Illumina Inc (ILMN) 2017 Q4 法說會逐字稿

Hello, and welcome to the Fourth Quarter and Full Year 2017 Illumina Earnings Conference Call.

My name is Michelle, and I will be your operator for today's conference.

(Operator Instructions) Please note that this conference is being recorded.

I would now turn the call over to Ms. Jacquie Ross.

Ma'am, you may begin.

Thank you, Michelle.

Good afternoon, everyone, and welcome to our earnings call for the fourth quarter and full year 2017.

During the call today, we will review the financial results released after the close of the market and offer commentary on our commercial activity, after which we will host a question-and-answer session.

If you have not had a chance to review the earnings release, it can be found in the Investor Relations section of our website at illumina.com.

Participating for Illumina today will be Francis deSouza, President and Chief Executive Officer; and Sam Samad, Chief Financial Officer.

Francis will provide a brief update on the state of our business, and Sam will review our financial results.

This call is being recorded and will be archived in the Investors section of our website.

It is our intent that all forward-looking statements regarding our financial result and commercial activity made during today's call will be protected under the Private Securities Litigation Reform Act of 1995.

Forward-looking statements are subject to risks and uncertainties.

Actual events or results may differ materially from those projected or discussed.

All forward-looking statements are based upon current available information, and Illumina assumes no obligation to update these statements.

To better understand the risks and uncertainties that could cause actual result to differ, we refer you to the documents that Illumina files with the Securities and Exchange Commission, including Illumina's most recent Forms 10-Q and 10-K.

With that, I will now turn the call over to Francis.

Thank you, Jacquie.

Good afternoon, everyone.

Illumina finished 2017 on a high note with revenue of $778 million, growth across both our sequencing and arrays businesses and notable strength in sequencing consumables.

For the full year, revenue grew 15% to $2.75 billion.

Sequencing consumables grew by more than 30% year-over-year to $432 million with contributions from each of our product families.

The primary contributor to the more than $100 million year-over-year increase was NovaSeq, followed by NextSeq and the HiSeq family and then our other desktop platforms.

NovaSeq consumable revenue almost tripled sequentially, with strong demand for the newly released S4 flow cell in addition to growing interest in S2.

Even excluding a $19 million stocking order, NovaSeq consumables grew by more than 75% sequentially.

Fourth quarter HiSeq X consumables were up modestly from Q3.

Most of our HiSeq X customers have taken at least 1 NovaSeq.

And now that S4 has been launched, some of these customers will transition their fleets in 2018 while others will move in 2019 or later, depending on the cadence of their clinical, translational and research activities.

Excluding HiSeq X, HiSeq consumables were up, driven by clinical customers.

Similar to HiSeq X, transition plans among the approximately 800 customers are staggered across a multiyear horizon.

That said, we continue to expect HiSeq consumables in aggregate to start to trend downwards in the coming quarters.

NextSeq consumables had another strong quarter, and average pull-through was once again above the high end of our target range, driven by production clinical customers.

Nextera DNA Flex got off to a strong start in October as growing appreciation of the simplified workflow led to orders from more than 300 unique customers.

Since the launch of AmpliSeq for Illumina a few weeks ago, I'm pleased to share that we have received and shipped our first orders for the latest addition to our library prep portfolio.

We also reported a strong sequencing systems quarter with revenue of $131 million.

NovaSeq shipments were in the high 80s, the most we shipped during any quarter in 2017, reflecting customer readiness for their systems.

As expected and consistent with the third quarter, we did not ship any HiSeq Xs in the fourth quarter, and we shipped just a handful of HiSeqs as most high-throughput customers are now looking to NovaSeq.

With approximately 285 NovaSeqs shipped in 2017, we are very encouraged by the system's performance in its first year of launch.

Interest to date has included an impressive cross-section of HiSeq X customers, HiSeq customers, new customers as well as customers moving up from desktop instruments to take on higher-output applications or increase sample volumes.

NovaSeq is demonstrating a very broad appeal.

For the third consecutive quarter, about 1/3 of NovaSeqs orders received were from new-to-Illumina or straight-from-benchtop customers.

New-to-Illumina customers included a number of academic and translational labs looking to build out their offerings for clinical testing, whole-exome sequencing and whole-genome sequencing.

Straight-from-benchtop customers included a diverse set of labs building on their MiSeq and NextSeq capabilities to extend into exomes, methylation arrays, RNA, ctDNA and whole-genome sequencing.

Beyond this group, the HiSeq conversion opportunity remains very much ahead of us, with only 15% of the approximately 850 HiSeq or HiSeq X customers having ordered their first NovaSeq as of the end of 2017.

NovaSeq represents a compelling opportunity for these customers, most of whom will recognize a significant economic benefit from the transition.

Beyond the price-per-sample economics, NovaSeq offers higher throughput, easier workflow and the ability to easily scale output to support a diverse suite of applications and sample volumes.

So we expect that the vast majority of these HiSeq customers will transition to NovaSeq over time.

NextSeq continues to perform well, and the system and associated consumables remain a very important growth driver for Illumina as customers use the system for a broad set of research, translational and clinical applications.

Rounding out our benchtop systems, we continue to see stable win rates and solid adoption by new-to-sequencing customers who represented approximately 60% of MiniSeq and MiSeq shipments in the fourth quarter.

We look forward to shipping the first commercial units of our latest desktop system, the iSeq, towards the end of the first quarter.

iSeq delivers exceptional accuracy at a low capital cost of less than $20,000, making our SBS technology accessible to any lab.

You will have seen that the beta units are operating very well in the field, and we have already received customer orders reflecting strong initial demand.

As is routine for a product launch at Illumina, we will be shipping a small number of iSeqs this quarter to be followed by a manufacturing ramp through 2018.

That said, we do not plan on sharing order or shipment details on the iSeq on a quarterly basis.

While this is an important strategic platform for Illumina, it is unlikely to have any material revenue driver in the near term.

In other product updates, we look forward to adding the S1 flow cell to the NovaSeq portfolio to further extend the flexibility for our lower-throughput customers and those looking for rapid turnaround on higher-output applications.

We've completed internal validation and started beta testing of S1 in anticipation of commercial release later this quarter.

S1 offers very compelling per-sample economics versus HiSeq and the ability for HiSeq customers to easily transition projects due to its lower output.

S1 also delivers the fastest run times of all NovaSeq configurations, delivering 500G runs in just over 24 hours.

Reflecting our investments and focus, our clinical business continued to grow in the fourth quarter, notably in oncology and NIPT.

In total, clinical and translational customers represented about 45% of our shipments in 2017 compared to 39% in 2016, and we expect incremental adoption of our products and technology in the coming years, given the importance of genomics in solving human health challenges.

As we have shared previously, our strategy to further catalyze clinical adoption is to provide clinical-grade instruments and reagents to customers who wish to develop their own solutions while developing sample-to-answer solutions in areas of the market where we are uniquely positioned.

As part of this, we launched NextSeqDx in mid-November to bolster our clinical instrument offerings, and we shipped our first systems in the fourth quarter.

These initial customers plan to use their systems for routine non-small cell lung cancer liquid biopsy testing and clinical assay development.

Oncology is the largest and fastest-growing area within clinical, driven primarily by customers focused on therapy selection and liquid biopsy.

In total, our oncology testing shipments grew about 40% in 2017.

Development work on our sample-to-answer solutions progressed in the fourth quarter, with TST 170 now available as an IUO as well as RUO.

Building on our TST 170 panel, we have launched an immuno-oncology program as we believe that Illumina is uniquely positioned to deliver a distributable IVD solution, offering a turnkey solution to pharma partners and customers.

We continue to see positive coverage decisions in the field of oncology.

Most recently, Cigna announced a significant policy change, extending previous coverage for panels of less than 10 genes with proven clinical utility to larger panels where the medical necessity criteria is met for at least 1 gene on that panel.

We believe that this will be particularly helpful for clinical trial enrollment, increasing the utility and utilization of multi-gene panels over time.

NIPT is another focus area for Illumina, and we are very pleased with the performance of VeriSeq NIPT, which received CE Mark in April 2017 and is our first true sample-to-answer IVD product.

Sales are expected to more than double in 2018, driven in part by the Dutch National contract, which we were awarded during the fourth quarter and some other key competitive wins.

We will submit NIPT IVD solutions in many more countries throughout 2018 and look forward to the continued adoption of this product globally.

Our microarray revenue grew 21% from the same quarter in 2016 due to strong growth among our direct-to-consumer customers.

Full year growth was 20%, driven by the acceleration in the number of consumer samples processed during 2017.

Our consumer affiliate, Helix, announced a partnership with a Healthy Nevada Project to expand one of the first community-based population health studies in the United States.

Led by the renowned Institute for Health Innovation, the project plans to enroll 40,000 Nevadans and utilize the Helix platform to collect and sequence the samples.

The Healthy Nevada Project started in 2016 with an initial 10,000 participants who submitted samples for genotyping analysis.

Phase II of the project has expanded to exome-plus analysis in an effort to gain additional health determinants data.

Before I hand over to Sam, I'd like to formally welcome Aimee Hoyt, who joined Illumina to lead our Human Resources team earlier this month.

We believe that our corporate culture is one of our strongest competitive differentiators, and we've invested a good amount of time in finding our Chief People Officer.

I am confident that Aimee is the right person to evolve our unique culture and our commitment to innovative people practices.

With that, I'll hand the call over to Sam for a review of our financials.

Sam?

Thanks, Francis.

As discussed, fourth quarter revenue grew 26% year-over-year to $778 million, driven by growth in sequencing consumables that once again exceeded forecast as well as strong performance in sequencing instruments and our microarrays portfolio.

Geographically, Americas grew revenue 32% versus the prior year period, driven by strong growth in both sequencing instruments and consumables, while EMEA grew 25%.

Asia Pacific grew 9% overall, with 12% shipment growth in Greater China.

Revenue from sequencing instruments grew 18% year-over-year to $131 million, driven primarily by NovaSeq and offset in part by muted shipments of the HiSeq family of instruments as expected.

Microarray instrument revenue was $8 million, down as expected from an unusually strong Q3.

Total instrument revenue was, therefore, $139 million, an increase of 22% year-over-year and represented 18% of total revenue.

As Francis noted, fourth quarter sequencing consumable revenue was $432 million, up more than 30% from last year.

Microarray consumable revenue of $82 million was up 8%.

Total consumable revenue of $514 million represented 66% of total revenue.

Fourth quarter total product revenue, which includes freight, was $659 million, up 26% from the year-ago quarter.

Service and other was $119 million, up 27% from the same quarter last year, driven by strength in genotyping services due to consumer demand as well as sequencing services and instrument maintenance contracts.

Moving to gross margin and operating expenses.

I'll highlight our non-GAAP results, which include stock-based compensation.

I encourage you to review the GAAP reconciliation of non-GAAP measures, which can be found in today's earnings release and supplementary data available on our website.

Please note that all subsequent references to net income and earnings per share refer to the results attributable to Illumina shareholders.

Non-GAAP gross margin of 70.9% improved 210 basis points from last quarter due to a more favorable sequencing consumables mix as well as improved cost savings from NovaSeq.

Year-over-year gross margin increased 140 basis points and was impacted by higher sequencing consumables mix and a favorable impact associated with the lack of product transition reserves that occurred in Q4 of 2016.

Non-GAAP operating expenses were $307 million, up $8 million from last quarter with higher stock-based compensation expense partially offset by lower Helix expenses.

Non-GAAP operating margin was 31.4%, up 460 basis points sequentially and higher than the 25.1% reported in the fourth quarter of last year.

Excluding Helix, operating margin was 33.9% compared to 29.7% in the third quarter.

For the fourth quarter, GAAP net income was $68 million or $0.46 per diluted share.

Non-GAAP net income was $212 million or $1.44 per diluted share, and Helix dilution was $0.06.

The non-GAAP tax rate was lower sequentially due to a higher foreign mix of earnings.

Cash flow from operations equaled $294 million.

Continued reductions in our collection cycle led to Q4 DSO of 48 days compared to 49 days last quarter.

Capital expenditures in Q4 were $76 million, and Q4 free cash flow was $218 million.

We ended the quarter with approximately $2.1 billion in cash and short-term investments.

During the quarter, we repurchased approximately 368,000 shares under our previously announced buyback program at an average price of $204.

Moving to guidance and as we previously shared, total company revenue is expected to grow between 13% and 14% in 2018, including a modest revenue contribution from Helix.

Mix remains the primary driver of our gross margin, and we therefore currently expect full year non-GAAP gross margin to be up slightly from 2017.

GAAP EPS is expected to be between $4.14 and $4.24, and non-GAAP EPS is expected to be between $4.50 and $4.60.

This includes an approximately $0.10 benefit associated with tax reform and about $0.25 of Helix dilution.

Some other items for you to keep in mind.

First, and as Francis discussed, the activity we are seeing among our customers is consistent with our belief that the NovaSeq upgrade cycle will span multiple years.

Going forward, we will report NovaSeq shipments on an annual basis in an effort to minimize volatility associated with normal quarterly fluctuation.

For 2018 and compared to the approximately 285 NovaSeq shipments reported in 2017, we expect to ship between 330 and 350 NovaSeq systems.

With regards to Q1 specifically, I'll remind you that we do not expect the $19 million stocking order to repeat.

Additionally, you should expect sequencing system revenue to be down sequentially, in line with normal first quarter seasonality.

As a result and after an exceptional fourth quarter, we expect first quarter revenue to be up to $35 million lower on a sequential basis.

With that, I'll hand the call back to Francis for a few comments before we open the Q&A session.

Thanks, Sam.

Looking forward, I'm excited about Illumina's opportunity in 2018 and beyond.

As I discussed at JPMorgan a few weeks back, we're seeing the regulatory and reimbursement decisions that will ultimately enable sequencing to move firmly into the clinic, where it will positively impact patient outcomes.

In many cases, these changes are emerging faster than we could have expected, which is a positive indicator for Illumina's growth in the years to come.

For 2018 specifically, there are many factors that will contribute to Illumina's continued growth.

First, our NovaSeq system that remains in the earlier stages of the upgrade cycle, and as we expected, looks set for a multiyear transition.

Notable within this category, Xp workflow and S4 were launched late in 2017, so we will benefit from a full year of release in 2018.

Additionally, we have the launch of S1 this quarter, so we expect NovaSeq consumables to continue to grow rapidly in 2018.

Beyond NovaSeq, we have an impressive portfolio of products that are new or just launched, including iSeq, Nextera DNA Flex and AmpliSeq for Illumina.

In population genomics, we have a number of national initiatives that are ramping in 2018.

And of course, our consumer business continues to grow through our array customers, in addition to a modest contribution from Helix, still only in its first few quarters of launch.

In clinical, 2018 will represent our first full year of sales of VeriSeq NIPT CE-IVD and the ramp-up of some new large customers.

NextSeq is proving to be a workhorse for our clinical customers, so we expect to continue strong -- to see strong pull-through performance and consumable growth.

We expect to see continued and potentially accelerating traction in RUGD, given the expanded insurance coverage for whole-exome sequencing achieved in 2017.

Finally, we expect to see continued growth among our oncology customers as they continue their important and potentially transformative work.

I look forward to updating you on our progress throughout 2018.

2017 was a terrific year, and I'd like to thank the Illumina team for the commitment and execution that delivered such a strong year.

Operator, we'd now like to begin our Q&A session.

(Operator Instructions) The first question in the queue comes from Doug Schenkel from Cowen.

You closed out 2017, clearly, with a lot of momentum in addition to having the NovaSeq out there without manufacturing constraints, you now have the S4 and the single-lane loading device out in the field, and as you just mentioned, the S1 coming soon.

Essentially, you have all the tools at your disposal out there to drive placements and utilization, and probably visibility is improving.

With that said, for those of us outside the company, there are still lingering questions on how replacing HiSeqs with NovaSeqs is going to impact overall consumable revenue.

Based on what you're seeing in the field about a year into this launch, on a dollar basis, do you expect same-store consumable sales to increase at labs that replace HiSeqs with NovaSeqs?

Essentially, based on what you're seeing out there with placements and replacements, how confident are you today that volume elasticity will be enough to offset pricing reductions afforded to these customers?

Sure.

So thanks for that question, Doug.

Even as we were planning for the launch of NovaSeq, we were spending a lot of time with our HiSeq customers and our HiSeq X customers.

And as we do with all major, especially high-throughput launches, we spend a lot of time understanding the drivers of demand for the new platforms.

And we released new platforms when we feel that we are about to tap into sort of the next wave of elasticity.

And so as we got into the NovaSeq wave, what we heard from customers very clearly was that with the power that they would have access to with NovaSeq -- and remember, for a lot of these customers who are HiSeq customers that haven't bought the Xs, so the 800 HiSeq customers that were outside of the 35-plus X customers, they really hadn't had power -- the access to the power of the X. What they were telling us is when they got access to NovaSeq, what they would do is they'd be able to drive more sequencing in a number of ways.

They'd look at doing larger cohorts.

They'd look at doing broader sequencing, so panels to exomes, exomes to genomes.

And they'd look at doing deeper sequencing.

And so that's the -- that's what built the model around NovaSeq and the demand for NovaSeq.

And as we look back sort of a year from having launched NovaSeq, it's playing out as we expected.

As you heard me talk about at JPMorgan, we're now seeing examples of customers that are tapping into this elasticity in each of those buckets.

We're seeing customers doing larger cohorts than they would have planned.

We're seeing customers doing much deeper sequencing, especially in areas like liquid biopsy but other areas of oncology, too.

And then we're seeing customers doing broader sequencing.

If you look at projects, for example, like the consortium with Regeneron announced, in terms of sequencing, the U.K. biobank, that's the kind of project that really is made available through NovaSeq that, frankly, wouldn't have happened on the HiSeq.

And so in each of those categories now, we have examples that NovaSeq is delivering on the elasticity of that market.

Okay.

That's helpful.

And this one's a bit quicker.

Based on your installed base disclosures the last couple of years, it looks like you maybe retired only about 100 HiSeq and HiSeq Xs in 2017 or at least something in that ballpark.

This is far less than what you retired the year before.

Again, assuming my math is right here, why did the pace of retirements moderate in 2017?

And maybe more importantly, looking ahead to 2018, what is your expectation for HiSeq and HiSeq X replacements that you embedded into full year guidance?

Sure.

So in Q4 itself, we had another 20 -- about 20, sorry, HiSeqs that we retired from the installed base, and we didn't actually take out any Xs in the installed base from Q4.

So I think that adds to the numbers you heard from the rest of the year.

As we were thinking about the upgrade cycle and the pace of retirements and transitions from the HiSeqs and the Xs, this really is a 3- to 5-year upgrade cycle.

And the way we sort of measured it out is we said in the first year, so 2017, the focus was really on the HiSeq customers, not the X customers.

And the focus on the HiSeq customer base was to get them comfortable -- in the higher utilization HiSeq customers get them comfortable with the NovaSeqs, get them starting to validate their workflows on the NovaSeqs.

And so the way it showed up in demand for us in 2017 was, certainly, I talked about it in Q3, especially we saw lots of orders that were onesie-twosies, and those were driven by some of the higher-throughput shops, looking at NovaSeq and starting to get comfortable with the workflow in NovaSeq as well as in some of the smaller labs that were looking to upgrade.

We expect the bigger part of the transition to happen starting this year, driven by 2 things.

One, some of those HiSeq customers are going to start to move more of their workload onto NovaSeqs.

And so you'll start to see more instruments being retired as we go through 2018.

And then as you know, we launched the S4 in Q4.

And so the intent was always 2018 was the year for X customers to start their transition to NovaSeq, and so you should start to see Xs start to be taken out of commission this year going forward.

And so that's what led to the lower retirements -- part contributor to the lower retirements of those instruments in 2017.

The next question in the queue comes from Ross Muken from Evercore.

As we think about your conversation with your customer base around the NovaSeq from an application perspective, it seems like the size and the scope of the types of studies one would want to do and sort of even things that a few years ago were kind of envisioned as impossible are being made possible by the ramp in the technology.

Sort of now that we've had some time where this has been introduced and with the S4 out in particular, what are the types of discussions you're seeing globally in terms of some of these newer projects?

And I would assume some of the things coming to market now were probably derived 12, 24 months ago, and so we will probably see some of the newer studies or new scope studies come over the next few years.

So just help us understand how that has changed and how folks after they've had time to understand what's possible and what they can do with it, particularly with the S4, how they're planning.

Yes.

So let me start by saying, look, some of the things you heard about last year, the -- I talked about Regeneron, but the work that GRAIL is doing, for example, those size of studies have been fundamentally enabled by having access to the power of NovaSeq.

Now if I look back at some of the new customers and new applications, and I think you touched on globally that are being enabled.

Let me give you some examples, right?

So we saw, for example, a new service lab customer in Eastern Europe that purchased a NovaSeq and is planning to establish whole-genome, whole-exome pipeline for genetic disease testing.

We saw a large Canadian university that's going to use NovaSeq for population genetics and repatriating clinical exomes.

We saw a benchtop customer that's working to develop ctDNA technology used for early cancer screening, and they moved to NovaSeq considering the sequencing costs and the higher throughput that they would need for that application.

We also saw, for example, a couple of benchtop customers that originally were HiSeq 3000, 4000 opportunities that purchased the NovaSeq.

To your point, that's an opportunity that's been percolating for a while, but then really what got it over the line was when they saw NovaSeq and what it would do for them.

We also saw a new customer, which is a recently opened cancer center that wants to do whole-genome sequencing or high-throughput exomes on NovaSeq.

So globally, those are examples of new applications, new customers that have been enabled because of NovaSeq.

That's super helpful color.

And I guess, just maybe in terms of how much of this outperformance you're sort of willing to drop through, I mean, this quarter, obviously, big uptick in gross margins, but also OpEx was held pretty tight.

What I think is kind of implied in the guide is a little bit more reinvestment.

I mean, how do you think about, particularly on the R&D line, where incremental dollars are being allocated just because you've obviously got such a great presence in your space, and there's so many dynamic opportunities.

It seems like your R&D budget is already quite sizable.

But considering some of the opportunities ahead of you, there's still a lot to invest.

And how are you sort of balancing that?

And what you're willing to kind of show on the drop-down?

That's a really good point, Ross, because we spend a lot of time internally and have just gone through a few months' process where we were sort of laying out our 3- to 5-year strategic plan.

And the reality of the opportunity facing us and where we are in that opportunity is that there is a lot more that we could go after than we want to go after than frankly we'll be able to go after.

And so a big part of our strategic planning process is to be very thoughtful about, okay, what are the opportunities we actually want to go after?

What are the opportunities we want to go after ourselves?

And then how do we enable the rest of the opportunity to develop maybe through a partner ecosystem, through things like the accelerator and ventures?

And obviously, we don't talk about the things that we have in development.

But you've noted, we have a commitment to innovation and R&D, and we think that there is opportunity in continuing to expand the market through innovation, through innovation and improving the workflows and improving the cost of sequencing.

There is opportunity for innovation in delivering in targeted areas these sample-to-answer solutions we talked about.

There is opportunity from continuing to get our products cleared in markets around the world.

And then the other area we're looking to expand is we believe that it will be money well spent to continue to expand our commercial infrastructure.

That means feet on the ground, especially in growing markets like China.

That means continuing to invest in our infrastructure, like our digital and e-commerce capabilities.

That also means expanding our partner ecosystem.

And so those are some of the things that we consider, the kinds of things that we consider as we prioritize our investments.

But we truly believe, again, we're still at the very early stages in almost every market we're in of this very large opportunity in front of us.

And so the hard thinking is what you invest and when?

The next question in the queue comes from Dan Arias from CTI.

Francis, just on the instrument side.

I mean, we can kind of talk to how much of the NovaSeq customer base has placed a multisystem order but is still kind of waiting for a portion of that order.

It seems like there's a good number of labs that have machines, but they don't have everything that they bought and that maybe that's a meaningful part of the placement dynamic this year.

Do you agree with that?

And if so, is that something that maybe you can put numbers around?

It's definitely a part of the dynamic.

If I look at the -- certainly, the large genome centers, the X shops, the majority of them have purchased at least 1 NovaSeq.

But for most of them, the actual, both delivery and then re-platforming, is still in front of us.

Some of that will happen in 2018.

Some of them will start delivering additional units even in Q1, and then we'll work over the course of the year to move more of their workloads onto NovaSeqs.

But that's a dynamic that will play out not just this year.

It will play off next year and the following years as well.

That dynamic is definitely a factor in our backlog.

So as we walked into 2018, if I look at the NovaSeq backlog we were carrying, definitely, the remaining units of a multi-unit order from a large genome center, so that category makes up a chunk of the backlog we're walking into 2018 with.

But if I look at overall NovaSeqs for this year, a large number of the orders are still going to come from the onesie-twosie orders from the customers, the HiSeq customers that are moving to NovaSeq.

So if I look at the order profile over the year, I'd say a bigger chunk of it is coming from the HiSeq customers, not the X customers.

But if I look at the backlog we're walking to 2018 with, there's more of the X remaining units to be delivered in there.

Yes.

Okay.

That's great.

And then maybe just on NIPT.

Curious about your sense for the capacity needs of labs there if you do take a big step forward on volumes.

And then to the extent that, that does impact system demand, how much do you think goes to the NextSeq versus the NovaSeq?

Yes.

I definitely think that the NIPT segment will see a continued growth in demand, and mostly because of 2 things.

One, the continued growth in reimbursement around the world.

And we talked about how that's building, especially globally.

It built nicely over '17, and we'll start to see that play out in '18.

In that segment, the first area that we see impacted is consumable growth, and so the first thing you'll see as you start to see NIPT reimbursement expand is you should see that show up in our consumable revenue.

And then obviously, they continue to add systems as well as they need more capacity.

That is primarily a NextSeq business, I'd say.

And so if you start to think about where it shows up in terms of instruments, say, today, it's primarily sort of NextSeq.

The next question in the queue comes from Tycho Peterson.

Francis, appreciate all the color in NovaSeq.

I just got 1 or 2 clarifications.

Now that you've got the Xp Sample Prep out and the S4 chip, can you give us a sense of how much of the funnel incorporates the Xp device?

I'm just wondering on kind of some of these larger multisystem orders.

And then separately, how do we think about the S1 launch driving Nova demand?

I mean, these are really for kind of newer to high-throughput sequencing and benchtop customers that you're trying to upgrade?

Is that the right way to think about it?

Yes.

So let me start with S1, and that's exactly it.

So S1 is already targeting the HiSeq customers that have either the lower sample volume.

They're the ones that would be most attracted to S1.

And for them, that could be the trigger that causes them to start to look at moving from HiSeqs to the S1.

But it's also going to be customers that wanted a fast turnaround time.

As you remember from our HiSeq customer base, there was a group for which rapid turnaround was a key criteria.

And so S1 does appeal to that segment of the HiSeq customer base, too.

We think that is a segment that some of them were waiting for an S1 because that's what makes the most economic sense for their business model.

I think Xp is helpful to this market, and there are definitely customers who needed it for their workflow.

And the way they ran their business was they were doing, in some cases, if they were a service lab, they were doing 1 lab per customer, or 1 lab per genome.

And so the Xp workflow does help them with their business.

I think it's going to be incrementally helpful.

I don't think we're going to call out sort of the attach rate or how many of those units we sell, but it definitely is helpful in terms of having NovaSeq embedded into the workflow of some of our -- certainly, some of our larger service customers.

Okay.

And then if we think about iSeq, is MiniSeq a good proxy?

I know you talked about having 600 MiniSeqs out there.

And how do we think about cannibalization and maybe where MiniSeq fits in going forward?

Yes.

I think it's somewhat of a proxy.

I think, though, it's not a great proxy.

Because if you look at the economics, there's a pretty big difference between iSeq and MiniSeq, right?

So iSeq's $20,000.

MiniSeq is 2.5x as much at $50,000.

But MiniSeq is 2.5x as expensive, but it delivers 6x the output.

It delivers 6x the number of reads per run.

And so if you have the capital, if you have the sample volume, you'd go for MiniSeq.

And so iSeq is really targeted at customers that are saying, look, a $50,000 instrument is just out of reach for our lab or they just don't have the sample volume to warrant even a MiniSeq.

And so I really think of it as opening up a different segment, and I -- there may be some cannibalization, but I don't expect there'll be a lot of cannibalization.

And in that sense, there is a bit of an analogy where we didn't end up seeing MiniSeq cannibalize the MiSeq.

We thought maybe up to 25%, and they ended up not being that at all.

And I think maybe that's similar.

Okay.

And if I could just ask one last one, can you comment on what's baked in the guidance for arrays, given the inflection you saw in samples last year.

Could arrays be up 20% again this year?

Or how should we think about it?

Yes.

So I would say it's in line with overall guidance, Tycho.

So basically, building some of the very strong demand and growth that we saw in 2017, and I'd say you would expect the same trends in 2018 as well.

The next question in the queue comes from Derik De Bruin.

When you look at the overall instrument number for 2018, when you think about the revenue number, it's like do you expect total revenues for instruments to be up in 2018?

Because I'm trying -- it sort of is a follow-up on Tycho's question.

I'm just trying to get any sense of the sort of like the underlying demand for the MiSeq and the NextSeq as you sort of look at that.

And what's sort of the ASP are you looking at for -- or are you using for your calculations on the Nova?

Yes.

So Derik, I'll talk about the instrument demand for 2018 or what's basically embedded in guidance.

So obviously, we gave you the NovaSeq shipment expectations for 2018.

So that's self-explanatory there.

So we do expect growth in placements of NovaSeq.

We also expect growth in placements across really all of our instrument platforms with the notable exception being the HiSeq family of instruments.

And for -- iSeq is also going to be a tailwind, driven by the fact that it's a new product introduction.

So basically, in summary, we do expect growth in the instruments line for 2018.

Great.

That's helpful.

On the -- is there anything embedded in terms of for GRAIL?

I mean, you called out Helix.

I assume that Helix is in the DNA -- is Helix being recorded in the DNA sequencing services line?

Helix, essentially, shows up across our revenues and shows up across our whole income statement.

And then we essentially record a noncontrolling interest adjustment when we're calculating the effect on EPS.

And for GRAIL, I mean, the -- are any of the -- are you shipping instruments to GRAIL right now as part of that -- for the ramp-up in placements for this year?

So the way it works, GRAIL is just like any other customer of ours, right?

A great customer and a large customer, but like any other customer.

So as they need additional capacity, they buy instruments from us and they buy consumables from us.

So we talked about the fact that they were one of our larger customers last year, and we expect them to continue to be a big customer of ours going forward.

We don't share specifics about what instruments they're buying when obviously.

The next question in the queue comes from Steve Beuchaw from Morgan Stanley.

This is Steve Beuchaw in for Steve Beuchaw.

I wonder, Sam or Francis, if you could give us just a little bit more modeling help with the trend on sequencing consumables that you anticipate for 2018.

Any sense for the distribution first half versus second half?

Does it look -- I mean, 2017 was a little unusual, given the first half destocking dynamics.

What is the growth rate for sequencing consumables look like in '18?

And any view on distribution would be super helpful.

Yes.

So Steve, we're not going to give a specific numbers in terms of sequencing consumables growth, but I'll give you some of the underlying drivers at a high level.

So we do expect growth over the course of 2018 in terms of overall sequencing consumables versus '17.

That growth is going to be driven by the NovaSeq uptake and by also growth across NextSeq and other platforms as well.

Again, with the exception being that HiSeq, we do expect across the HiSeq family to see a reduction in terms of the overall consumable dollars across HiSeq.

But that's going to play out in a, I would say, over time, and it's going to be fairly choppy.

So there's going to be some, I would say, fluctuation over the course of the quarter.

So I'm not going to give you a specific number in terms of what to expect for each quarter.

There will be seasonality, though, mostly in Q1, and so you will see a lower sequencing consumable number in Q1 driven by seasonality.

But across the course of the year, you will see growth.

Okay.

And then just 2 housekeeping items from me.

One, apologies if I missed this, but could you give us the -- or give any sense directionally for the NovaSeq order number in the fourth quarter?

Did we get a step-up there relative to third quarter?

And then any commentary on China, given the growth commentary there, just thinking through the drivers would be really helpful, and I'll get back in queue there.

Yes.

For orders, Steve, we're not providing order information anymore, as we had mentioned last quarter.

So won't give you much there.

But in terms of China, we had 12% growth in terms of shipments in Q4, as I mentioned.

And this was lower than the Q3 number that we had reported, but China is essentially -- is choppy as well in terms of overall growth, driven by sometimes the timing of instrument placement.

So back in Q2, I believe, we had in the mid-teens of growth.

So it's fairly in line with what we saw in Q2, but a step down from Q3.

And then what we said, Steve, in terms of NovaSeq shipments, so we said that in Q4, we shipped in the high 80s NovaSeq units, and that was the highest we've shipped of any quarter over the course of the year.

So you can see that's continuing to build.

The next question comes from Patrick Donnelly with Goldman Sachs.

Can you put out any color on the utilization mix of the NovaSeq customers you continue to see new-to-sequencing customers make up about 1/3 of the customer base?

Is it a wider range for pull-through between low utilization and high ones compared to prior systems?

I guess, I'm just curious if the new-to-sequencing customers are meaningfully lower than kind of the existing customers who buy them.

The trend is not, frankly, in my opinion, very different than what we saw before.

And so what happens is and we saw this with the X. And what we see is the first customers who come on board, certainly it was true with the X, too, are higher utilization customers.

So they are the ones that, they understand sequencing.

They typically had been using another high-throughput platform before.

And so the first few instruments, you see utilization that's extraordinarily high.

And then you start to see some of the newer-to-sequencing customers come on.

And so initially you see a little bit bimodal and then very quickly, it sort of normalizes to almost a line, not quite a line, but almost a line where people will come in, and they will start to build up their pull-through on the instrument and work their way up that line.

And so we're still -- we're seeing that with NovaSeq.

We know with experience now that it will still take maybe a couple more quarters before the -- that cohort settles down to a more normalized state where we can call a pull-through that you guys can use to model.

Okay.

Helpful.

And then maybe just on the balance sheet and the cash continues to build now that you guys are over $2 billion in cash.

How should we think about the capital deployment strategy?

I mean, maybe just talk about the pipeline, the assets that are out there and what makes sense to add to the portfolio here?

Yes.

So I'll start, and I'll just say that our capital deployment philosophy hasn't really changed.

We remain very focused on, first of all, we need to invest in our capabilities.

We need to invest in potential improvements to our technology, be it through partnerships or other opportunities that we have.

And then in terms of other opportunities out there, obviously, hard to comment and give you specifics there, but they need to essentially be aligned with our strategic priorities and focus and be also accretive to us from an overall growth standpoint.

So nothing really has changed materially from what we've discussed before.

It really remains focused on how do we improve our technology?

How do we upgrade our technology and look for partnerships in that space as well?

And then finally, on a probably less material note, we also continue to focus on offsetting any equity dilution with share repurchases as we did in Q4 when we bought back some shares.

Yes.

And if I could just sort of underline what Sam said.

Look, our top priority continues to be organic innovation to accelerate the market.

And so we're doing a lot of work there.

But I mean, frankly, I think, it's almost every 6 weeks, the leadership team meets, and we review what's happening outside Illumina.

And we review the landscape to see if there are innovations that we think would be better as part of Illumina, and we've done this for many years now.

And so we are very willing, as we've demonstrated in the past, to go out and acquire something that we think will be long-term accretive to us.

You've seen us do technology tuck-ins.

You've seen us do bigger ones like even way back when there was Verinata or Solexa.

Obviously, in the last couple of years, we haven't seen anything out there that we feel we had to have, but we're -- we look at it all the time.

The next question in the queue comes from Tim Evans with Wells Fargo.

Now that we are at the end of 2017, I was wondering if you might update us on the mix of your consolidated revenue when it comes to clinical, translational and research.

And what I'm really trying to get at here is how fast is research growing versus how fast is clinical translational going?

I know you've given us a few pieces here and there, but is there -- could we get a little bit of a higher-level view of that?

Yes.

So let me give you some top-level sort of observations.

I'd say that what we said now is that if you look at the mix of our business, the clinical part of our business, that includes translational, is about 45% of our business, and that's up from the high 30s in 2016.

So we are continuing to see more of our business come from the clinical markets.

That's clear because we're seeing higher growth in those markets, right?

So we talked about the growth rate that we've seen in oncology, that has been over 20% as we look at the growth rate.

We talked about the growth rate that we see in NIPT.

We also have talked about the fact that the clinical markets are also just very much larger over time, and so they are growing faster and they are larger over time.

So they will be, over time, the majority of our business.

They will represent the majority of our business over time, so we do expect it to tip over the 50% mark over time.

The research markets are going to continue to be an important market for us.

They are important because, one, today, they represent over half our revenues.

But also they are very important customers because very often, everything that shows up in a clinic, at some point, started in a research lab.

And so our research customers represent not only an important source of revenue for us, but they are an important source of thought leadership.

They are great customers in terms of helping us think about the direction that our technology needs to evolve.

In terms of growth, we've talked about the fact that, overall, the research markets, rough and tough, are growing in the high single digits.

It's how you should think about it.

Now there's still an opportunity for us to grow faster than that in the research markets to the extent that you're seeing research dollars moved into genomics research.

And so in the U.S., for example, it's been a good sign for us that you're starting to see more and more of the NIH agencies take on projects that are genomics projects.

It's not just coming from the NHGRI, but you're starting to see incremental dollars from NHLBI, NCI and a whole bunch of other agencies.

And so the extent that you can see more research dollars moving into genomics, there's a chance to grow faster than the overall research envelope as well.

Real quick question on tax rate.

A lot of moving parts here, and I hear you on the $0.10 from tax reform, but can you just tell us what the tax rate is that's embedded in the 2018 guide?

So we haven't communicated an exact tax rate.

All I would tell you is there is some improvement in the tax rate from 2017 to 2018, driven by the tax reform benefit.

But as you said, there's a lot of moving parts.

In Q4 of 2017, we actually had significant benefit related to the mix of -- foreign earnings, driven by consumables and where we manufacture consumables.

We also had some benefit in Q4 related to some release of reserves, what we call provision to return benefit from prior year returns.

So that was a onetime benefit in Q4 that, at this point, at least we haven't modeled to repeat in 2018.

But there is moving parts.

But I would say just overall, some benefit in 2018 in terms of the total tax rate versus 2017.

The next question in the queue comes from Bill Quirk with Piper Jaffray.

Appreciate the color around NIPT expectations for 2018 and the color around the Dutch tender.

Francis, I wanted to talk a little bit about, I guess, expectations for the oncology franchise, given that we obviously have some important, both label and reimbursement milestones, that we crossed in 2017?

Sure.

Oncology has been a really important story for us as we look at 2017, right?

If you look at even just in Q4, oncology testing was up 60% in Q4.

And oncology testing, if you look at the whole year, was up 38% over 2016.

And so we're seeing a -- the real emergence in oncology, there's a number of factors that are driving it.

One is just the emergence of liquid biopsy.

And so some of our liquid biopsy customers are driving some of the demand that we're seeing.

But as you pointed out, we're also seeing a number of trends play out in terms of regulatory approvals, in terms of reimbursement that may not have yet had the impact in 2017, but I think are setting us up well for future growth in oncology.

So we talked about the change in guideline from and reimbursement procedure at Cigna, which we think is a good step in the right direction towards the routine adoption of multi-gene panels.

And as we said, what's happened there is Cigna said that they will reimburse multi-gene tests as long as the clinical utility for one of the genes in the test has been established.

That's a good step forward.

We talked about the CMS decision where they're going to reimburse genomic tests for more than just non-small cell lung cancer but for all solid tumors, which they used to not do before.

We also like what's happened with both of the regulatory decision and also the reimbursement decision around the FoundationOne test.

And so there are a number of things that got put into place in 2017 that may not have been responsible for the results in 2017, but we think set us up well for the future.

Got it.

And then just a follow-up on the array business.

Given some of the increases in traffic around Ancestry and 23andMe, particularly relating to the Christmas holiday, I guess, I was a little surprised that the array consumables weren't up a little higher than they were.

Is there some sort of delay by any chance?

And we tend to see, say, a Christmas benefit coming in the first quarter as the samples are processed?

Yes.

That's exactly right.

So the way it works and there are sort of 2 things that I'll call out.

One is just sort of where you would see the results.

So the Ancestry business really shows up in terms of array services.

So it's not the only thing in there, but it's one of the bigger things in our array services line.

So that's where you would have seen it, not necessarily array consumables.

But the other consumer, customers we have, like 23andMe, would be in array consumables.

Now in terms of how it actually plays out, so the way it plays out in the market is a consumer will buy the kit, and they'll buy the kit either over Thanksgiving or for Christmas holiday.

Some of that buying is to be gifted.

And so what happens is the kit is purchased, it shows up as an order for Ancestry or 23andMe and sort of they talk about it as orders that they got in Q4.

That kit then has to be gifted.

The person who receives it has to spit in a tube and send that tube to us before we start to work on it.

And therefore, it shows up as revenue for us, as array revenue.

So you can see with that process why a Thanksgiving or a Christmas purchase could show up as Q1 or even Q2 revenue for us.

And so there's a natural delay in how that works between what shows up for them and what shows up for us.

And the last thing I would just add to Francis' delineation between where the array revenue shows up -- or where the consumer revenue shows up, the microarray consumables is also not just consumer business.

There's a lot more other stuff in there as well.

So when you're thinking about growth rate, just keep that in mind.

The next question in the queue comes from Dan Leonard with Deutsche Bank.

So just want to press on the NovaSeq ship guidance a little bit.

It looks like you're allowing for a lighter quarterly shipment trend in 2018 than is recent trend.

But presumably, you ended the year with a big backlog.

You're manufacturing at full tilt, so you can draw down your backlog.

You have the S1 chip, the S4 chip -- or S1 soon, S4 chip already, Xp already, so you can open up all the applications.

Why couldn't the NovaSeq shipments be a bit higher than 330 to 350?

Sure.

So we shipped last year 285 -- approximately 285 NovaSeqs.

And so the range we're talking about this year, 330 to 350, is a step-up from last year, but what we said before is you should think about it as a step-up more because you're seeing a full year effect in 2018.

And we only -- we were constrained for, at least the first 2 quarters of last year.

And so what we have said is we expect this to be a more measured sort of upgrade cycle playing out over 3 to 5 years, and we've done a number of things to make that happen.

So shipping the S4 flow cell and the Xp device in Q4 rather than earlier in the year, having S1 come out now.

And so what we want is, and what we're expecting to see in the market, is this more measured sort of upgrade cycle that plays out over the 3- to 5-year period.

So yes, you will see more shipments this year at 330 to 350 than you saw last year, approximately 285.

But again, think about it as more of a full year effect rather than we do not want to see a pop and then sort of a decline.

Okay.

And then Francis, for my follow-up, a year into the NovaSeq launch, do you have any better sense of what the swap ratio looks like?

So 1,900 HiSeqs could convert to what number of NovaSeqs?

Yes, what we said before, and I think it's even more true than was last year, the way we think about it is not the number of instruments back.

So we don't think about it as an instrument swap ratio.

We think about it as number of customers out.

So you can start with the, what, 850 HiSeq and HiSeq X customers and work out.

And say, look, some of them, certainly some of the HiSeq customers, may go to NextSeq, but the majority of them will go to NovaSeq.

And certainly, on the X customers, they will go to NovaSeq, right?

So the way we thought about it and now that I've spent a lot of time talking to customers, I think it is the right way to think about it is really start from number of customers out, assuming that the majority of them will purchase at least one and likely more than one.

And the reason for that is when customers think about purchasing the NovaSeq, their demand profile isn't static, right?

So as they start to think about the NovaSeq, they start to think about what else they could do.

I mean, Regeneron is the perfect example, right?

So they took on the 500,000 sample U.K. biobank project and said they could get it done by 2019.

But it's not something they would have done on their HiSeqs.

And so as we have those conversations, the dynamic happening there is that the demand profile changes in many cases as they start to think about what a NovaSeq allows them to do.

And so taking the old demand and then trying to figure out what capacity of NovaSeqs they'll need would get you to the wrong answer.

And it comes from Sung Ji Nam from BTIG.

Just one kind of a 2-part question on microarray.

You guys showed the inflection point for the consumer genomics in terms of the volume at the sample volume in 2017.

And was curious as to what -- from your standpoint, what's driving that?

Is that kind of more consumer genomics providers entering the market?

Or is it -- just kind of curious as to -- I'm just trying to get a better sense of where this could go.

And then also, I think Sam alluded to other applications within microarrays also driving growth for that business.

And if you could provide more color, in the past you talked about ag bio, et cetera, so if you could provide more color there as well.

Yes.

Sure.

I think one of the biggest factors that's driving this inflection point we saw last year is the elasticity in the market.

And we talk about elasticity a lot at Illumina, and we talk about the elasticity in the oncology market and NIPT, and it's very true in the consumer market, too.

If you talk to the leading players in this space, 23andMe, Ancestry, they'll tell you, and I don't know if you heard Anne speak at JPMorgan.

She has a terrific story about just the huge spike in demand she saw when the 23andMe raised -- lowered the prices of their kits.

And she'll talk about when they went from about $150 to $100 a test to, in some cases, now $50 a test.

It opens up a vast demand for them.

And so probably the biggest thing that drove the inflection point last year was that the tests reached prices, the $99 mark, and in some cases, over the Black Friday weekend, the $50 mark, that really tapped into this elasticity of demand in the consumer market.

I think that's simply the biggest part.

There are a number of entrants in that market, but in the U.S, I think it's primarily Ancestry, 23andMe, and then maybe Family Tree DNA, a little further behind there, so that's the big players.

So it's not that there's many more entrants in the space.

It's just that you have price points now that are really attractive to customers.

Now Helix allows more experimentation in that space, and so the players that I talked about are all targeting the genealogy space, and that's emerged as the first killer application in consumer genomics.

What Helix allows us to do as a market is really touch all the other markets, right, the wellness market, the nutrition market, the fitness market and see what's the next killer application in consumer genomics beyond genealogy.

Great.

And then a bit more color on kind of other applications where you might be seeing a lot of growth as well.

Yes.

So as Sam touched on, if you look at the numbers we report, we report an array service line, and there, the majority of it is the consumer genomics space.

And so they -- that business does drive the growth rate of that line.

If you look at array consumable line, there are a number of other segments that show up in the line.

So our ag business, our agricultural business, for example, shows up in that line.

And so while that's an important segment, it's growing at a much lower rate than the consumer genomics, right?

So if you look at the blended growth rate in array consumables, you'll end up with something in the single digits because the majority of that line is actually not consumer genomics.

Thank you.

I'll turn the call back over to Jacquie now.

Ms. Ross?

Thank you.

As a reminder, a replay of this call will be available as a webcast in the Investors section of our website as well as through the dial-in instructions contained in today's earnings release.

Thank you for joining us today.

This concludes our call, and we look forward to our next update following the close of the quarter.

Ladies and gentlemen, thank you for participating.

You may now disconnect.