Illumina Inc (ILMN) 2017 Q3 法說會逐字稿

Welcome to the Third Quarter Fiscal Year 2017 Illumina Earnings Conference Call.

My name is Adrienne, and I'll be your operator for today's call.

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

And I'll turn the call over to Jacquie Ross.

Jacquie, you may begin.

Thank you, Adrienne.

Good afternoon, everyone, and welcome to our earnings call for the third quarter of fiscal 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; Sam Samad, Chief Financial Officer; and Marc Stapley, EVP, Strategy and Corporate Development.

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 the audio portion 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 results 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, and good afternoon, everyone.

I'm pleased to report that the Illumina team delivered a strong third quarter, with revenue growing 18% year-over-year to $714 million.

With solid results across our portfolio, it is clear that Illumina is both enabling and accelerating the global adoption of sequencing applications.

At Illumina, our mission is to improve human health by unlocking the power of the genome.

And our commitment to innovation, while clear across our portfolio, is most recently demonstrated with the launch of the NovaSeq System.

We believe NovaSeq is a transformative platform that enables existing sequencing customers to increase output and lower costs, while at the same time lowering the barriers for new-to-sequencing customers.

Ultimately, our goal is to unlock the benefits of sequencing for as many people as possible.

With that in mind, we are very pleased with NovaSeq's performance since its launch in early January.

In total, we have shipped approximately 200 systems in the first 3 quarters of 2017, including more than 80 in the third quarter.

As expected, NovaSeq shipments increased sequentially from last quarter, with shipments no longer constrained by manufacturing capacity.

With approximately 70 orders in the quarter, we ended Q3 with a backlog of more than 100 systems.

Going forward, and as expected, backlog should start to trend downwards now that manufacturing constraints are behind us.

Overall, we are currently on track to significantly exceed the 2017 forecast we targeted at launch.

As expected, most of our third quarter NovaSeq orders came from existing HiSeq customers.

We're seeing robust demand from HiSeq customers upgrading their infrastructure and are still in the early stages of the upgrade wave.

We remain confident that most of our 800 HiSeq customers will upgrade to NovaSeq to access its power, improved workflow and lower sequencing costs, especially with the introduction of the S4 and S1 flow cells.

Additionally, NextSeq will continue to be the go-to platform for a subset of HiSeq customers.

Moving to HiSeq X. Many of our HiSeq X customers have ordered their first NovaSeqs to start validations in new projects.

It's worth noting that these very high-volume customers will generally align their fleet transitions with the completion of existing projects.

As a result, we expect to continue to see strong NovaSeq uptake through the fourth quarter of 2017 into 2018 and beyond.

Additionally, we are seeing continued signs of elasticity, with about 1/3 of NovaSeq orders in Q3 coming from new-to-Illumina sequencing or previously benchtop-only customers.

Examples include a children's hospital in North America that has purchased a NovaSeq to bring sequencing in-house, and an existing MiSeq and NextSeq customer that is moving to NovaSeq to begin providing exome sequencing and liquid biopsy tests.

We also saw a growing multiunit NextSeq customer that is moving to NovaSeq to leverage the increased throughput on a system that occupies a smaller footprint overall, requires less FTE time and offers a lower cost per sample.

We therefore believe we are still in the earlier stages of a multiyear adoption cycle that extends well beyond our HiSeq customer population.

Moving to sequencing consumables.

We continue to see strong adoption, with growth of 14% year-over-year to $380 million.

The $42 million sequential growth in sequencing consumables represents the largest-ever sequential dollar increase in our sequencing consumable business.

This was driven by growth in our installed base and strong NextSeq utilization, in addition to stronger-than-expected HiSeq consumables.

HiSeq consumables increased sequentially, driven by a few of our large clinical commercial customers who are seeing growing sample volume.

Excluding this group of customers, HiSeq consumables declined.

Our expectation remains that HiSeq consumables will trend lower going forward, given the transition to NovaSeq and NextSeq.

As expected, we saw a slight sequential pickup in HiSeq X consumables, driven by translational studies and growth in China.

With the S4 flow cell now broadly available, we expect HiSeq X consumables to follow a similar trend to HiSeq consumables, with the decline to begin sometime in the coming quarters.

Moving to our benchtop portfolio.

System shipments were essentially flat from the second quarter.

Our win rates remains stable, and new-to-sequencing customers represented over half of NextSeq, MiniSeq and MiSeq placements.

Utilization was within each benchtop instrument's respective guidance range, except for a record NextSeq performance that exceeded the upper end of our range, and was once again driven by production clinical customers.

In terms of new product launches, we delivered the new S4 flow cell to early-access customers as expected.

I'm pleased to share that feedback on initial runs at 7 customer sites has been very positive, with performance exceeding specifications.

Building on the S2 flow cell, the new S4 flow cell enables a lower price per sample compared to HiSeq X for customers in our highest tiers of utilization, and is therefore ideally suited for high-intensity sequencing applications.

S4 is now commercially available.

We are also gearing up for the release of the NovaSeq Xp workflow and reagents.

This is the individually addressable lane workflow that we first announced back in April.

The device enables libraries to be loaded directly into each lane of the flow cell, allowing customers to partition different libraries, projects, samples and applications.

We expect to begin beta testing in the coming weeks, ahead of a full commercial release expected before the end of the year.

The next flow cell we're working to add to the portfolio is the S1, which we now expect to bring to the market in the first quarter of 2018.

The S1 is the lowest output flow cell for NovaSeq at 1 terabase per run, or 500 gigabases per flow cell.

This is expected to be a very attractive upgrade option for HiSeq 2500 customers, including those who utilize the rapid run option.

The addition of S1 will nicely round out our flow cell portfolio, which can now accommodate data yields of between 500 gigabases and 6 terabases.

We have suspended plans for an S3 flow cell with an output of 4 terabases, given strong customer preference to use S4 for high-output applications and either S1 or S2 for lower-output applications.

You may also recall that we had plans for a lower-output and marginally lower-cost NovaSeq system, the 5000, to run just the S1 and S2 flow cells.

Based on customer feedback and our experience over the last 3 quarters, we believe that the existing 6000 fully meets all customer needs, in addition to enable greater flexibility to scale in the future without an additional capital purchase.

Consistent with our broader objective to streamline our product portfolio, we have therefore decided not to launch the NovaSeq 5000, instead focusing on the 6000.

At Illumina, we strongly believe that sequencing offers the promise to transform lives and are encouraged to see sequencing continue its progress into clinical application.

In the third quarter, shipments to our clinical customers grew 35%, largely driven by our liquid biopsy customers.

More specifically, oncology testing shipment growth accelerated from last quarter.

The FDA's approval of Kymriah as the first immune cell therapy is, of course, a very exciting development for the field of oncology, opening a whole new world of possibility for gene and cell therapies.

Sequencing is an important enabler in the discovery of immunotherapies.

And with approximately 1,000 immuno-oncology drugs and clinical trials in the U.S. alone, Illumina is committed to supporting our customers as they endeavor to deliver new hope to cancer patients and their families.

Another important area for Illumina is rare and undiagnosed diseases, where patients, including newborns, and their families can find themselves on a multiyear odyssey searching for a diagnosis.

There are more than 7,000 genetic disorders that aren't well known enough to be routinely identified by physicians, but that can now be identified by genomic analysis.

The right diagnosis can reduce pain and enable earlier and potentially life-impacting or life-saving therapy.

It was therefore great to see UnitedHealthcare's recent decision to cover whole exome sequencing for the diagnosis or evaluation of genetic disorders in patients and their parents and siblings.

Starting on November 1, UnitedHealthcare will cover whole exome sequencing for patients where clinical presentation is nonspecific and does not fit a well-defined syndrome for which a specific or targeted gene test is available.

Once the UnitedHealthcare decision takes effect next week, more than 100 million lives will be covered for whole exome sequencing for rare and undiagnosed diseases in the U.S. This compares to almost none a year ago.

Finally, in its first full quarter of availability following the CE-IVD release, our VeriSeq NIPT CE-IVD solution delivered another strong quarter, with the number of shipped samples and revenue almost doubling sequentially.

The product is performing well in competitive tenders, and we added a sizable German customer during the quarter that will start ramping around the end of the year.

Moving back to the quarter.

Microarray revenue grew 27% to $121 million, with continued momentum among our direct-to-consumer customers.

Helix, formally launched in July, creating a completely new marketplace for emerging consumer genomics.

During the quarter, they added 6 new apps from 4 new partners, bringing the total number of products available to 26.

At the same time, Helix is exploring new commercial avenues with a selection of products now available on Amazon.com and has teamed up with Illumina Accelerator to support innovative start-up companies working to create breakthrough DNA-driven products for consumers.

Before I hand the call over to Sam, I want to share with you a change in one of our leadership roles.

I'm happy to announce that Marc Stapley has accepted the role of EVP of Strategy and Corporate Development, responsible for corporate strategy, corporate and business development, government affairs and global infrastructure, including IT and facilities.

Marc has been an invaluable partner to me as Chief Administrative Officer and has hired some exceptional leaders while helping our G&A functions develop in scale.

As a result of the change, our CFO, Sam Samad; General Counsel, Chuck Dadswell; and the new Chief People Officer, when hired, will report directly to me.

I'm looking forward to have Marc now focus on driving our strategy process and getting more deeply involved in key transactions, including our population sequencing efforts.

Finally, I'll highlight that Illumina Ventures raised $230 million for its first genomics-focused fund.

To date, Illumina Ventures has made investments in 7 early-stage companies in industries ranging from novel therapeutics, diagnostics and research tools, to food security and synthetic biology.

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

Sam?

Thanks, Francis.

As discussed, third quarter revenue grew 18% year-over-year to $714 million, largely driven by growth in sequencing consumables and instruments as well as strong performance across our microarrays portfolio.

Geographically, Americas grew revenue 20% versus the prior year period, while EMEA grew 15%, driven by growth in both sequencing instruments and consumables.

Asia Pacific grew 15% overall, as continued weakness in Japan was more than offset by the more than 30% shipment growth in Greater China.

Revenue from sequencing instruments grew 21% year-over-year to $128 million, driven primarily by NovaSeq and offset in part by muted shipments of the HiSeq family of instruments, as expected.

Microarray instruments had an unusually strong quarter at $12 million, driven partly by shipments to our direct-to-customer -- direct-to-consumer customers.

Total instrument revenue was therefore $140 million, an increase of 24% year-over-year and represented 20% of total revenue.

As Francis noted, third quarter sequencing consumable revenue was $380 million, up 14% from last year.

Microarray consumable revenue of $71 million was up 13%.

Total consumable revenue of $451 million represented 63% of total revenue.

Third quarter total product revenue, which includes freight, was $596 million, up 16% from the year-ago quarter.

Service and other was at $118 million, up 26% from the same quarter last year, driven by strength in genotyping services due to consumer demand and sequencing instrument maintenance contracts.

Moving to gross margin and operating expenses.

I will highlight our non-GAAP results, which includes 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 68.8% improved 180 basis points from last quarter due to a more favorable sequencing consumables mix with growth in NovaSeq and NextSeq.

Year-over-year gross margin decreased 320 basis points and was impacted by an increase in array services mix and lower instrument margin associated with the NovaSeq introduction.

Non-GAAP operating expenses were $299 million, up $2 million from last quarter with higher Helix launch expenses and increased headcount, offset in part by lower stock-based compensation expense.

Non-GAAP operating margin was 26.8%, up 470 basis points sequentially and lower than the 28.7% reported in the third quarter of last year.

Excluding Helix, operating margin was 29.7% compared to 24.4% in the second quarter.

For the third quarter, GAAP net income attributable to Illumina stockholders was $163 million or $1.11 per diluted share.

Non-GAAP net income attributable to Illumina stockholders was also $163 million or $1.11 per diluted share, and Helix dilution was $0.07.

Of note, our non-GAAP tax rate of 21.6% was lower than expected, with a onetime benefit associated with the increase in income attributable to overseas manufacturing compared to what we forecasted earlier this year.

We saw additional benefit with R&D credits and foreign tax credits on prior year returns filed in the third quarter.

Cash flow from operations equaled $235 million.

Improved revenue linearity and continued reductions in our collection cycle led to Q3 DSO of 49 days compared to 51 days last quarter.

Capital expenditures in Q3 were $82 million, and Q3 free cash flow was $153 million.

We ended the quarter with approximately $2 billion in cash and short-term investments.

During the quarter, we repurchased approximately 385,000 shares under our previously announced buyback program at an average price of $195.

Moving to guidance.

We now expect full year revenue growth of approximately 13%.

For the fourth quarter of 2017, we expect a more typical sequential increase in sequencing consumable revenue.

We expect fourth quarter non-GAAP gross margin to be slightly up from the third quarter and fourth quarter non-GAAP operating expenses to be flat to slightly up on a dollar basis from the third quarter.

Finally, we expect a more normalized tax rate for the fourth quarter.

We therefore expect full year 2017 GAAP earnings per share in the range of $5.56 to $5.61 and non-GAAP earnings per share in the range of $3.73 to $3.78.

Before we start the Q&A, I'd like to formally welcome Jacquie Ross to the Illumina Investor Relations team.

As most of you know, Jacquie has taken over from Rebecca Chambers, who has moved to a new leadership role in our finance organization.

We thank Rebecca for her many contributions as Head of IR at Illumina since 2012, and I know you will join me in wishing her every success in her new role.

Operator, we are now ready to begin the Q&A session.

(Operator Instructions) And our first question comes from Derik De Bruin from Bank of America.

So first question.

So you sort of alluded to it, Sam sort of alluded to it.

But 14% consumable growth in the quarter, up 24% quarter-on-quarter, 14% year-over-year.

I guess, with that stocking of NovaSeq reagents in the quarter, and you said more traditional ramp on Q3 to Q4?

Can you just -- the consumable number has been a little bit choppy the last few quarters.

I'm just wondering what's a normal ramp?

Yes.

So I think when we look at Q4 going forward, Derik, I think we will go back to a normal ramp, which is consistent with what we've seen typically, the growth of our consumables.

We did have in Q1 and Q2, a slight dip in consumables, driven by what we thought was destocking of inventories.

And as we look forward, we expect to see the growth of NovaSeq consumables take off.

We expect to see HiSeq consumables come down, per expectations.

And with HiSeq X consumables going forward, that's going to be a little bit lumpy, as we said before, because we expect -- we've had some noise with some of the continuation of experiments on the HiSeq X platform.

But that's going to be a bit lumpy as we go forward.

Now we have the S4 introduction, so we might see ups and downs.

But eventually, we will see HiSeq X consumables come down and be -- and see that offset by the growth in NovaSeq consumables.

So that's kind of the trends going forward, without giving you a number in terms of our growth rate.

But that's what you should expect in terms of directionally for sequencing consumables.

I'd just add, Derik, that on the NextSeq side, essentially, we did say that, that was above the high end of the range for the quarter.

And so it was -- it's not clear what will happen to that in Q4 as well.

I'm sorry.

NextSeq -- you said NextSeq at the high end -- above the high end of the range.

It was above the high end of the range this quarter.

Got you.

Okay.

And I guess, the delay in the S1 ship is -- and can you talk a little bit about that?

I thought it was -- my understanding that essentially it was going to be released about the same time as the S4.

Can you talk about that?

And I guess, is there any concern about meeting flow cell capacity?

Or is it just a research delay because of where the demand for flow cells or some manufacturing issues that you're trying to overcome?

Yes, it's a modest delay.

And the team prioritized in this quarter getting the Xp workflow and the S4 flow cell out, and so the contingency team is currently working on it, optimizing the S1 chemistries.

And so we expect that to come out in Q1.

And our next question comes from Tycho Peterson from JPMorgan.

I guess, a question on the instrument side.

As we think about NovaSeq mix here, I know you talked about 1/3 of orders from new-to-sequencing customers kind of consistent with last quarter.

But can you talk a little bit about how much is coming from benchtop-only versus the legacy HiSeq base versus X customers?

And then, given the price gap between the 5000 and 6000, I think it was about $200,000 when you first announced it, what gives you confidence that lower-end customers that may have been 5000 customers will actually trade up into the 6000 now that you've delayed the 5000 -- or canceled the 5000?

Yes.

So for now, 2 quarters in a row, we've seen a good demand, a healthy demand come from new-to-Illumina sequencing as well as benchtop only.

And we've talked about the fact that about 1/3 of the orders that came in for NovaSeq in these 2 quarters coming from that segment.

The majority, therefore, and just do the math, the majority did come from typically HiSeq customers and HiSeq X customers.

The -- we are definitely seeing the beginning of the HiSeq upgrade wave.

In the last quarter, we talked about the fact that about 10%, or just over 10% of the HiSeq customers, have bought NovaSeqs.

We obviously added to that number in this quarter, and we're starting to see X customers buy and validate their first NovaSeqs and start to think about the fleet re-platforming, that now that we've put S4 out, we expect to see more of those happening in the coming quarters.

And so we expect to see more multiunit orders come in now that we've put S4 out there.

In terms of the 5000 and the 6000, as you will recall, Tycho, typically, we do have sort of a lighter-end version of most instruments we put out, right?

So we had the HiSeq 3000 and the 4000.

We have the NextSeq 500 and the 550.

And we fully expected that would be the dynamic here with NovaSeq as well.

And what we found from talking to customers and our field force really sort of amplified that feedback back to the leadership team is that, in this case, there was a lot of demand for the next -- for the 6000.

So for the next platform to be, the 6000, for our high-throughput customers.

And there weren't a lot of customers asking for, and certainly not waiting, for the 5000.

And so what's playing out right now is that the choice our customers are having is, if they want a high-throughput instrument, they're looking at the 6000.

Otherwise, they're looking at the NextSeq.

And so, as we think about the upgrade opportunity for HiSeqs, we expect the majority of them will go to NovaSeqs.

But then there'll be a number of them that opt to get one or more NextSeqs instead.

Okay.

And then just a follow-up, can you just talk a little bit about the types of work being done on Nova?

And now you've had it out a couple of quarters, how much is exome versus whole genome versus other application?

We haven't broken it out.

But certainly, some of the anecdotal feedback I can share is that we're seeing sort of all of the above right now.

We are seeing customers that are new to sequencing getting into the game and are looking to do whole exomes or whole genomes.

An example of that would be a new-to-sequencing customer that we got out of the Middle East that bought NovaSeq to participate in the second phase of the Saudi whole genome project, for example, or the hospital that I talked about in North America that previously didn't do any of their sequencing in-house and now bought a NovaSeq so that they don't have to outsource the sequencing that they do.

We're seeing customer out in Greater China, for example, that purchased NovaSeq to support their collaboration with a genome alliance in that region.

And in that case, that was genome.

And then we talked about what's playing out with RUGD [ph], rare and undiagnosed diseases where there's reimbursement now for whole-exome trio sequencing.

And so hospitals that look to provide that will be looking at potentially NovaSeqs to do exomes.

And so at this point, it's too early to break it out specifically, but we're seeing all of the above.

We're seeing people getting into from panels to exomes.

We're seeing exomes to genomes, new to genomes.

And then we talked about demand coming from oncology applications like liquid biopsy.

And our next question comes from Bill Quirk from Piper Jaffray.

Great.

First question is, I guess, just on the S4 launch.

Certainly, our diligence has picked up a couple of times, there appears to be a healthy customer segment that's been waiting for that particular flow cell to be released.

And so I recognize you don't want to talk intra-quarter orders or anything like that, but can you just talk about the pace of interest and any color you'd like to add with respect to orders since you've released the S4 just a couple of weeks ago?

Yes, and what I'll share with you will be anecdotal.

But certainly, the economics makes sense for HiSeq X customers to really get excited by the S4 in a way that, frankly, the other flow cells aren't as exciting to them.

And so if you look at how the conversations are playing out with some of our HiSeq X customers, some of them did purchase NovaSeqs.

But there are more NovaSeqs that they can get familiar with the instruments and the platform, understand the workflow, understand the differences.

But for the majority of the HiSeq X customers, moving the fleet over to NovaSeq really makes sense once the S4 is out.

And that's how this market is playing out.

In some cases, the HiSeq X customers waited for the S4 to get into NovaSeq at all.

And in other cases, they got in, but are looking at the S4 as the enablement that they were waiting for to move their fleets over.

Got it.

And then, I guess, slightly changing subjects here.

Helix, Francis, are we at a point where we can talk about kind of the relative contribution of that, now that we have some additional apps launched and it appears to be kind of a broader rollout at this point?

And if so, anyway you can help us think about kind of the magnitude of that business, maybe relative to some of the other DTC customers that you support.

It's still too early for us to talk about the contribution that Helix will have.

We are -- they announced -- they launched last quarter, so only a few months post their launch.

And so they're still in their ramp-up phase.

I think sometime over the course of the next of the year, it'll make more sense for us to give you more details on that business.

But again, it's too early right now.

Yes.

It's still not material in terms of our overall revenues to discuss.

And the next question comes from Doug Schenkel from Cowen and Company.

Okay.

I have a -- I guess, it's one multipart question, going back to sequencing consumables.

First, you provided a lot of detail on legacy consumables revenue.

What you didn't say is how much of the $42 million sequential jump in consumable revenue growth was driven by NovaSeq.

Can you provide additional detail?

I guess, basically, I'm just asking how much of the increase is attributable to NovaSeq specifically.

Second, a bigger but more important question, and I think it's one of, if not, the biggest questions with the NovaSeq launch, is whether or not it's going to lead to labs spending more on consumables than it did before.

For example, if a lab spent $700,000 annually on consumables on 2 HiSeqs in 1 year and swaps out their 2 HiSeqs for a NovaSeq, will that lab end up spending more than $700,000 on consumables on that new NovaSeq?

How confident, based on what you're seeing now, and the answer being yes, that NovaSeq will lead to more consumable spending per lab?

And when can we expect to see this?

And the third part is really just a clarification.

When you comment on sequential growth per sequencing instrument, is that growth in both total revenue per instrument and overall revenue per instrument class?

All right.

Let me take a cut at the answers.

So the $42 million sequential consumable growth was driven by growth across all of our platforms.

So if you look at every single one of our platforms, we saw sequential growth in sequencing consumables.

But obviously, the growth wasn't equal.

So if you look at the high throughput part of our portfolio, the biggest part of the growth in that part of the portfolio came from NovaSeq.

And so that was the biggest contributor on the high throughput part of the portfolio to the sequential sequencing consumable growth that we saw.

On the benchtop side, again, although every platform grew, the biggest contributor to that growth was NextSeq.

And so while all our platforms grew, it was really NovaSeq and NextSeq that were the big drivers of the consumables growth that you -- the sequencing consumables growth that you saw sequentially.

In terms of will NovaSeq lead to lab spending more in consumables, we are at the stage where we can share some anecdotal feedback.

And what we are hearing is that, certainly, with S4, there are applications, there are samples that become accessible at the price points that S4 affords that will lead, over time, to labs being able to access additional funding and drive more consumable spend.

We're at the beginning of that, obviously, with S4 coming out and giving customers access to price points for those large-scale projects that they didn't have before.

And so we are encouraged by seeing a lot of new customers come in to buy NovaSeq because they believe that thesis.

Now what's still in front of us then is to see that pull-through show up in NovaSeq consumables.

But the early indicators in terms of people buying the instruments, in terms of new customers coming in, is very positive.

And then, actually, I didn't capture the third question.

I don't know if you guys did.

Can you repeat it, Doug?

Yes.

Just clarifying, when you commented on sequential growth per sequencing instrument, is that both total revenue per instrument and overall revenue per instrument class?

I guess, I'm just trying to see if what you meant by that was that pull-through per instrument across all instruments increased sequentially.

Let me check into that and I'll get back to you Doug.

And our next question comes from Dan Arias from Citigroup.

Yes.

Francis, maybe just to take a shot at the decommission trade-in dynamic around the Xs and the HiSeqs.

I'm curious whether you're finding that there's sort of a level of utilization that tends to trigger customers shutting down the box.

Obviously, some of these guys are keeping their existing systems and others aren't.

So I'm just wondering whether you're noticing any sort of commonality amongst those that are doing the trade en masse.

Yes.

So let's talk about both the trade-in and the decommissioning.

It isn't really -- our conversations aren't really driven as much by the utilization rates.

What we're finding is it's driven by what customers expect in terms of their business plans going forward.

And so, again, some of them are new that didn't have the history but are getting into NovaSeqs, and it's based on a business plan that they see accessible to them now, whether it's the new customer we have that decided to get into the whole-exome, whole-genome sequencing servicing market.

And so it's much more driven by business plans going forward rather than the utilization (inaudible).

There's some overlap between the 2, but it's not a perfect overlap there.

In terms of decommissioning, just for the numbers, we removed approximately 23 instruments from the HiSeq installed base, and we did not decommission any Xs from the installed base.

Great.

That's helpful.

And then maybe just sort of switch topics a little bit.

Francis or Sam, can you maybe help us with our models on the SERF line, I'm thinking specifically on Verinata as the NIPT volumes transition out of house.

Would you be willing to sort of talk volumes or maybe even revenue expectations for this year versus last year?

At this point, we're not talking about the specific volumes.

So we moved away from reporting those volumes, so we don't have that to share with you.

What we are sharing is the market dynamic, right?

And so from a market dynamic perspective, we continue to be very encouraged by the progress we are seeing in terms of NIPT reimbursement, both in the United States where we're now at 39% of the population, about 115 million lives that are covered for NIPT testing; and then the progress we're seeing in Europe, where we have some countries that are covering NIPT testing for the entire population.

And so we see that happening in the Netherlands, for example, in Belgium, and then countries like England and France, where we're starting to see parts of the population getting covered.

So we continue to see progress in terms of people having access to the test, which is then driving the sample volume growth that we talked about.

And that, obviously, will then drive revenue for us.

And the next question comes from Mark Massaro from Canaccord.

So last week at ASHG, Bill Gates gave a keynote and he talked about the gut microbiome as being something he's most excited about.

But then he went into, surprisingly, RNA sequencing, metabolomics and proteomics.

I guess, my question for you is, which area of research are you most excited about that you think that the NovaSeq can unlock and drive consumables utilization near term?

Yes.

And there are a number of areas that are emerging that we're excited about.

We're excited about the whole area of single-cell research.

We think that's a new research paradigm that is driving a lot of very exciting projects to understand the differential expression across cell types, even in a single tissue.

We're excited about the research that's happening, and this is more commercially oriented, but in immuno-oncology, for example.

We think that, that's driving some very exciting research.

You know, I'm never going to be on the side of arguing against Bill Gates.

And in this case, I actually agree that metagenomics is also a very promising and exciting area, and for genomics, in addition to proteomics.

So I think that's an interesting area, too.

It has, in the long term, some really exciting potential.

Great.

And just as a quick follow-up.

Can you give us a sense for the interest from biopharma, if that's picking up at all as it relates to doing some of their own sequencing?

We continue to see demand growing over the long term from that segment.

We talked about the fact, specifically, for example, that sequencing is a core tool in immuno-oncology and immunotherapies.

And so that area continues to be an exciting one that we are watching, that holds promise not just for us, but for patients as well.

And so we see them as a growing part of our business, and I expect that will continue to be so for a long time.

And our next question comes from Patrick Donnelly from Goldman Sachs.

Looking at the market size for NovaSeq, I guess, with another quarter of 1/3 of the customers being new to sequencing as the kind of orders here, I mean, has that changed your expectation for the market size given this recent trend?

And then it seems that's boosting the market size from the bottom and then maybe NextSeq is capturing a few customers you expected on the lower utilization side.

So are those 2 dynamics offsetting?

Or maybe just kind of how you think about the market relative to your initial expectations?

Yes.

So I'll start by saying that -- and we've said this, this quarter, we said last quarter, that we are very pleased with how NovaSeq is performing, and it is on track to exceed the full year plan we had for NovaSeq coming into this year and exceed it comfortably.

So overall, we are very pleased with how it's playing out in the market.

We continue to believe that in the long term, the vast majority of our HiSeq customers will upgrade and most of them will move to NovaSeq.

Now we expect some of them to move to NextSeq, and that's good for us, too.

But we do expect that (inaudible) will move to NovaSeq.

The 1/3 new to sequencing is incremental, obviously, to the upgrade wave that we are looking at, and that's encouraging to see it play out the way it's playing out now for 2 quarters in a row.

So overall, I'd say we are very happy.

It's exceeding our plan for the year.

We do believe -- we continue to believe, though, that this is a multiyear upgrade path.

So coming into the year, we thought it was a 3- to 5-year upgrade wave, and we continue to believe that.

And just to clarify, within that, it includes also previously benchtop-only customers.

So we see evidence of customers moving up to NovaSeq, as well as, I agree with Francis' point, a portion of that 800 customers will move down to NextSeq from HiSeq.

Okay.

Helpful.

Maybe one for Sam, just on the margin side.

When we think about Nova, obviously, still dilutive early in the ramp.

But how should we think about the margin profile over the long term relative to maybe some of the other instruments out there?

Yes.

As you said, I mean, as with any product introductions, I think in the first few quarters, you will see some of the initial costs, which we've been working through, improving.

The margin profile has been improving for NovaSeq.

That's helped also drive our overall gross margin improvement for the quarter.

I think going forward, what you need to expect, which will play out over the coming quarters, is that NovaSeq will essentially approach the -- and I'm talking about the instrument here, will approach the margin profile of HiSeq and close to HiSeq X, but not quite at the HiSeq X level.

So that's at the instrument level.

And our next question comes from Amanda Murphy from William Blair.

So I just had, I guess, a higher-level question as you think about long-term growth.

We get asked a lot, I'm sure everyone does, about just, I guess, thinking about what new applications or new markets that NovaSeq enables.

And so you talked a bit about how people are using the platform.

I just was curious how you're thinking about it over time and what you've seen historically.

Obviously, there's this idea of doing more samples because it's lower cost and then more complexity in terms of the samples that you have currently.

But if you think kind of longer term, would you agree that to drive growth that you would need to expand into liquid biopsy or something that's kind of new?

Or is that not the right way to think about it?

We definitely believe that NovaSeq will enable new types of sequencing to be done on a number of fronts.

So NovaSeq fundamentally enables larger experiments, larger cohorts.

And that could be at the macro scale, so population sequencing efforts, I think, are more accessible to more customers than they were before because of NovaSeq.

But also at the micro scale.

So we expect to see, and we are seeing, lots of exciting single-cell projects being contemplated because of the economics and the power of NovaSeq.

And so we expect, on the one hand, NovaSeq will enable access to larger cohorts at every scale in a way that hasn't been done before.

We also expect, and we're seeing, that even with the same size of cohorts, we're seeing a lot of sequencing-intensive applications being done more.

So the growth in liquid biopsy is enabled by people having access and more customers having access to more power at lower cost, and that's what NovaSeq does.

So the liquid biopsy market, I think, has accelerated by something like NovaSeq.

We will, we believe, continue to see the movement from single-gene panels to multi-gene panels, to exomes, to genomes, right?

And actually interestingly enough, I think the recent PAMA decision also helps that where there's a reimbursement framework starting to be set for the larger multi-chain panels.

But I do think you'll start to see this migration to doing bigger panels and then doing exomes and doing genomes, and we're starting to see that.

But it's early in that journey.

But over the course of multiple years, we all know you do better science, you get higher diagnostic yield as you move up that continuum.

And NovaSeq is helpful in that dimension.

And then the applications that require sort of needle-in-a-haystack type of work, and that's again like liquid biopsy where you need to do very, very, very deep sequencing to find the rare event, the rare molecule.

And there's a set of those applications that can now be enabled with the power of NovaSeq.

And so on a multiyear horizon, we do believe that NovaSeq fundamentally enables new types of sequencing at the cohort, larger cohort scale, but also at the more intense, deeper sequencing scale.

Got it.

And just one quick follow-up on the liquid biopsy side.

So obviously, I guess, GRAIL had talked about a product longer term and there's products already in the market for later-stage patients.

But do you think, just thinking about the research dollars or the move into liquid biopsy, is that enough to move the needle for you guys near term, so thinking the next 12 months?

Is that enough to be reasonably additive to whatever it is, consumable or instrumentation growth, over the next year?

I think on both categories, I think if you look at the screening side, there are a number of companies now that have been resourced and financed to look into that market.

So it's not just GRAIL anymore.

And in aggregate, I think that is a -- that's additive to our business as a whole, and it's sort of a positive.

We've talked before about the fact that GRAIL had the potential to be among our biggest customers.

And to the extent that you have more companies that are doing sequencing-intensive applications and really large-scale studies like that, having more customers that show up among our biggest customers, is definitely good for our business.

It's -- that'll probably play out over time, and I'm curious to see what the actual overall impact in 2018 will be.

But that is a very healthy dynamic, in general.

Similarly, if you look for later stage, whether it's therapy selection or monitoring, again, that brings NGS into segments of oncology where there was no NGS before.

And oncology remains one of the big market opportunities where NGS can really make a difference.

And so I think it's all incremental in terms of driving our penetration into oncology.

And that remains a very big opportunity, we believe.

And our next question comes from Steve Beuchaw from Morgan Stanley.

Well, I'd first like to start on a bit lighter note and just say I'll reiterate the welcome to Jacquie.

It's great to have you back.

Thank you.

I appreciate that.

My first topic, on the more serious note, comes back to the very popular topic of sequencing consumables.

I wonder if we could do 2 things.

One, can we do just a little bit of retrospective?

As you went back and looked at the trends in the first half of the year, I mean, do you still think that there was a significant amount of the consumables destocking in the first half of the year?

And have you given any more thought to how big an impact that had?

And then sort of a related question here in the third quarter also on sequencing consumables.

I mean, it seems like there are a number of different constituents who had pretty significant ordering patterns.

Ordering patterns above trend in the quarter, setting aside the obvious NovaSeq stocking.

I mean, what would be the binding thread?

And why would the growth across these different constituencies be simultaneously so strong?

And then I have a follow-up.

Sure.

We did anticipate coming into the year that, as we launch NovaSeq, you would see some destocking.

And that's what played out in Q1.

And if you look at it, it's primarily the handful of X customers potentially that we felt would destock as they looked at NovaSeq.

If I look at the consumable growth -- the sequencing consumable growth that we saw in Q3, it was across the board, across every platform that we had.

And so I think when we started to break down, okay, could the destocking have played a factor here?

It certainly could have and it likely did probably maybe in one of those platforms.

I think for HiSeq X, it's possible for example, or HiSeq.

But if you look across the board, I said, we saw sequencing consumable growth across every single platform.

And the story -- there are stories around each segment of the market and why that's growing.

For example, we continue to see our clinical commercial customers grow.

And that's driven by growth in samples in the markets that they're addressing, whether that's in oncology or in NIPT.

And that's starting to flow through, for example, in NextSeq, or in MiSeq or it could be some in HiSeq.

And so what's happening is, across the board, we're seeing demand grow, and that's starting to show up across our different platforms.

In X and in NovaSeq, you're starting to see the dynamic of NovaSeqs ramping up.

And so it's a little bit potentially of the destocking that's playing out in that part of the market.

But again, I think across the board, what we're seeing with our customers is they're seeing demand, and that's driving them spending more on consumables.

Okay.

I really appreciate all that color.

The follow-up is much, much simpler, and it's on microarray service.

It's been a really good driver of growth this year.

And we know, of course, there's some good success stories behind that.

What do you think is a medium-term reasonable growth rate for microarray services given everything you see going on out there?

It's definitely been a good growth story for us, and there are a number of drivers from it.

A big driver, obviously, has been the direct-to-consumer market, those customers that continue to exceed their own plans for growth, and that's then translated into the demand that you're seeing from us.

But there are other segments that are playing out, too.

For example, agriculture is a very interesting segment and continues to drive demand for our microarray products.

The -- we expect to see the demand continue to grow from their customers' perspective.

You talk to ag customers, they're looking to expand either the types of species that they're looking to sequence.

They're looking to actually expand the production sequencing that they do.

If you talk to our DTC customers, they believe that they are still in the very early stages of this big consumer genomics market.

And so from a demand perspective, our customers will tell us that if you look out into the horizon, they expect demand to continue to grow, and that we're at the very early stages of those markets.

We're also, though, talking to them about moving to sequencing.

And if that happens, obviously, then you'll see a decline in the microarray line.

But that's really good news for customers, and their customers, because that means they are getting more value from the sequencing.

And it's obviously terrific news for us at Illumina.

It's hard to predict we're already seeing sequencing in the ag market, for example, and that's already starting to play out.

So it's hard to know how much of this microarray demand will move to sequencing over time.

And obviously, we'll do everything we can to encourage that.

And then the same thing on the DTC side.

It's hard to predict if and when those markets will move to sequencing.

And when that happens, obviously, you'll see microarrays come down and you'll see sequencing go up, which, again, is terrific for us.

And the next question comes from Dan Leonard from Deutsche Bank.

First question on the NovaSeq orders in the quarter, the 70 you mentioned.

How are you thinking about that from a run rate perspective?

Is that a pretty good run rate?

Or are you thinking that, that should accelerate or decelerate as you progress through in the launch here?

So it's hard to call a specific run rate.

I mean, we've been on the market now for these 2 quarters not fully quite 3 quarters yet, and each one of them has had their own dynamic.

Obviously, there was the launch quarter.

There was a quarter when we were constrained by manufacturing, and this is the first quarter we're exiting without being constrained by manufacturing.

And so I don't think we're at the stage yet where we can call an order run rate.

I will say, though, that as we looked -- coming into this year, we looked at the curves -- the order curves by quarter for some of our previous instruments, including the X. And what we're seeing with NovaSeq is consistent with what we saw with the X. So at this stage, it's not yet as a regular run-rate basis, but the curves are similar to what we've seen before.

It will take a few more quarters, I think, before we get to a steady-state of orders per quarter.

And I would just add, exiting Q3 as well, we have a very strong back flow -- backlog, I mean, in terms of NovaSeq instruments, which has built up because of some of the production constraints that we had, which we are now past.

Okay.

And for my follow-up, I want to touch on the 35% growth in shipments to the clinical market.

Francis, you said that was largely from liquid biopsy.

Can you comment from other segments of your clinical market, like tumor profiling or germline genetics?

Are you seeing anything in the customer base around the prior authorization requirements for genetic testing?

Is that causing any hesitancy in that market?

Or is that not a factor?

Look, we believe that prior authorization will add friction into the market as a whole.

And so over time, that's something we want to help the industry work through.

It hasn't showed up in terms of impacting our demand at this stage.

And when I talked about the growth in the clinical markets, one element certainly was the translational liquid biopsy studies that are going on.

But we talked about the growth we're seeing in oncology testing as a whole, even beyond liquid biopsies.

We talked about the growth that we're seeing in NIPT in terms of the covered lives continuing to expand, both here and internationally.

And so there are a number of factors that are driving that 35% growth, not just the large liquid biopsy studies.

And the next question comes from Puneet Souda from Leerink Partners.

So could you give us a sense of if there's any change in the timing of the $100 genome or reaching that price?

Clearly, liquid biopsy has taken off here in the market with NovaSeq already, so -- and I asked that because, at ASHG, we saw R&D presentations where clinically diagnostic genome is more of a near-term reality, and so the pediatric labs and pediatric researchers are highlighting elements that are missing in exomes.

So just wanted to get your thoughts on the timing of that genome.

Yes.

I'll start by saying that I saw some of those things at ASHG, too, and I think it is exciting to see the momentum that's building around the value of whole-genome sequencing.

I was talking to a customer at ASHG who said -- look, it's clear to them that even if you wanted to be the best exome you could, the way you do it is you do a whole genome, because you actually would find variance that you wouldn't find if you did whole exome sequencing, which is an interesting conversation.

And I think from our perspective, it's exciting to see the momentum building around whole-genome sequencing.

From a $100 a genome perspective, we continue to believe that, that is attainable with the architecture that we have in NovaSeq.

It is a multiyear goal, but we have line of sight into the technology improvements that get us from here to there.

The thing that we are also working on then is trying to catalyze the sequencing-intensive application demand so that we can continue to drive prices down, and it will be met with a demand that will be overall additive to the market size.

And so we have work going on from a technology perspective to enable that, but the rate at which we step the price down will be equally informed by our belief that the market -- the elasticity of the market is there to extend it at each price point that we take down.

We think that S4, for customers that reach the highest volumes, do give you access to prices that are lower than you have got with the X. And we believe the market is ready for that.

And so we think the elasticity is there by taking the prices per gene, so therefore the prices per genome, down below the X. It overall expands the markets.

So that's first step towards the $100 genome, as we put out S4 right now.

And we'll continue to step it down, knowing that we can get there with this architecture.

But it will be gated somewhat by -- in our belief that the market is there at any given point for the price point we put out.

Okay.

And a quick question on the consumables.

Just trying to understand, so if we look at the theoretical capacity of HiSeq 2500, 4000s and HiSeq X and the theoretical capacity of NovaSeq for an entire year, it seems to me that number of permutation combinations, whatever you can try from S1 to S4, the consumable pull-through is going to be significant and likely to be higher than HiSeq X or NovaSeq in the long run.

Have you given thought to, I mean, essentially, how that could potentially shake out longer term?

And then any sense, any color there would be helpful.

So the -- sorry, the question you're saying is that -- let me make sure I understand it.

So you're saying, look, if you looked at the high throughput part of the portfolio, what is the dynamic that plays out as they move to NovaSeq?

Is that the question that you're asking?

Yes.

So if you have the theoretical capacity and the practical capacity, which we have known in the past as -- for HiSeq, it was utilized at maybe -- HiSeq 2500, maybe utilized at 20%, 25%, whereas HiSeq X was utilized at a more higher utilization rate, maybe potentially 50% when you average it across the world.

So I was just trying to get a sense of when you -- how are you thinking about the practical utilization of NovaSeq when you look at the theoretical versus the practical capacity?

And any color there would be helpful.

Essentially, trying to get to NovaSeq consumables in the longer run, if you have any color there.

Yes.

So there are a number of things that are going to play out.

One thing that we believe with NovaSeq is that it will push the elasticity of the market.

And that by enabling customers to have access to lower price points, they will go after larger cohorts.

So if you are a high throughput customer, if you are an X customer, for example, you will look to do larger experiments.

You will look to do single-cell experiments.

You will look to do liquid biopsy experiments.

And some of those you couldn't do on the X, by the way, because it was constrained to whole genomes, right?

And so one thing that I think will drive elasticity -- sorry, utilization beyond just moving your workflows from an X, for example, to NovaSeq, is that part of that transition for a number of our customers includes a contemplation of workloads that they wouldn't do before without a NovaSeq.

And that's true whether you're an X customer moving to NovaSeq or a 2500 customer moving to NovaSeq.

Our -- as we think about the market, we think sort of overall, we talked about the fact that utilization is in the 50-ish percent and has been sort of approaching that now for a number of years and fairly stable.

And so we're never designing to say what we want to have them running at 80%, 90%.

And if that's happening, it would indicate to us maybe that there was too much centralization of the market.

And so the other dynamic that NovaSeq enables in the market is it democratizes access to high throughput sequencing.

And so you're seeing customers that were outsourcing a lot of their samples and, in some cases, and they're looking to move those in-house.

And so the other dynamic that we're targeting with NovaSeq is to democratize access to high throughput sequencing.

And your last question comes from Jack Meehan from Barclays.

I just want to start on the much-discussed NGS consumables.

If I take Sam's fourth quarter guidance, that would imply low-teens level of growth for the full year.

But I think that fully encapsulates the inventory stocking issues.

Just compared to prior years' deceleration, can you give your thoughts on just the level of sequencing that's taking place?

And do you think this can reaccelerate in 2018?

Hard to comment on 2018 right now.

I think you're in the right ballpark there in terms of your expectations.

But again, I go back to some of the factors that I mentioned earlier, which are really the important fundamental factors for us to look at here looking forward, and that is as you look forward in terms of the progress or the growth sequentially on sequencing consumables, what you would expect, and it's early right now to comment definitively, but what you would expect is you would see NovaSeq to continue to increase in terms of the level of consumables quarter-over-quarter.

Had a strong quarter this quarter, but will continue to see that going forward, especially now also with the S4 launch.

What you will see on HiSeq is that you will see it start to come down, and what you will see on HiSeq X is also eventually it will start to come down, but obviously with some lumpiness.

What you'll also see, and what we saw this quarter, is that NextSeq also will experience or has experienced pretty strong sequential growth in Q3, and we would expect to see growth in NextSeq consumables as well.

So I won't give a specific figure in terms of what the 2018 expectations are, but those are the dynamics shaping our sequencing consumables' sequential run rate.

That's helpful, Sam.

And then just want to dig in on one component that the HiSeq, how that's going to trend, can you just give us color in terms of the replacement cycle, how long you think that validation takes for NovaSeq?

Just trying to pare the commentary for the quarter, HiSeq stepped up and some of the more limited replacements you've seen so far.

Yes.

I think, so when we talked about the HiSeq sequencing consumables, we said, look, that the -- we expect that to go down over time.

And the dynamic we saw play out last quarter was we had these large clinical commercial customers that were seeing growth in their business, and that drove the demand for HiSeq consumables last quarter that resulted in overall HiSeq consumables growing.

We expect that to be a temporal dynamic, and we expect that, over time, there will a cutover.

Now the validation point -- you know, with every customer the validation at NovaSeq is unique to them.

And one customer is telling me, in 2 days, they were ready to go.

And other customer cases it could take months for them to do the validations they need, and it's driven a little bit by whether they're a research customer or clinical customer.

But it's very much driven by the practices of that lab.

That, I think, is not going to be the gating factor of how the HiSeq consumable bucket changes.

I think it's going to be driven by the dynamics of their business.

And so if you have customers that have large experiments that are running or that have extensive validation practices that will determine how quickly they move over.

And this concludes the question-and-answer session.

I'll now turn the call back over to Jacquie Ross for closing comments.

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 fourth fiscal quarter.

Thank you, ladies and gentlemen.

This concludes today's conference.

Thank you for participating.

You may now disconnect.