-reads, and so the differentiation really comes from the read length that we have. That read length lets us look at structural variations 'cause we sequence across all of them. That looks at, lets us look at tandem repeats and lengths of tandem repeats that you can't look at or assemble 'cause we sequence through the entire region. It also looks at phasing, so we know what DNA comes from the maternal versus the paternal haplotype, which is really important for disease and cis or trans disease associations, and because we're also doing it with a single molecule, we're not amplifying that DNA.
And that single molecule is unaltered, so we can start to look at different methylation effects on that DNA. And so when you do your DNA sequencing, you get your epigenome and your genome at the same time. And so we look at 5-methyl-C, which is becoming really important for aging and disease association and a lot of these tandem repeat disorders. The single molecule also means we're not amplifying it, and that's where you start to see issues coming from a technological approach to amplifying DNA or having to manipulate that DNA. And so we get a very nice evenness of coverage across that genome, so there's no real GC bias of any sort.
If you step back and think about PacBio years ago, it was deemed as error-prone, and it's because we only looked at a single pass with that DNA. When we launched the Sequel II and we came up with the HiFi approach, we now get a consensus of that single molecule. We sequence that same 20 kb molecule of DNA multiple times over the course of 24 hours, which gives us that really high consensus accuracy, which is really what's helping to drive us into clinical applications.
Gotcha. You know, just moving back, I mean, you have been in the market for a while, so I'll ask some move on to some of the, like, the topic questions of the day. I mean, you just moved your maturity back. What, what trade did you make, and, and what kind of feedback did you get from investors?
Yeah, so it's been very positive. I felt like it was a little bit hidden after earnings in that we had just launched a new SPRQ chemistry for Revio and a benchtop Vega, which I'm sure we'll talk about some more. And we'd lowered some guidance, and so I think those were the main topics right away after earnings. But moving that note with SoftBank back another 18 months was really important. And so over the last two years, we've been strategically managing that. And so for those of you that don't know, we started off with a $900 million convertible note with SoftBank. Last year, we pushed $440 million of that out with a single holder to mature at the end of 2030. And we still had $459 million that was due in February of 2028.
As we got closer to February 2028, we still feel like we had time, but we wanted to make a move on some of that note. And so we did a note exchange, and $200 million of that is pushed out 18 months. So now the SoftBank note is $200 million, October of 2029. And the other $259 million that was remaining, we paid $50 million in cash, $50 million in stock, and they wrote down the other $159 million. And so we turned $459 million into a $200 million note through the end of 2029. So we felt good about that trade. They felt good about being an equity investor in us and giving us some more time to work through that note.
Gotcha. Well, well, you know, you mentioned Vega. You launched it a couple weeks back, benchtop sequencer. What kind of market does Vega open up for you? Obviously, it's a lot smaller, can go in a lot more labs, lower price point, but, you know, add a little bit more to that. And can you talk about some of the outsourcing dynamics in the market and, you know, hoping to see lower prices overall? So how do you manage the relationships with the service providers?
Yeah, Vega and a lower CapEx instrument is a really important next step for PacBio. Over PacBio's history, it's only ever had a single platform on the market and just every few years launched an upgrade to that, to cannibalize the old system and upgrade that same customer base. And this is the first time we've had both a low-throughput system and a mid-throughput system in the market at the same time. That starts to address different customer bases with different project sizes and different capital budgets. It was really important when we launched Vega that we also launched a SPRQ chemistry to improve the performance of Revio. So Revio's the higher-throughput version on the market. And so with the SPRQ chemistry, for a $1,000 list price run, you can get 120 gigabases versus 90 gigabases.
And so that was an important step because we want to further differentiate Revio from the benchtop Vega. And so for $1,100 on Vega, you get 60 gigabases. So it's half the output of a Revio SMRT Cell for about the same list price of reagents. And that gap in pricing is important to make sure that we're addressing a different customer base with, with Vega than we are with Revio. So far, the early feedback's been great. I see, I see Vega, really starting to do what we would hope it to do is to decentralize the use of HiFi and get more people aware of what PacBio can do. And, and that's gonna be a really important aspect of how we, how we think about, global access to, to our sequencing data.
We're already seeing it through the smaller core labs that couldn't afford a $600,000 Revio, turning towards Vega. A lot of our Sequel II customers that we haven't been able to convert to Vega are now looking at Revio, and so we're seeing that pipeline build up for some existing Sequel I and Sequel II customers to convert. And importantly, I think it's gonna become a really good targeted RNA and targeted DNA box. I don't see it as a whole genome platform, but sitting beside Chromiums for single-cell with a long-read readout to look at the full transcriptome or some of the bulk RNA momentum we're getting for looking at isoforms and alternative start sites and all the great science on the RNA side of the world, it's gonna be a great fit for. Then all of the targeted panels for tandem repeat disorders and carrier screening. We're really encouraged by early targeted desire for Vega.
Got it. Well, you know, we can probably talk about the elephant in the room of life sciences, which is, you know, CapEx cycles and, you know, obviously, it's been a brutal environment for everyone from Thermo to, you know, your sequencing peers like Illumina, all of it down are down year over year. How do you think about the CapEx cycle next year and beyond? What are some of the things that investors can look at for, you know, light at the end of this tunnel?
Yeah, it's been a tough year for us and a tough year for us to predict as well. I don't have any predictions for 2025 CapEx turning around, and I think we'll responsibly keep that pretty cautious as we move through. Now, a lot of the moves we made in the last month with Revio, extending the capabilities of Revio and lowering the list price, were because of the headwinds we're seeing on CapEx. Our average selling price on Revio was closer to $600,000, but our list price was $779,000. Customers just self-serving on our website would be pricing themselves out of it earlier.
It was really important that we've made so many improvements on our cost of goods in making Revio, that we could afford to sell at a $600,000 or less. It was important for us to reset that list price so that the general academic community would see that it's a $600,000 box, not a $779,000 box. So that is helping move some of our funnel forward, 'cause they didn't realize the types of discounts we were giving. And then on the operating cost of it, it all goes into the same equation when they look at the return on investment. You know, we just increased the output of the chips by 33%, but we kept the price the same.
And so because we're still deemed as too expensive, we passed all of those gains on to our customers so that they're now gonna get that effective 33% cheaper cost of a genome. And so now at list price, we are truly at a $500 HiFi genome, which was important for us to get into the market to help drive the momentum around it. So less around just the capital concern, but just the general awareness for what PacBio has with now affordable and scalable long-reads so that we can continue to drive through that capital conversation with them.
Gotcha. And then, now moving on to the next thing that you can't predict, which is NIH funding in the years ahead, but you know, something you do have access to, which is customer end markets, or your customer concentration to NIH funding.
Yeah, I can't speculate on what new leadership to HHS or NIH might mean yet, but we'll stay close to it. So it's good news and bad news for me that, you know, the good news side of it is that we're not that exposed to NIH funding here in the U.S. Now, it's hard for us to triangulate because a lot of people with NIH grants may be outsourcing to other companies that we don't track as NIH dependent, but we think we're in around the mid-teens exposure of the U.S. business. And so it's not so much exposure that it gives me heartburn, you know, the bad news is that it's still a small piece of our business and it should be higher.
I actually look to Vega as a way for us, even in a flatter declining NIH budget year, Vega opening up opportunities for us with, within the academic community here in the United States because of the price point. And the NIH is doing a great job of funding smaller grants and young investigator grants and startup grants for PIs that just, you know, they're not getting grants large enough to spend $600,000 on capital, but $160,000 is much more affordable. And so we do think this is gonna give us a better opportunity to get into the smaller young investigator grants and smaller grants than the NIH tends to fund at a higher rate these days.
Gotcha. Can you talk about, you know, I think that was interesting on the Revio and the decrease in list price. Can you, I mean, it was already being sold at six hundred, so, you know, no decrease to your gross margins, but, you know, can you just talk about your gross margins overall? Are they improving? How are you improving them to get them closer to your peer group?
Yeah, so as much as we've struggled this last year on revenue and predicting the revenue, for things that we can control internally, I think we've really done a wonderful job. You know, the main focus initially was on gross margin, and then it really was on getting through our products and product launch and development timelines, and so I feel like we're controlling the internal operations well at this point in time. On the Revio, we took out, you know, 10%-15% of the COGS this year, mostly through insourcing. You know, we had some contract manufacturers, and when the volumes were coming down, we took the opportunity to unwind some of those agreements and fully insource.
And so as of this quarter now, we're fully from chassis up insource building our own Revios in our Menlo Park facility. And so that was the first wave of cost reductions that we've seen. And we've got a second wave of cost reductions hitting in the first quarter next year. We thought it was initially gonna hit this year, but with the slower demand for Revio, we won't get to those units until you know early next year, unfortunately. But we've been able to dramatically reduce the compute costs in the system. And that's just through this new SPRQ chemistry and some more efficiencies in our compute algorithms. We've been able to reduce the number of GPUs in the system, which is one of the most expensive components of a Revio because these have fully capable compute infrastructure built into them.
And so we've been able to reduce the cost of the compute, so we'll gain another 10%-15% of margin improvement through the COGS or of COGS improvement through that. So all those in combination let us feel really comfortable with a new list price at $600,000 for Revio, and our margins from that should be flat to up. Vega's also important for the margin on the Revio instrument because, in the past, it would be, you know, a stalled sale or a slow sale or can I get one for really cheap, and it would be a long drawn-out negotiation where now we can point to Vega. And if they can't afford the Revio at this point in time, then maybe Vega's the right platform.
Gotcha. And just a clarification on that, on the gross margins, is that 10% reduction on gross margin or just, or sorry, on the COGS?
The cost of goods.
Was that on cost of goods sold generally or cost of goods sold on Revio?
On Revio.
Okay.
That was on Revio. Yeah, yeah. No, that was all, that was all specific to the Revio platform. On the consumable side of Revio, you know, the margin increases there are, you know, as we continue to get better yields off of our manufacturing of those. And then volumes. And so we were hurt a little bit with the volumes this year as those came down on the Revio consumables, but I expect that to all be a tailwind for us next year.
Got it. And then, just, I think, you know, we've gotten to some of the cost cuts that it sounds like you've made, but, are you on the right side and what would you do from here, if you did make any cost cuts? Like, what would be the low-hanging fruit that would be kind of remaining?
On the product gross margins or?
No, no, no, just generally in the business on the OpEx.
Yeah, yeah. So you'll see the full effect of our cost cuts in 2025. And so we've done a lot of work to gain, you know, $75+ million of cost savings for next year on the OpEx side. I know those were difficult decisions. We shut down both our Baltimore site as well as our San Diego site. And so we laid off, you know, around 200 of our 800 people. But we've consolidated all of the manufacturing from those sites as well as all of the R&D from those sites into our Menlo Park facility. And with a smaller force, we're actually gaining a lot of efficiencies. And it's a lot easier when you don't have multi-site management.
And so the manufacturing efficiencies come through as labor and overhead rates, which will help all of our gross margins 'cause we don't have to spread those across three bigger facilities. So you'll see some margin improvement from labor and overhead reductions next year because of the consolidation. But the big thing to me is the R&D. And San Diego R&D was mostly on the short-read side. And with our Apton acquisition, we had enough critical mass in Menlo Park that we felt like we could consolidate short-read R&D up there without an impact on that timeline. And happy to say that our timelines for getting that high-throughput short-read Apton box to market haven't been impacted by the San Diego site closure. So real gains of efficiencies and cost savings. It's that discipline that I think is gonna be really important for us over the next two to three years.
Gotcha. You know, going back to the market demand side here, you know, can you talk about the play out between, you know, the $500 genome that you're giving to customers that are doing Revio at service providers versus the, you know, the $1,000, $1,100 genome you're gonna get at Revio? It does seem like customers do tend to find those service labs, you know, pretty well in terms of their ability to find them and access them. So can you just talk about that trade-off dynamic, that calculus? And I don't really know necessarily what you can say, but just cost of goods and.
I don't think they were finding them fast enough. That is probably where I'll start. Vega is $1,100 list price. Yes, you could do a genome because it is that 60 gig of output. They could do their own genome for around $1,100 plus some cost of capital in there. The reality is most of the core labs, if you go to the websites and/or get quotes for whole genome sequencing core labs, they're charging somewhere between $2,500 and $3,000 a genome. The world still thinks that a PacBio genome or HiFi genome is closer to $3,000. It was really important that we put that message out there that no, it's now $500 for that data.
And/or if you wanna do it yourself, it's closer to $1,000 with your own Vega. I think that's a really important next step. Ultimately it will drive more business to those core labs because they'll do some small studies and they'll use their Vegas. Ultimately when the sample sizes become large enough or the scientific project becomes large enough, they'll want to find a cheaper source. That's the dynamic you see. You know, you're used to projects of thousands of samples that get outsourced all the time. You know, when we get thousands of samples of projects, we talk about it on our earnings call. We need to get to a point where a thousand sample projects are more routine for us. That's why I think Vega's important to expose more people to it.
Got it. Now, let's just maybe talk about the Constellation product. Now, synthetic long-reads, I mean, probably have a 10-year history of not really working very well. Though I would argue at least that I think Illumina has done a good job of getting more data out of, you know, library prep and, you know, maybe some of that stuff that could have been long-reads before, maybe stayed with them a little bit longer than it naturally should have. So, I mean, is there that at least maybe that dynamic in Constellation? And you know, I am kind of making you compliment my recommendation.
So I don't know enough about Constellation yet. You know, they, they've had some very early access users of that. The NovaSeq X, you know, the NovaSeq line does give them enough real estate on a flow cell to make that more affordable than what they might have tried with it eight, 10 years ago. You know, and so we'll have to wait and see. I still, you know, in my mind, it's still not necessarily gonna be the most economic long-read version or synthetic long-read version of a genome. And I still see using short-reads to assemble a longer read as a synthetic long-read, which to your point, you know, 10x has tried. Illumina's tried a couple of times. Element's tried. So I don't think that's a novel concept.
I just think that what we're doing is we're trying to make a single-molecule long-read genome, you know, price parity with short-reads and simplify the workflow. And anytime a short-read technology has to add in library prep and add in complexity and bioinformatics to do different assemblies, they're adding complexity and they're adding cost to it. And so as we and other long-read, single-molecule long-read technologies try to drive prices down to become price parity, I think we're gonna continue to be viewed as the long-read technologies to choose in the future.
Gotcha. And so let's just talk about the complementarity of short reads. What would you consider in terms of what's kind of the current pathway with Onso? And would you kind of consider any kind of strategic change with Onso or short reads?
Yeah, I still love the strategy of having a short-read technology as well as a native long-read technology. I think that the application should determine what technology choice that they make. And so when we got into the short-read technology with Omniome and then the Apton acquisitions, we did it intentionally to push down the idea of more accurate short-reads. We'll open up markets like infectious disease, wastewater management, microbial surveillance, you know, a lot of cancer applications and around early detection or MRD, liquid biopsy applications and infectious, you know, transplant genomics is one of those. And so we saw this market need where short pieces of DNA that didn't fit with our long-read technology, we're gonna benefit from more accurate short-reads. And so we've proven that out with Onso.
So there wasn't about selling 500 Onso a year in my mind because those applications need billions of accurate reads, and Onso does 500 million. So there's a mismatch in the market need. But we've now sold enough Onso into the world to validate our thesis that people that are doing wastewater management and infectious disease and oncology want to use Onso for the inherent more accurate data that they're getting off that, which improves the sensitivity, and it means they don't have to do a whole bunch of, you know, molecular biology and/or hurdles to get to the error correction. So they're proving to us that there's a real need for the Apton high-throughput. And so strategically, you know, as a stepping stone, I think it's done a really good job of validating our thesis.
We acquired Apton so that we could gain time and start integrating that same chemistry on a high-throughput box that was already sequencing billions of reads, as well as the R&D investment through that stock acquisition of it. You know, that's also why we were able to cut our R&D expense so much is we had bought Apton to leapfrog into having a platform to integrate. I'm really excited about pushing on that high-throughput version of the Onso box, and taking that out for specific markets. I would take it out to specific markets 'cause you know, I've never thought that you're gonna win the short-read space, you know, based on price or throughput. I think you have to do it differentiated, and the last vector of differentiation was accuracy.
Gotcha. Very interesting on the Apton. Do you have any? What's the latest in terms of timeframe and for high-throughput?
Yeah, as we mentioned when we first acquired them, you know, it was always gonna be a multi-year development program, and I still think it's within a couple of years. But we're sequencing on it every day. We're integrating the chemistry. We're pushing on the density and the read length. And so, I'm feeling really good about what the team has done with that box. And we're fully staffing that program right now to get there as quickly as we can.
So, high-throughput SBB, probably price parity with some of the other two, you know, Ultima, Illumina.
Yeah, you have to be able to show that you can do it at that margin, like an acceptable margin, you know, at that $1-$3 per gigabase, you know. And in that time horizon, I expect most of the short-reads to be in that $1-$3 per gigabase. And so we have to be thinking about where the market's gonna be. And we feel really good about both the acquisition and the status of that program.
Gotcha. What's the potential for multi-unit orders right now in terms of your revenue customers? You know, kind of what's the dynamic then? You know, is there a chance for upside from?
Yeah, we didn't, we didn't have as many multi-unit orders last quarter as we had, had we had prior quarters. And, and you've seen that come down since the initial launch of Revio. Now there, now it still exists, but I think it's, you know, is it two or three? I don't, I think it's fewer of the five and 10 at the gate. I think it's, you know, people come in and they've got a big project or they've, they've got a big use case for it and they're probably gonna need, you know, two or three Revios. I, and so that would be upside. You know, we, we generally forecast them as individual units by customer and, and the multi-units become upside for us on the quarter. But, it has come down over time. But when I look at the pipeline of large projects that we're focused on, I'm encouraged by what that could mean for the next 12 to 18 months.
Got it. Maybe just talk about, as we end it, the long-term revenue target growth targets and getting into cash flow breakeven. Now when you gave your long-term targets, you know, nobody really knew what the capital cycle was gonna look like in the broader market. You know, again, can you reiterate what those targets were at that time? Do you have the levers if, say, capital cycle continues to be where it is to get to cash flow breakeven, you know, in a constrained and the unfortunate assuming the unfortunate just continues like what happened with capital cycle?
Yeah. I'm not gonna give any new guidance today. We'll get to that next year, but we still are committed to exiting 2026 as cash flow breakeven with 2027 being cash flow positive. Our timeline from our long-range plan has not changed even though the revenue is lower, and a big part of that is the discipline that we have and how we go about our spend. We have shrunk the workforce. We have shrunk the number of projects we're doing internally. We've controlled the spend where we are. I'm not saying we're not still innovative. We still spend a lot of money on R&D, and innovation's important for the future.
So we haven't stifled innovation, but we've really prioritized what we're doing across the organization to make sure that we can still hold that positive. Now we have to return to growth, and I think 2025 we'll grow over 2024 and 2026 we'll grow over 2025. So we're obviously returning to growth and then continue expansion on gross margin. So I wanna get those gross margins up to start with a five, and I feel really good about then just managing the OpEx and the spend in a disciplined way to make sure we still hit that same goal.
Thank you so much for joining us.
Great. Thank you.