Illumina, Inc. (ILMN) Earnings Call Transcript & Summary
January 29, 2020
Earnings Call Speaker Segments
Jacquie Ross
executiveGood afternoon, everyone, and thanks for joining us on the call covering the fourth quarter and full year of 2019. Today we'll go over the financial results that were published following the market close, and we'll also share some commentary regarding our commercial activity, and after that we'll open things up for a question-and-answer session. Should you not have had an opportunity to look through the earnings release, you can locate it within the Investor Relations portion of our site, illumina.com. With us from Illumina today are Francis deSouza, our President and Chief Executive Officer, along with Sam Samad, our Chief Financial Officer. Francis will share a brief status report on where things stand with the business, and then Sam is going to take us through the numbers. Please note that this call is being recorded, and a replay of the audio will be posted in the Investors section of our website. We want all statements made on today's call that look ahead — whether about financial performance or commercial activity — to receive the protections of the Private Securities Litigation Reform Act of 1995. Such forward-looking statements carry risks and uncertainties. What actually happens could end up differing materially from what we project or discuss. Every one of these statements reflects information available right now, and Illumina has no duty to revise them later. For a clearer picture of the risks and uncertainties that might make actual results come out differently, please look at the filings Illumina submits to the Securities and Exchange Commission, which include its latest forms 10-Q and 10-K. Having said that, I'll hand the call over to Francis now.
Francis deSouza
executiveThanks, Jacquie. Good afternoon, everybody. Illumina finished 2019 on a strong note, posting fourth quarter revenue of $953 million, which represents a 10% increase over the fourth quarter of 2018. Among the highlights: sequencing consumables grew 22% year over year, and both IVD partnership revenue and non-DTC array revenue came in above what we had anticipated. Those gains more than made up for softer sequencing system revenue and DTC revenue that fell short of expectations. Some degree of mix variability is always going to be part of operating in our fast-moving industry, but we're happy to report that fourth quarter revenue came in above our expectations. Looking at 2019 as a whole, we brought in $3.5 billion in revenue. That's a 6% increase, and it lines up with what we told you to expect back in July of last year. On the sequencing side, we sent out over 2,400 systems — more than in any prior year for Illumina. When it comes to NovaSeq, we achieved our goal for 2019, shipping about 320 units, slightly more than we did in 2018. Revenue from sequencing consumables overall grew 14%, passing $2 billion for the first time ever. Within that, high throughput sequencing consumables accounted for more than $1 billion. Meanwhile, the amount of data our sequencers generated went up 50%, which reflects just how much demand there is right now for genomic information. Looking at Q4 again, consumable usage on NovaSeq hit its peak for the year and actually the strongest mark since we introduced the platform, aided partly by the U.K. Biobank, which has now scaled up; during Q4 we sent out more than 100 NovaSeq systems. That number, as we expected, was over twice the amount shipped in Q1 of 2019. For 2019, pull-through per system on NovaSeq was roughly $1.2 million, and as we move ahead, our goal range for pull-through in 2020 sits at $1.1 million to $1.2 million for each NovaSeq system. So during NovaSeq's fourth year on the market, HiSeq consumables continued to shrink the way we expected — in Q4 they dropped under $100 million for the first time since NovaSeq came out. When the year closed, we wrapped up a look at our HiSeq customer roster, and it revealed an active HiSeq installed base of roughly 1,300. Through that same review, we identified about 600 HiSeq systems that are currently inactive but had been included in our installed base before. Beyond that review, Illumina received reports of around 200 decommissions across 2019. Turning to mid-throughput: 2019 was a record year for NextSeq shipments, with a worldwide installed base of approximately 3,600 as it stands now. NextSeq shipments in the fourth quarter landed under our expectations, which came down to customer timing. Demand for NextSeqDx keeps expanding and accounted for almost 1/4 of 2019 shipments, up from approximately 10% last year; it was a record quarter for NextSeq sequencing consumable revenue, once again fueled by oncology and NIPT. NextSeq pull-through per system was better than last quarter and landed at the lower end of the $130,000 to $160,000 target range. With our new NextSeq 1000 and NextSeq 2000 systems now launching, we'll revise our target pull-through range once we've built a sizable installed base of the new systems. We expect to ship approximately 500 NextSeq 1000s and 2000s this year. Although the majority of NextSeq 550 customers are likely to move over to the new platforms, we believe the NextSeqDx pipeline will stay robust, because our regulatory cleared system occupies a distinctive position. We can't wait to begin shipping the initial NextSeq 2000 units later in this quarter, and it's a real pleasure to report that orders have already started coming in. Now on to low-throughput. Revenue from systems came in under what we had anticipated for Q4, largely because of MiniSeq. Interest in MiSeqDx keeps surpassing our forecasts, especially in China, where the China NMPA granted clearance back in August of 2018. During the first part of this month, the PMDA in Japan gave its approval to MiSeqDx, which we anticipate will add to placements during 2020. Shifting to low-throughput consumables. It was a record quarter for revenue. MiSeq consumable pull-through climbed nearer to the $40,000 to $45,000 goal range, while MiniSeq pull-through landed at the bottom of the $20,000 to $25,000 band. Returning to sequencing consumables. Revenue for the total reached $2.1 billion, an increase of 14%, or more than $250 million, in 2019. Slightly above 40% of the sequencing consumable units we shipped went toward clinical use, covering testing in oncology, reproductive health and genetic disease and other; altogether, clinical sequencing consumables rose roughly 20%, or about $130 million, in 2019, reaching approximately $830 million. Oncology still accounts for roughly 20% of all sequencing consumables, and it expanded more quickly than clinical consumables overall in 2019, thanks to greater uptake of panels, comprehensive genomic profiling among them. With additional tests such as FoundationOne CDx and Guardant360 gaining coverage as companion diagnostics, the need for Illumina sequencing keeps expanding. On top of that, clinical trials such as Guardant's LUNAR fuel higher sequencing consumable usage in oncology testing. Reproductive health, once again, made up slightly over 10% of sequencing consumables, mainly a reflection of ongoing NIPT expansion from wider coverage in EMEA, where sample volume for our VeriSeq NIPT solution rose 80%, along with increasing adoption in China. In the U.S., reproductive health keeps expanding at a rate that is more modest than what we see in EMEA and China. Last of all, within clinical, close to 10% of our sequencing consumable revenue ties to genetic disease testing, which expanded a bit under the clinical average. Companies such as CENTOGENE are fueling that growth; it has assembled a genomic repository covering more than 450,000 patients across over 125 countries, and it is striving to help diagnose patients' genetic disease while teaming up with pharma partners in the search for cures. This also takes in some revenue from genetic testing firms such as AMRI, which provide tests customized for our genetic disease diagnosis, along with other clinical tests. Now moving to research and applied. Shipments totaling more than $1.2 billion accounted for slightly under 60% of the sequencing consumable units we shipped. Taken together, research expanded by more than 10% in 2019, with genetic disease and cancer leading the way. Research into genetic diseases includes population genomics projects like the Million Veterans Program and the U.K. Biobank. That expansion was also driven significantly by cancer research, which encompasses efforts such as the cloud initiative at St. Jude Children's Research Hospital. A database is being assembled by the St. Jude team, allowing whole genomes from 10,000 pediatric patients and survivors to be accessed, and this will assist researchers in gaining valuable insights into what genetically causes pediatric cancer. Additional research areas include microbiology, infectious disease testing, and work in cell and molecular biology. This covers efforts such as J-GRID, the Japan Initiative for Global Research Network on Infectious Diseases — a collaboration involving 9 countries across Asia and Africa that uses Illumina sequencing for research into microbial diseases. Now turning to sequencing services and other. We brought in $124 million in revenue. That's up 19% compared to the same quarter last year. The main driver was upfront revenue from the Roche deal. That was partly offset by GeL, which dropped to almost 0 in the fourth quarter of 2019, as we expected, ahead of the clinical ramp-up later this year. And then arrays delivered $116 million in revenue, which is down 12% from the same quarter in 2018. That's due to continued headwinds from our direct-to-consumer customers, partly offset by array growth in genetic disease research. Before I hand the call over to Sam, I want to say something about the novel coronavirus outbreak. Our immediate focus has been on our colleagues in China, and our thoughts are with the families and communities that have been impacted. Over recent weeks, Illumina has taken part in multiple efforts aimed at helping address the coronavirus outbreak. Researchers have already leveraged Illumina sequencers to determine and share the coronavirus's genomic profile in public databases — an essential initial move that supports creating diagnostic tests and, in time, a possible vaccine. Our team is collaborating closely with Chinese CDC laboratories to develop NGS testing protocols for the coronavirus and deliver the needed training. In addition, we're coordinating with our supply chain group so that systems and consumables reach labs handling the novel coronavirus as fast as we can manage. Our intention is to make these NGS testing protocols available to our customers, helping the worldwide infectious disease community as it ramps up its response to this danger. Beyond that, we are looking into charitable initiatives and partnerships so that sequencing of the novel coronavirus can be accessed, supplying sequencing and consumables to anyone requiring them in the battle against this epidemic. Having said that, I'll now pass things to Sam.
Sam Samad
executiveThank you, Francis. As we covered, revenue for the fourth quarter rose 10% compared with the same period a year earlier, reaching $953 million, with 14% growth in sequencing driving that result while a 12% drop in microarray partially offset it. Sequencing revenue overall came to $837 million, an increase of 14% over the fourth quarter of 2018, and accounted for 88% of total revenue, versus 85% in last year's comparable quarter. Sequencing consumables brought in $572 million, up 22%, or more than $100 million, against the fourth quarter of 2018, whereas sequencing systems revenue slipped somewhat on a sequential basis and fell 12% from a year ago. On the sequencing systems side, we are going through installed base counts so we can proactively determine which systems are active. Our work on HiSeq is finished, and we are now examining NextSeq, MiSeq and MiniSeq. Those installed bases will be revised over the coming quarter or 2 once the analysis is done. Importantly, reported revenue is unaffected by this in any way. That said, we do anticipate system counts will come down, meaning pull-through per system will rise as a result. Sequencing service and other revenue totaled $124 million, a sequential decrease of $14 million stemming from lower IVD licensing and milestone revenue as well as lower GeL volumes in the fourth quarter, though it was up 19% year-over-year. Results for the fourth quarter included an upfront payment tied to the Roche partnership we announced recently. Arrays made up 12% of total revenue in the fourth quarter, compared with 15% in the fourth quarter of 2018 and 16% in the fourth quarter of 2017. Array systems rose $2 million sequentially but were down $5 million against a particularly strong fourth quarter of 2018. Array consumables increased $18 million sequentially because of DTC seasonality, with one customer ramping ahead of the holiday season, yet they were modestly lower than the same quarter last year. Array services declined $6 million sequentially and dropped $8 million, or 32%, year-over-year, on weaker demand from our DTC customer. Turning to regional results. Americas revenue was up 5% versus the prior year quarter, as growth in sequencing consumables and IVD partnership revenue was partly offset by DTC headwinds. EMEA posted a record revenue quarter, growing 19% from the prior year quarter, helped by a strong contribution from the U.K. Biobank, which is now sequencing in full production mode. Greater China rose 21% from the prior year quarter, benefiting from an easy year-over-year comp tied to tariff-related stocking in China during the fourth quarter of 2018. The region grew 2% in 2019, as slower research offset very strong clinical growth. Lastly, APJ revenue came in at $73 million, up 4% from the fourth quarter of 2018, with genetic disease research and microbiology fueling sequencing consumable growth both sequentially and year-over-year. For the full year, the region grew 6%. Shifting to gross margin and operating expenses. I will focus on non-GAAP results, which include stock-based compensation. I'd encourage you to look over the GAAP reconciliation of these non-GAAP measures, available in today's release and in the supplementary data posted on our website. Just a reminder that whenever I talk about net income and earnings per share from here on, I mean the amounts attributed to Illumina shareholders. Our non-GAAP gross margin came in at 70.2%, which was pretty much what we had anticipated, and it fell by about 230 basis points versus the third quarter. The decline came from reduced IVD licensing and development revenue along with fluctuations in production and lab service absorption, though product mix offset some of that. Looking at it year over year, non-GAAP gross margin for the fourth quarter rose more than 100 basis points, mainly on product mix and greater IVD licensing and development revenue, with lower DTC service volumes providing a partial offset. Non-GAAP operating expenses reached $372 million, an increase of $42 million over the third quarter of 2019, mostly because of how OpEx timing was weighted toward the back half of the year, and that figure came in ahead of what we had expected. As a result, non-GAAP operating margin stood at 31%, compared with 36.1% in the prior quarter. Our non-GAAP tax rate was 18.5%, which was higher than last quarter and above expectations, driven by income mix across different tax jurisdictions. In the fourth quarter of 2019, GAAP net income came to $239 million, or $1.61 per diluted share, while non-GAAP net income was $252 million, or $1.70 per diluted share. Now turning to cash flow and balance sheet. Cash flow from operations totaled $443 million, aided partly by a $58 million quarter-over-quarter reduction in inventory; that reduction is part of an effort spearheaded by our operations and supply chain teams to make our working capital more efficient. DSO came in at 55 days, versus 54 days in the previous quarter. Capital expenditures for the fourth quarter were $57 million, and free cash flow reached $386 million. We also bought back $63 million in stock during the fourth quarter, which leaves $226 million still available for repurchases under the plan we currently have. We closed out the year with roughly $3.4 billion in cash, cash equivalents and short-term investments. Our weighted average diluted share count for the quarter was 148 million. Now on to guidance. For the full year 2020, we anticipate revenue growth between 9% and 11%, or $3.86 billion to $3.93 billion, which works out to an increase of about $354 million at the midpoint. With the new system launch and continued softness in DTC, we expect revenue linearity to resemble 2017, implying slightly under 22% of revenue in Q1, roughly 24% in Q2, roughly 26% in Q3, and roughly 28% in Q4. For the full year 2020, at the midpoint of our revenue guidance range, we anticipate sequencing revenue growth of about 14%. That figure includes sequencing consumable growth of around 17%, and we expect sequencing system revenue to rise year over year, reflecting the launch of our NextSeq 2000 and 1000, which should more than make up for the anticipated step-down in NovaSeq shipments; we also expect sequencing service and other to be about flat year over year. We anticipate array revenue declining roughly 15%, reflecting continued weakness in DTC. Keep in mind that DTC revenue accounted for approximately 50% of total array revenue in 2018 and fell to about 40% in 2019; for 2020, we expect DTC to account for approximately 30%. Looking at things by region, we're anticipating growth in China at a high teens rate, with clinical sequencing as the main driver; for the full year, non-GAAP gross margin should come in at approximately the same level as 2019. Operating margin is projected at around 30%, and our 2020 tax rate should land above what we saw for full year 2019, since a number of onetime discrete tax benefits showed up in 2019 and we don't anticipate them recurring in 2020. As a result, GAAP earnings per share should fall between $6.45 and $6.65, while non-GAAP earnings per share should be in the range of $6.80 to $7; diluted shares outstanding for 2020 are expected to be roughly flat relative to Q4 2019. Keep in mind that after the Pacific Biosciences merger agreement was terminated earlier this month, Illumina made a $98 million reverse termination fee payment to PacBio. On top of that, Illumina will pay $34 million in continuation advances, which could be repaid to Illumina should PacBio sign change of control agreements or bring in at least $100 million within a given time frame. These payments aren't factored into our EPS guidance, since we're still waiting on the valuation of those amounts this quarter. Turning to the first quarter of 2020; total revenue should land somewhere between $850 million and $855 million, given that customers are evaluating and revising their plans in the wake of our NextSeq 2000 launch. Non-GAAP operating expenses as a percentage of revenue are anticipated to rise by roughly 600 basis points on a sequential basis, largely because our bonus accrual resets at the beginning of the year. And first quarter non-GAAP EPS should come in between $1.20 and $1.25, with GAAP EPS at $1.11 to $1.16; with that, I'll turn the call back to Francis.
Francis deSouza
executiveThanks, Sam. We've gotten off to a really solid beginning in 2020. The U.K. Biobank is doing sequencing at scale, and we have a contract with GeL to supply sequencing services beginning midway through the year, which meaningfully boosts our PopGen visibility compared with where things stood a year back. We brought out our most innovative platform so far, the NextSeq 1000 and 2000. That platform provides the highest cluster density flow cell among all NGS systems, which pushes down cost per gigabase for mid-throughput users; and we hit our ambitious goal of completely integrating the hardware accelerated best-in-class pipeline that we picked up at Edico only 18 months ago. We're enthusiastic about our 2-site software suite, a potentially transformative solution that makes genetic disease diagnosis simpler and lowers barriers to adoption. I'm eager to share updates on it once it launches. We keep broadening our clinical portfolio, with both TSO 500 and 2-site NIPT moving through regulatory review. And we're growing our capabilities via partnerships with the world's leading clinical companies, among them Qiagen, Roche and Adaptive, to deliver the most compelling IVD menu offered on our clinical-grade sequencing system. As our customers uncover more biological insights by sequencing at greater depths and volumes across new and emerging applications, the expanding clinical utility of genomic information is growing ever clearer. That, combined with rising community awareness and physician adoption, will let more patients gain from the promise of genomics. With that, I'll ask the operator to begin the Q&A.
Tycho Peterson
analystI'll attempt to squeeze in a few fast ones right at the start. Regarding Q4, could you put a number on that Roche milestone? The partnership fees appear to have risen by $15 million, reaching $20 million; so did Roche account for the entire amount in the fourth quarter? Then, thinking about instruments going forward, after 4Q when instrument revenues came in weak. Are you taking any different approach with NovaSeq pricing to help push through what's left of the HiSeq upgrade cycle? NovaSeq ASPs seemed to slip somewhat in the fourth quarter. That's my first question; and then about NextSeq. I'm just wondering, when you consider it, is this an upgrade cycle for the existing installed base or is it market expansion, given that those 2 new systems are being introduced to the market?
Sam Samad
executiveYes. Thanks, Tycho. This is Sam; regarding the Roche milestone, we didn't put a number on it. It actually landed above what we had anticipated. So the sequencing and other line outperformed our expectations in Q4. I'd say roughly $20 million or so above expectations when it comes to sequencing and other. But we didn't specify precisely what portion of that came from Roche. You can view it as IVD revenues during Q4, with the bulk of those coming from Roche. As for your second question, I'll handle that one, and then Francis can weigh in on NextSeq. On your second question regarding the ASP for NovaSeq; it's not as though we're doing anything unusual by cutting the ASP or offering — or lowering the price or handing out discounts to customers. What we — this is simply a result of multiunit orders we have throughout 2019 as well. You can think of it as the U.K. Biobank, a number of placements we have with the U.K. Biobank that typically carry a lower ASP; we've had some NovaSeq placements in emerging markets, which also carry a lower ASP; so those are natural dynamics when you place multiunit orders with certain customers. But there's nothing particular about reducing the price across the board.
Francis deSouza
executiveAppreciate it, Sam. So Tycho, when it comes to the target base for the NextSeq 1000 and 2000 that we're pursuing, our focus is really — like you mentioned, this is an upgrade cycle for customers moving off the NextSeq 550 today. That's the source we anticipate most of the demand will come from for the NextSeq 2000, which launches this quarter, and then the 1000, which comes out later in the year. We also anticipate a smaller bolus of orders arriving from certain MiSeq customers. A couple of quarters back, we pointed out that we expect — that we're beginning to see MiSeq customers who are migrating up and purchasing bigger instruments. Those customers, we expect, will be considering the NextSeq 1000 and 2000; so that's the place we expect the bulk of the customers to come from. It's possible there could be a small number of HiSeq customers who take a look at the NextSeq 2000, but that's probably going to be a very small number.
Dan Brennan
analystI was hoping to get a bit more detail regarding NovaSeq — specifically that cycle center you mentioned, with placements declining in 2020. Could you help us understand what portion of the installed base you anticipate will have been upgraded by year-end, and provide additional detail on the magnitude of that decline? Then, second, regarding HiSeq consumables. Given that you're offering greater granularity there, what has been assumed within your consumables guidance for HiSeq consumables in 2020?
Francis deSouza
executiveOkay. Looking at the NovaSeq upgrade cycle, here's how things are shaping up. Our belief is that fewer than 50% of HiSeq/HiSeq X customers have begun moving to NovaSeq, meaning more than 50% of that installed base remains to be converted. This will unfold from now through 2024, when we end-of-life the HiSeqs. Those are the figures we anticipate. Regarding NovaSeq unit counts, last year proved very strong for NovaSeq placements, which has us genuinely excited, particularly since it marks year three of the NovaSeq upgrade cycle. We anticipate that figure will decline this year quite naturally, and we expect a similar pattern over the coming couple of years.
Sam Samad
executiveWhen it comes to HiSeq, Dan, we aren't providing particular details about the consumables level or what anyone should anticipate, including where pull-through is concerned. On HiSeq, we've stopped giving pull-through ranges, and the same goes for total consumables. As you heard during the prepared remarks, what we did say about Q4 was that HiSeq consumables came in under $100 million for that quarter. The trend keeps heading downward. Naturally, the intuition — and this is what you'd expect looking at 2020 — is that HiSeq consumables will keep declining. That said, we haven't offered specifics about what that dollar figure actually is.
Steve Beuchaw
analystTwo topics, please — one about China, one about the NextSeq launch. On China: I know that a few weeks back you shared a view that growth there should recover somewhat. Could you walk through how you're weighing the funding trends against, perhaps, simply softer year-ago comparisons? And then the bigger issue we're all facing right now with the outbreak — how much is that, or isn't it, something you're factoring in? Second, we're now a few weeks past — and I realize it's still early — the NextSeq launch, so how are your higher throughput customers, the ones who may also purchase or already own NovaSeq, approaching use of that system? Would it be right to say you're at ease with NovaSeq order momentum staying mostly unaffected by NextSeq, since the price per G economics on the P3 flow cell look quite appealing?
Francis deSouza
executiveYes. Thanks, Steve. I'll take the China one first, then move to NextSeq. Regarding China and what we anticipate, Q4 delivered solid growth there. Part of that reflected an easy year-over-year comparison, yet we are seeing — yet we are seeing expansion, and we understood it would be uneven. For this year, as we've stated, we anticipate growth in China. The main driver will be the clinical markets, and in particular oncology remains a very strong area for us in China, and we expect that to keep unfolding this year. On the coronavirus, let me say up front that our thoughts go out to the families and communities affected by the virus, and one of our top priorities is ensuring we do all we can to help while making certain our teams remain safe. As for the financial effect, calling it for Q1 or for the full year is premature. Once we have greater clarity as things develop, we will of course speak openly with investors. In thinking through the impact, it's possible that if the epidemic spreads and patients begin avoiding hospitals, NIPT and cancer testing in China could be affected. Still, it's far too soon to determine that impact, so we'll keep monitoring the situation. And once more, we'll share with you whatever we observe. We're putting a lot of effort into the coronavirus; we've got teams across the globe collaborating on that with the Chinese CDC teams and CDC teams worldwide; and then regarding the NextSeq launch and what our high throughput customers, particularly certain legacy HiSeq customers, are thinking about it. So like I mentioned, we truly anticipate that existing NextSeq customers will be drawn to the NextSeq 1000 and 2000, and that'll make up the bulk of the upgrade cycle. Some MiSeq customers who are upgrading will show interest, and a small amount of interest will come from current high throughput customers, though not a lot. The explanation for that is when you examine the output from NovaSeq and compare it to what NextSeq produces, we've aimed to span the range, yet overlap is minimal. So at the top end, if you use the high output flow cell on the NextSeq 2000, you can reach 300 gigabases of output; and at the bottom end of NovaSeq using the Prime Flow Cell, you can achieve 500 gigabases of output. Thus there's an obvious gap in between. And so you can — you figure out where you fit. And we believe most of our high throughput customers will choose the smaller flow cell on NovaSeq, while the NextSeq 2000 and 1000 will keep drawing that mid-throughput customer base.
Bill Quirk
analystSo real quick, when we look at how clinical oncology might grow across the coming 2 years, a number of quite innovative noninvasive screening choices, plus recurrence monitoring tests, are arriving. What's the right way to frame how this business trends overall versus 2019? Did it perhaps swing back up to — I believe it was 30%, perhaps 1 or 2 years back? Then, on a separate note regarding coronavirus, do any parallel cases exist with other outbreaks, such as MERS? If that's the case, did you examine what the effect was, let's say, in the Middle East region on your business while that outbreak occurred, and compare that against coronavirus?
Francis deSouza
executiveGreat. Thanks, Bill; so I'll begin by discussing the clinical oncology market, in particular addressing how the screening opportunity and the emerging monitoring opportunities could affect the growth rate, and after that I'll cover the coronavirus. So when it comes to clinical oncology testing, right now, most of the revenue we receive from clinical oncology and most of the growth, honestly, comes from customers who use NGS-based testing, different kinds of panels for therapy selection. And so that's roughly the majority of the revenue we're seeing in clinical oncology. That's the most penetrated portion of the clinical oncology market; yet honestly, it remains in its early days, it's still roughly 8% penetrated. When we consider the other areas within clinical oncology, we observe truly exciting opportunities and possibly far bigger opportunities arising from monitoring and screening. Each of them sits in the extremely, extremely, extremely nascent stages. Customers such as Guardant and GRAIL and Freenome are doing some truly exciting work there. The monitoring opportunity is beginning to unfold perhaps somewhat further along than the screening opportunity right now, where we're seeing the early use of liquid biopsies to monitor how effective the therapy is, to look for MRD and for a cancer's recurrence. And so there, customers such as Guardant are beginning to sell tests into that space. Once again, very early stage. We're beginning to see revenue from it, and we believe this represents a potentially large growth opportunity for us in future years. Screening is somewhat earlier, again, super nascent, but hugely exciting. I don't know whether you've had a chance to look at results that companies such as Freenome and GRAIL have been publishing over the last few months at ASCO and ESMO, but the GRAIL results, for instance, show that they can, in those studies, achieve very low false positive rates. So their numbers indicate perhaps under 0.5% false positive rates and high levels of specificity for over 20 different types of cancer. So they're still in the study stage. They're talking about having a test, maybe end of this year, some time next year. Clearly, that's a huge market opportunity when you consider the effect on people's lives, and it's a huge revenue opportunity for those companies as well as for Illumina. But that's something that won't really play out much during this year — it'll happen in years to come. If you consider the coronavirus and how things are unfolding, and we examined how MERS unfolded and how SARS unfolded — it's somewhat different, because the market is somewhat further along when it comes to the state of the art in NGS testing. The way we think this will unfold is that NGS is going to have a number of critical roles in the fight against this coronavirus outbreak. While our belief is that first-line testing of patients will be performed using RT-PCR, NGS is going to be utilized in a number of specific ways. First — and we've already watched this happen — NGS is going to be utilized to sequence the virus at the outset and to create the publicly available reference that's needed to help develop diagnostics and, in the end, cures. So we began to see that happen already with coronavirus, and we saw it before with SARS. NGS is also going to be utilized to confirm the strain in patients. It's going to be utilized for detecting novel mutations in the viral genome; we witnessed that in other outbreaks — not only MERS and SARS, but we witnessed that with Ebola. And beyond that, it's going to be utilized for epidemiological surveillance, which is going to be very important when it comes to tracking this outbreak. And then, last, we expect NGS is also going to be utilized as last resort testing for patients whose tests are inconclusive. So, for instance, when a patient shows up with suspicious symptoms, but the RT-PCR tests come back negative for coronaviruses, and also for other viruses with similar symptoms such as influenza, then we expect an NGS test is going to be utilized. And that's our perspective, based on what we witnessed with previous outbreaks. As I mentioned earlier, our teams around the globe are currently engaged and have been working — for instance, in China with the CDC teams since December — and we're working with CDC teams around the globe now. We're also engaged with philanthropic organizations in collaborations to make sure that NGS testing is available globally to monitor this outbreak.
Sam Samad
executiveBill, to highlight a few figures regarding the clinical business and oncology, returning to what you asked at the start. For 2019, as we mentioned before, our clinical sequencing consumables increased roughly 20%. As for oncology, sequencing consumables rose beyond that figure, meaning greater than 20%. Also, looking at our biggest oncology customers, their growth exceeded that average as well — surpassing the oncology average; so those are a few figures concerning growth in oncology and clinical.
Derik De Bruin
analystI have a few questions. To begin with, could you elaborate a bit on the HiSeq decommissions? The figure caught me off guard, honestly. And maybe discuss the labs that aren't utilizing it — are those clients now sending more work out? Or did they purchase NextSeq, or perhaps something different? That kind of leads into my second question, which is: I was taken aback to see that the NovaSeq consumable guidance for 2020 came in at roughly $1.1 million, $1.2 million for each box. So I'm curious whether your HiSeq buyers, with those machines declining, make me wonder why that NovaSeq figure isn't greater than what it is.
Francis deSouza
executiveOkay; thank you, Derik. I'll begin with the HiSeq decommissioning figure. That figure stems from an emphasis we've placed on drawing nearer to our clients, gaining improved insight into the way our clients utilize the machines. Consequently, we devoted time to genuinely examining the machines to determine which ones remain active and which ones don't. That represents a fresh approach for us — comprehending which machines are in use out in the field. Therefore, the overall count, 800, isn't what was decommissioned during the last year. That figure is roughly 200. The overall count is somewhat a reconciliation to say, fine, among all the machines we have out in the field, how are they being utilized? Which ones see heavy use? Which ones see lighter use? And what we arrived at is roughly 800 that we believe are mostly inactive. It alters nothing regarding the revenue we receive from HiSeqs or consumables. What it does signify is that fewer active machines exist in the field, and the ones our clients are using are actually being utilized more than the earlier model had projected, and that the pull-through for each machine at those labs is in fact greater than the old model had projected. So those clients aren't doing anything differently. They aren't sending more work out. They still possess both the active and inactive machines. But when we examined it, they were utilizing certain machines more; and so that's how the dynamic unfolds.
Sam Samad
executiveAddressing your follow-up, Derik, about NovaSeq: indeed, for 2020 our guidance stands at $1.1 million to $1.2 million in pull-through for each instrument. Our 2019 results made us quite happy, particularly what we observed regarding NovaSeq pull-through. As stated earlier, that figure reached $1.2 million across the year. Q4 marked a record quarter. Now, entering 2020 — the instrument's fourth year on the market — we expect that as some lower-throughput clients begin converting and transitioning, plus even newer buyers employing NovaSeq alongside the S Prime and S1 flow cell, reduced throughput from such customers will begin dragging down the average. That pattern is typical once an instrument reaches its fourth launch year. Even so, it accounts for why our pull-through guidance sits between $1.1 million and $1.2 million.
Francis deSouza
executiveBeyond that — Sam mentioned a number of the smaller clients that should be starting up during this year, and that's part of why we provided the guidance we did. An opposing effect is going to show up as several of those big PopGen accounts keep sequencing, such as the U.K. Biobank, which is truly operating — truly operating at maximum capacity — along with NHS commissioning getting underway. So that's going to be the counterbalancing element that unfolds over this year.
Dan Arias
analystSorry, I've got a bit of a cold. Francis, regarding NextSeq, could you explain why you're launching the 1000 unit, especially when we think back to Nova 6000 and 5000? At that time, it appeared that buyers who upgraded chose one of the larger capacity choices and — so they wouldn't have to wait to get it. So I suppose I'm just — I want to know how you think this time will be different. And then, perhaps on a related note, is there anything more precise you can share about the P3 Flow Cell launch coming later this year, in terms of timing? And I'm wondering how much demand for that configuration we should anticipate being captured in 2020.
Francis deSouza
executiveSure. And I hope you feel better soon. I'll take the — I'll take the NextSeq question first. So why do the 1000 when we anticipate most customers will purchase the 2000, and after that I'll discuss P3; so the approach of offering a higher-end and a lower-end instrument is one we've employed many, many times; so you mentioned the NovaSeq 5000, 6000; we have the HiSeq 3000, 4000; we have the HiSeq 2000, 2500; and what that enables our sales teams to do is it lets us begin the conversation with customers at a lower price point; and some customers do purchase the lower instrument. But as you noted, what occurs during that conversation is the vast majority of cases may ultimately purchase the higher-priced instrument. And that happened in every one of those instances I mentioned going back a decade, correct? So — but it's essential for our reps to have that lower-priced instruments so they can begin the conversation. And that's been the approach. As for the P3 flow cell. The way we anticipate this unfolding is that customers for the first couple of full quarters of NextSeq 2000 are going to be thrilled about the fact that they can purchase a flow cell whose output is comparable to today's NextSeq. So they can make — they can evaluate the new machine on an equivalent output using equivalent workflows from the old machine to the new machine. But we do anticipate that a lot of our NextSeq customers are going to purchase the instrument with an eye to their future business and will be very drawn to the higher output and the economics tied to the P3. On timing, we said we anticipate that to arrive towards the end of the year. Obviously, though, it's public now and the specs of that flow cell are public now; so it's something customers are aware of and are planning for.
Doug Schenkel
analystSo a few points about NovaSeq. To start, when we look backward at the guidance for consumable pull-through of $1.1 million to $1.2 million per box for Nova, how does that stack up against 2019? And I suppose, does the drop in HiSeq consumable revenue under $100 million, which is somewhat lower than where I believe you've been trending across the last few quarters, suggest that bigger programs on HiSeq Xs are mostly finished now and that some of those large customers with such projects are increasingly shifting to Novas?
Sam Samad
executiveAlright, Doug, thanks for those questions. Let me handle the backward-looking items first. Then, turning to what's ahead: are you assuming fewer NovaSeq instrument installations in 2020 compared with 2019? Is that because you're trying to guard against the chance that Nova will eat into some of that business, because there are fewer HiSeqs in the field that will eventually need to be swapped out, so maybe you have a bit more clarity on the replacement picture, because you're out in front of the launch — or is it some mix of all those factors? And finally, is there any structural barrier that would keep you from bringing the blue/green chemistry to a higher-end system like NovaSeq, or even to NovaSeq at some later point? I'll take the first question, then let Francis handle the second and third. On NovaSeq: for all of 2019, we told you the pull-through came in at $1.2 million for each instrument. Going forward, our expectation for 2020 is somewhere between $1.1 million and $1.2 million. Essentially, that upper figure lines up with the 2019 result. As for what you asked about HiSeq, I'd say you've got it exactly right. As customers move further along in shifting from HiSeq over to NovaSeq, and as that transition work wraps up on the older platform and migrates to the newer instrument, it's only natural that HiSeq consumables decline. That's exactly what we've been saying: during 2020, we expect HiSeq consumables to keep falling, while NovaSeq consumables clearly go up.
Francis deSouza
executiveThen, Doug, you raised a few other points. First, regarding NovaSeq placements, why do we anticipate a decline this year after 2 consecutive years of exceptionally robust NovaSeq placements? Is it, as you wondered, being affected by NextSeq and some cannibalization happening there? You also asked whether there's any structural reason the blue/green chemistry could not someday be applied to NovaSeq. So on NovaSeq placements, when we examined this year, the effort we foresee for this year regarding the NovaSeq upgrade cycle is to truly target the smaller core labs, correct? That upgrade cycle was set in motion last year with S Prime, with S1, and then S2's new pricing. Consequently, looking at the pipeline for this year, we anticipate it will be far more filled with onesie-twosies among the small core labs; that's the dynamic we expect to unfold across our customer base and our sales teams; therefore the total will end up smaller than some of those large multi-unit purchases seen last year in Q3 and Q4 from customers such as the U.K. Biobank, for instance. Rather than 10 unit placements, we anticipate more deals involving 1 to 2 units. And honestly, that's the primary factor driving the placement figure lower this year versus last year. We don't anticipate significant cannibalization between NextSeq and NovaSeq; they serve quite distinct price points, both for capital and for cost per G. And there's no overlap when it comes to output. So once you combine those 3 variables, customers can self-select quite easily. If your sample volume supports buying a NovaSeq, you'll always choose that since the economics are superior. Even with the new NextSeq 2000 running the P3 flow cell, the NovaSeq still delivers superior economics provided you have the sample volume. And if your sample volume isn't sufficient, it will never make financial sense for you to move to NovaSeq; thus there's minimal overlap in the segment they target; so some overlap may exist, but we don't anticipate much. Then your question regarding a structural reason we haven't employed the blue/green chemistry in NovaSeq, or frankly anywhere. The answer is there isn't one. We regard that as a core architectural component in our toolkit now, just as we do other elements you saw in NovaSeq. For instance, the hardware acceleration toolkit we incorporated into NextSeq 1000 and 2000 based on the Edico technology — that's yet another core architectural component. Nothing prevents that from appearing in any of the instruments we release later, whether higher-end or lower-end across the lineup. Honestly, the factor that determines whether it appears is what design point we choose for the following instrument we launch, and whether our toolkit holds something superior. That is, more or less, an ongoing internal discussion. So whenever a given moment arrives, we'll select the finest components available to us, along with the design point we're pursuing. Yet as you note, there is simply no structural barrier that would stop those components from turning up in a later edition of a high-throughput instrument or in one of our lower-throughput instruments.
Puneet Souda
analystMy first question, then — I was hoping you could give me a clearer picture of the NHS clinical ramp and how much visibility you actually have into it. You mentioned clinical samples picking up in the back half of this year. This ranks among the biggest projects, maybe the biggest of all; and I appreciate that you were targeting 300,000 to 500,000 patient sequences by 2025; but how does that connect to the 5 million potential genomes the NHS mapped out previously? And what ought we to anticipate in the second half of this year? Then on NextSeq, I fully understand what you said about 2000 adoption, with ArcherDx indicating the same pattern among economic customers and smaller labs; but I'd just like to grasp, when it comes to those who own diagnostic instruments, or instruments sitting in a diagnostic setting — NextSeqs and MiSeqs — what do you expect from those customers regarding upgrading, now that there's new chemistry, considering these instruments are validated, and a number of them run NextSeqDx? Any thoughts about NextSeq 2000Dx?
Francis deSouza
executiveSure. Thanks, Puneet; so that's 3 questions. First, an update on the NHS ramp and what it looks like for the back half of the year, and then a comment on how to connect the fact that we've discussed 300,000 to 500,000 samples with Secretary Hancock speaking about 5 million genomes being completed, so how to connect those. You also raised NextSeq and that upgrade cycle, particularly the Dx instruments and whether Dx customers for NextSeq will form part of this upgrade cycle. So regarding the NHS, frankly, we're pleased with the progress. Over the past few months, we provided the update announced at JPMorgan that a contract has been signed with GeL to supply the lab testing services, the genomic lab testing services, for the NHS. Our teams are deeply involved; we brought a leadership group here, here in California, to go over plans moving forward. So right now, we feel quite good about how things are tracking regarding the ramp for the NHS, and we anticipate seeing that roughly midyear onward. The first phase is 300,000 scaling up to 500,000 genomes to be sequenced. And a bottoms-up analysis has been done with that team on how you reach that; and what we've stated publicly is that genomic testing will become a standard of care for a little more than 20 genetic diseases and 4 different types of cancers, beginning in the middle of the year when we ramp up with the NHS. So standard of care for the U.K. population, 55 million, 60 million people; but clearly, there are far more than 20 genetic diseases and far more than 4 cancers. And so Secretary Hancock is really pushing the NHS to be more ambitious about how we want to roll this out for the entire U.K. population. No — it isn't 20 genetic conditions we're talking about; the count is 6,000 genetic diseases. And cancers covered number far beyond 4. So once you begin adding up the effect of spreading this across oncology indications, plus additional genetic conditions and other fields where genomic testing can deliver improved results at reduced expense, that is precisely how you arrive at a figure such as 5 million. There's no committed roadmap or committed timeline yet, but that is what we're aiming for. So that covers the NHS ramp. Now, regarding NextSeq and the upgrade pathway, particularly for Dx. Our expectation for how the Dx market unfolds is this: if you're a NextSeqDx customer at present, or you're in the queue to purchase a NextSeqDx, our expectation is that you'll either continue with the unit you already own or you'll acquire a fresh NextSeqDx. And chances are greater that you won't be buying the 2000 or the 1000 at this moment. That connects to the point you raised: look, you've got validated workflows; frequently, you might even have a distinct assay constructed on our cleared box. Consequently, you'll definitely be watching for the moment the 1000, 2000 Dx instrument is announced and reaches the market. However, for the foreseeable future, whatever capacity you add will come from Dx boxes, NextSeqDx boxes. Therefore, we don't anticipate those customers fueling the pipeline for NextSeq 1000 and 2000 at this time.
Patrick Donnelly
analystSam, regarding the linearity in your guidance that you mentioned, could you give us a little more detail? In the first quarter, it clearly landed quite a bit under what the Street had anticipated—analysts were only expecting revenue growth of 1%; on the EPS side, there was a fairly significant drop compared to a year ago. So could you walk us through that further? And beyond that, when it comes to the guidance, I'd like to hear about the full-year visibility on certain POPSEQ revenues, plus the timing of some initiatives that haven't kicked off yet. As we all know, in 2019 there were some delays. Perhaps you could address how confident you are in those figures, especially the 200,000 from U.K. Biobank. What's the potential for that number to move, whether higher or lower?
Sam Samad
executiveCertainly. Thanks for asking, Patrick. Regarding the Q1 figure—first of all, when we examine Q1 specifically, and really all of 2020 as well, this pattern resembles other years in which we introduced new instruments. To take you back to 2017—I realize NovaSeq is a different instrument—the linearity closely mirrors what occurred in 2017. And thinking about Q1, if you're making a comparison with Q4, there are clearly 2 or 3 major elements behind the drop from Q4 to Q1. The first is seasonality combined with fewer instrument placements, particularly NovaSeq, in Q1 compared with Q4, along with seasonality affecting the majority of the business, including sequencing consumables. Then there's something we've witnessed previously: Q1 tends to see a significant jump thanks to the DTC business and the processing of samples that return following the holidays—well, that isn't happening this year, since we obviously have a far weaker DTC performance for both Q1 and the full year. And lastly, one more important element is that, with NextSeq 2000 launching, NextSeq shipments will be limited in Q1 as customers hold off, waiting for the new instrument to arrive; that new instrument won't be available until late in Q1. That means far fewer NextSeq shipments in total during Q1. Those are the main elements behind the quarter-over-quarter drop from Q4 to Q1. Looking at the remainder of the year, clearly, you'll see the NextSeq 2000 launch continuing to build momentum. NovaSeq instrument placements will increase throughout the year. You'll also have the population genomics opportunities we discussed. So U.K. Biobank has been operating at full production, and we'll keep processing those 200,000 samples we mentioned throughout the entire year. On top of that, there are the expectations we laid out at JPMorgan concerning All of Us, which begins in the middle of the year. We're beginning our commissioning work with the NHS and GeL around the midpoint of this year. As a result, we clearly have confidence in the ramp we anticipate throughout 2020; however, Q1 faces constraints for the reasons I gave earlier. Clearly, the NextSeq 2000 launch is a major factor in that. Still, this situation resembles other launch years we've experienced previously.
Dan Leonard
analystI just want to go back to the POPSEQ guidance for a moment. Sam or Francis, how much detail are you actually giving us here? In other words, the three programs you called out — U.K. Biobank, All of Us, and NHS — are those the three you've received approval to say something about, with possibly other significant ones emerging during 2020? Or are those truly the three that will move the numbers in 2020?
Sam Samad
executiveI can begin, and Francis may want to add something. This isn't a question of what we're allowed to reveal. It's that we want to stay focused on the main PopGen projects, and those three are the main PopGen projects we discussed a year ago and wanted to discuss now, giving very specific visibility into each. So U.K. Biobank, already underway, and we'll keep processing everything throughout 2020; All of Us, a major topic in 2019, and we wanted to clarify when we think it will begin and how many samples go with it; and clearly the NHS commissioning project, which also came up last year. We're now happy to report that we have this agreement with GeL and the NHS, and that will get started as well. U.K. Biobank has gotten started too. Other population genomics initiatives sit inside our 2020 financials. We won't be addressing those specifically. They're part of our business, a few began in 2019, a few should begin in 2020. But our real aim was to concentrate on the three key material ones.
Jack Meehan
analystJust perhaps on that point. After the PacBio potential deal fell through, I was hoping you might update us on Illumina's approach to long-read sequencing. It sounds as though Illumina will have to find its own path into that space through innovation; so what does the plan look like going forward? Where does this rank as an R&D priority? And could you offer some reassurance about how it stacks up against your other efforts, such as clinical?
Francis deSouza
executiveYes. Certainly. I'll take that one, Jack; we still hold the view that long-read will remain an adjacent segment within sequencing overall, making up roughly 5% of that total market. And as far as we can see ahead, the two markets will overlap only minimally, frankly. Our major markets, be it oncology, testing or NIPT, have no real requirement for long-read capability, while the major long-read markets, de novo sequencing, new species sequencing, genuinely require that capability and therefore will be served by a technology suited to them. So our view remains that this will be a complementary space alongside short-read, accounting for roughly 5% of sequencing overall. What got us enthusiastic about the PacBio deal was that our engineering groups excel at pushing the price, the cost of an instrument, downward. Our engineering and operations groups are outstanding, and so are our technologists. And our aim was to attempt to speed up that market since we recognize there's a knock-on effect whereby what benefits that market benefits ours. Because after you sequence a new species, say, you carry out the bulk sequencing for that species using short-read technology. And so we'll keep seeking ways to advance that market faster than its current pace, though it's a minor portion of sequencing overall; it's tiny next to the clinical market opportunity, which is far, far bigger than what long-read offers; so regarding internal trade-offs, there's simply no comparison when we consider allocating more spending to advance the oncology testing market, say, against the long-range market. That said, we've stated publicly that we have our own long-read programs underway internally. Those will keep getting work from us internally. But by a wide margin, our greater priority remains the short-read core markets, encompassing both the research markets and the clinical markets.
Jacquie Ross
executiveThanks. Just a quick note — you'll be able to hear this call again through the webcast posted in the Investors part of our site. We appreciate you being with us today. That wraps things up for now, and we're eager to speak with you again once the first fiscal quarter has ended.
Operator
operatorEveryone, that brings us to the end of today's call. We appreciate your participation. You're free to hang up now.
Read the full transcript via the API
You're viewing the first half of this call. Get the complete Illumina, Inc. transcript — plus 132,000+ transcripts from 7,000+ companies, speaker segments, AI summaries and full-text search — through the EarningsCalls.dev API.
Get the API View API docs →This call discussed
For developers and AI pipelines
Programmatic access to Illumina, Inc. earnings transcripts and 132,000+ others is available through the
EarningsCalls.dev REST API. Plans from $39.99/month — full transcripts, speaker segments,
full-text search, and the recently-added /api/v1/transcripts/recent polling endpoint for ETL pipelines.
More tools for stock research
OptionsBell: an unusual options activity scanner that tracks options flow across all US stocks and emails you when large trades appear in Illumina, Inc. or any ticker on your watchlist.
StockMarketScan: a stock screener with alerts: 24 pre-built screens refreshed every 5 minutes, watchlists and email notifications. Both offer API and MCP access and fair monthly pricing.