Illumina Inc (ILMN) 2006 Q2 法說會逐字稿

Good day, ladies and gentlemen, and welcome to Illumina's second quarter earnings conference call.

At this time I would like to turn the call over to Bill Craumer, Director of Corporate and Marketing Communications at Illumina.

Please go ahead, sir.

Good afternoon everyone, and welcome to our Q2 conference call.

At the close of market today we released our financial results for the second quarter of 2006.

During this call we will review our results, update our financial guidance for fiscal 2006, and provide commentary on commercial activity, after which we will host a Q&A session.

Presenting for Illumina today will be Jay Flatley, our President and Chief Executive Officer, and Christian Henry, our Vice President and Chief Financial Officer.

The call is being recorded and the audio will be archived on our website at Illumina.com.

During the call we will be updating our financial guidance and discussing plans for future activity.

Our intent is for these forward-looking statements to 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 information available, and Illumina assumes no obligation to update these statements.

To better understand the risk factors, we refer you to the documents that Illumina files with the Securities and Exchange Commission, including Forms 10-K and 10-K.

I will now turn the call over to Christian.

Good afternoon everyone, and thank you for joining us today.

I hope that you've all had a chance to review our financial results that were released after the close of market.

I would first like to walk you through our operating results for the second quarter, and then follow with an update of our financial guidance for fiscal 2006.

I will then turn the call over to Jay for a discussion of the Company's progress.

For the second quarter we are pleased to report our 20th consecutive quarter of revenue growth.

Total revenues were 41.6 million, which represents a year-over-year growth of 163%.

On a sequential basis, total revenue grew more than 43% over Q1 level.

On a year-over-year basis, Q2 product revenue grew 188% to 36.4 million.

This growth was driven by greater than 200% growth in both consumables and instrumentation.

Consistent with the first quarter, consumables represented nearly 50% of total revenue.

During the quarter we shipped 26 BeadStations and 2 BeadLabs, bringer our total cumulative system shipments to 178.

Many of our Infinium genotyping installations, including two booked this quarter, are substantially larger in value and throughput than the traditional BeadLab.

This blurs the distinction between these configurations.

Therefore, we believe that in the future it is more valuable to report our total number of BeadArray Readers shipped, rather than separately breaking out BeadLabs.

Our new cumulative shipment for comparison going forward is 181 BeadArray Readers shipped, taking into consideration the fact that several of the BeadLabs included multiple scanners.

In addition to the growth in BeadArray Reader placements, we had substantial shipments of our AutoLoader and other automation instruments during the quarter.

The increase in automation-related sales has increased our customers' throughput per system, enabling us to achieve annualized consumable throughput per system of more than $500,000 per BeadArray Reader in the quarter.

Finally, on a year-over-year basis we saw all oligo revenues grow more than 25% in the second quarter.

Service and other revenue increased 72% for the quarter to 4.8 million compared to 2.8 million in Q2 of last year.

This number includes revenue from our genotyping services business and revenue earned from instrument maintenance contracts.

Before discussing our operating expenses for the quarter, I would like to describe the effect of FAS 123R, which requires us to record the expense associated with stock options in our income statement.

The total impact for the P&L this quarter was approximately 3.3 million.

And as prescribed by the standard, the expense is allocated to each operating expense line.

We have not restated our historical financials for the impact of expensing these options.

In the discussion that follows I will discuss both our GAAP expenses, which includes the effect of FAS 123R, and the corresponding non-GAAP figures.

I urge everyone to review the schedule included in our earnings release that reconciles the non-GAAP figures with the GAAP figures.

Cost of products and services revenue was 13.6 million for the quarter compared to 4.7 million in Q2 of 2005.

This cost includes stock-based compensation expense of 350 K. Non-GAAP product and services gross margin was 57.9% for the quarter compared to 68.3% last quarter, and 69.3% in the second quarter of 2005.

As we previously indicated, product and services gross margin will vary primarily on the product mix.

In the second quarter revenue was more heavily weighted towards instruments, which had a small unfavorable impact on gross margin.

Research and development expenses were 8.6 million in the quarter compared to 7.3 million in the second quarter of 2005.

Of the $1.3 million increase, approximately 860,000 was related to stock compensation expense.

The remainder of the increase was due primarily to the increase in expenses associated with the development of the BeadExpress System, which is scheduled for launch later this year.

SG&A expenses were 12.9 million for the quarter compared to 6.5 million in the second quarter of 2005.

Of the 6.4 million increase, 2.1 million was related to stock compensation expense.

The remainder of the increase is related to the expansion of our commercial activities, which were required to support our revenue growth and increased spending on litigation.

For the second quarter we reported net income of 6.8 million, or 16% of revenue.

On a fully diluted basis, this represents earnings of $0.14 per diluted share.

This compares to a net loss of 18.5 million, or $0.46 per diluted share, in the second quarter of 2005.

To remind you, the net income in the second quarter of 2005 included a charge of 15.8 million, or $0.39 per share, related to our acquisition of CyVera Corporation.

Excluding the impact of stock compensation expense, we are pleased to report non-GAAP net income of 10.1 million, or 24% of revenue.

On a fully diluted basis, this represents earnings of $0.21 per share.

Turning to the cash-flow statement, we were cash-flow positive during the quarter, generating total cash of approximately 107.7 million.

Of this total, 96.5 million was generated by our follow-on offering that we completed in May.

The remainder was generated through operations, offset primarily by capital acquisitions.

During the quarter we generated cash flow from operations of 14.9 million compared to 2.4 million in the first quarter.

The significant improvement in operating cash flow was due primarily to record net income, strong cash collections, and the receipt of approximately 5 million in customer prepayments for future product shipments.

In the second quarter we purchased 5 million in capital equipment, giving us total free cash flow of approximately 9.9 million, or $0.21 per fully diluted share.

Moving on to the balance sheet, we ended the quarter with 157 million in cash and short-term investments.

Accounts Receivable DSO improved for the third consecutive quarter.

DSO for the quarter was 50 days compared to 67 days at the end of the first quarter.

Going forward we expect DSO to continue to fluctuate from quarter to quarter based on our geographic mix of revenue.

Inventory grew by 4.4 million in the quarter from 12.9 million at the end of the first quarter to 17.3 million at the end of the second quarter.

This change was primarily related to instrumentation inventory, as we prepared to ship a large system configuration to deCODE Genetics in the very early part of Q3.

Additionally, we increased our inventory of reagents to satisfy the significantly increasing volume of product shipments.

I will now update our financial guidance for the remainder of the year.

Consistent with our previous call, the following guidance excludes the impact of FAS 123R.

I will provide an overall estimate of the impact of 123R at the end of my remarks.

For additional detail, please refer to the table in our earnings release that reconciles our non-GAAP guidance to the related GAAP figures.

Let's talk about our view for the second half of the year.

We continue to see strong demand for our products and services as the particularly robust growth in the whole-genome genotyping market.

Given these market conditions and our performance in the first half, we now expect annual revenue to be between 160 and 170 million, which is an increase of $30 million over our increased guidance that we provided last quarter of 130 to 140 million.

Our updated revenue guidance represents growth between 118 and 131% over 2005.

Research and development expense is expected to range between 30 and 35 million for the year, while SG&A expenses are expected to range between 42 and 47 million.

We expect to generate quarterly net income in the range of 20 to 25% of revenue in each quarter during the remainder of 2006, with total annual net income between 30 and 40 million.

On a fully diluted basis this would yield earnings per share of $0.61 to $0.82, assuming weighted average shares of approximately 49 million.

We're continuing to forecast total stock compensation expense to range between 13 and 15 million; however, this estimate is highly dependent upon changes in our underlying stock price.

For the year we now expect to generate at least $25 million in cash flow from operations.

Additionally, we expect capital spending to trend to the high end of our original estimates of 15 to 20 million, primarily due to the continued expansion of our manufacturing capacity.

In the third quarter we expect total revenue to range between 44 and 48 million.

Excluding the impact of the expense associated with stock options, net income per share should range between $0.22 and $0.24 per diluted share, assuming weighted average shares of approximately 50 million.

At this point, I will turn the call over to Jay for additional comments before we begin the Q&A.

Good afternoon everyone.

What a quarter.

As Christian pointed out, we achieved revenue growth that significantly exceeded our expectations, maintained healthy gross margins, controlled our expenses, and delivered 24% after-tax income.

At this point we believe we can continue to generate operating income of at least 20% of revenue per quarter through the remainder of 2006, which has always been our long-term goal.

Additionally, we generated nearly $10 million in free cash flow for the quarter, which is the result of the efficiency with which we operate our business.

Our significant revenue growth is being driven by the continued rapid expansion of the whole-genome genotyping market.

During the quarter we shipped more than 40% more BeadChips than in the first quarter.

Also during the quarter average prices for our whole-genome genotyping products were stable when compared to the first quarter.

Finally, although we far exceeded our own revenue forecasts, we exited the quarter with a record backlog more than 85% larger than the next nearest quarter.

These results reinforced the view that the market for whole-genome genotyping is growing at a rate that far exceeds our original forecast and exceeds that of most any third party.

In the second quarter we signed several large deals that further demonstrate the accelerating growth in this market.

First, I would like to highlight our agreement with deCODE Genetics that we announced in May.

We're currently in the process of installing our largest system to date at deCODE, including seven scanners and ten robots.

Once installed, deCODE plans to use our HumanHap300 to perform whole genome association studies on its Icelandic founder population, including more than 100,000 participants covering some 50 common diseases.

It is expected that this effort will enhance deCODE's proprietary gene and target discovery programs, and may provide additional genetic markers for diagnostic development.

Additionally, we entered into a strategic alliance with deCODE to co-develop and commercialize DNA-based diagnostic tests in several major disease areas.

DeCODE has shown that carriers of particular SNP variance have a significantly higher likelihood of developing particular diseases.

The goal of the alliance is to develop diagnostic tests for these SNP variance in three major disease areas.

Under the terms of the agreement Illumina has gained access to disease-related biomarkers for joint validation as diagnostic panels to be marketed and sold on our forthcoming Bead Express platform.

The company will share development costs and split the profits from sales of diagnostic tests.

Initially we plan to focus on the development and validation and commercialization of diagnostic tests for heart attack, Type II diabetes, and breast cancer.

We will update you on the progress of this alliance over the course of the next several quarters.

The second significant deal I would like to highlight is the deal we announced in early June with The Children's Hospital of Philadelphia, or CHOP.

In this landmark deal CHOP purchased a large-scale system capable of producing 150 million genotypes per day.

The objective of the researchers at CHOP is ultimately genotype up to 100,000 children.

The knowledge gained from these studies is expected to provide significant insight into the genetic basis of common childhood disease, linking genetic variations with specific illnesses.

The results of their research are expected to guide therapy and drive development of new treatments based on the concepts of personalized medicine.

Including arrays and reagents for the first 30,000 samples, the initial value of the agreement will exceed $25 million, of which approximately $4 million was recognized in the second quarter.

This agreement can be viewed as an early indicator of the ultimate size of the genotyping market.

As researchers continue to improve our understanding of how genetic variation affects human health, it is our belief that genotyping will become the standard of care for screening every child born.

During the quarter we continued to expand our product portfolio.

In June we announced our HumanHap650Y BeadChip, which contains over 650,000 SNP markers on a single microarray.

This product provides the most comprehensive genomic coverage and highest data quality of any product currently available.

The Hap650Y is also a great example of the flexibility of the BeadChip platform.

The product was designed by first taking our commercially available Hap550 BeadChip and then adding an additional 100,000 SNPs targeted toward improving the genomic coverage of more genetically diverse populations, such as the [Aucuba].

These incremental 1,000 SNPs were developed in less than one quarter.

This demonstrates the efficiency of our development process which enables us to rapidly respond to the market's demand for various BeadChip products focused on different target populations.

Last week we announced the availability of two new BeadChip products, the HumanHap550 Plus and the iSelect Infinium genotyping products.

The Hap550 Plus allows customers to add up to 120,000 custom SNP markers to supplement the standard content provided on the existing Hap550 BeadChip, yielding up to 670,000 markers on one chip.

I would like to point out that in less than one year we've increased the number of SNPs on a chip nearly sixfold from 100,000 to 670,000, which in turn has created significant value for our customers.

Our iSelect Infinium productline is ideal for focused content applications.

With this product we can now create a custom array for customers, who can select up to 60,000 SNP markers per sample, with 12 samples per chip.

This multi sample offering is ideal for researchers who have already performed whole-genome genotyping studies, or have narrowed down SNPs of interest to a large yet focused set of markers relevant to the particular disease that they are studying.

In order to enable customers to more easily design these custom BeadChips, we have developed a powerful bioinfomatics tool that enables in silico SNP design and screening.

As with all of our multi sample offerings, researchers will also benefit from very favorable economics on a per sample basis.

We have already seen significant interest for this product, and note that Johnson & Johnson will be using this product through our services group to perform studies related to their clinical trials.

At this point as we look back on the nine months since the launch of our whole-genome genotyping products, I would like to take just a minute to highlight a few of the fundamental reasons why we believe our genotyping solutions are successful in the marketplace.

First, the development of the Infinium assays enabled us to select and deploy on our chips the most valuable and relevant content from the results of the International HapMap Project.

In the case of the Hap550, this means that with 550,000 SNP markers, we can obtain over 90% genomic coverage of the HapMap SNPs in Caucasian populations.

The Infinium assay also provides researchers with a very friendly workflow, reducing the labor required and thus the overall costs required to perform a SNP genotyping experiment.

The simplicity of the assay also has the added benefit of reducing experimental error.

Second, the flexibility of the BeadChip format provides possibilities that no other commercial array can offer.

Our custom, semi-custom, and multi sample formats, combined with our low cost oligo manufacturing process enable us to produce extremely competitive products.

Whether the researcher wants to steady 5,000 SNPs in 12 samples per chip, or 670,000 SNPs in a high-density single chip, single sample format.

Third, our automation and LIMS solutions enable researchers to increase the throughput of their laboratory and improve their ability to track samples without having to significantly increase their labor requirements.

This is extremely useful with whole-genome genotyping, as the study sizes tends to be much larger than in other types of microarray experiment.

These advantages, along with a manufacturing process that insures 100% quality control of every chip we ship, enables our genotyping products to have the highest performance in the industry in terms of call rates and reproducibility.

Turning now to operations, we have continued to expand our BeadChip manufacturing operation in our San Diego facility.

In our last call we set a goal to double our capacity from the April levels by the end of Q3.

I'm pleased to report that we have already achieved this goal.

More importantly, we now believe that we have the infrastructure in place to efficiently add additional manufacturing capacity as required to support our anticipated chip volumes.

During Q2 we made several improvements to our basic manufacturing processes, which effectively increased our capacity without incremental equipment.

I would like to now provide an update on our progress with the BeadExpress system.

As I indicated last quarter, the BeadExpress system is the reader that will be launched as a powerful system for low to mid multiplex applications.

This system will have important advantages over other platforms in both the research and molecular diagnostic markets.

In fact we plan on using the BeadExpress Readers as a key component for deploying diagnostic content in our strategic alliance with deCODE.

We have now manufactured the beta units, and the first beta reader will be shortly delivered from our operation in Connecticut to our San Diego facility.

In the second half of the year we will complete the beta testing program, as well as the assay integration and bead manufacturing programs, enabling a launch before year end, as we have previously communicated.

Initial customer response has been overwhelmingly positive, and I look forward to keeping you apprised of our progress as we get closer to launch.

I want to take just a minute to update you on the activities associated with our litigation with Affymetrix.

As many of you know, a Markman hearing was held on April 20.

In a Markman hearing the judge hears arguments by both parties on their respective definition of key terms in the patent claims.

There has been no ruling from the hearing as yet, but this ruling can come at anytime now.

The trial is currently scheduled to begin on October 16 of this year.

We believe that we have strong defenses and very strong counterclaims, and we will continue to vigorously defend ourselves.

From a commercial perspective, we continue to grow our infrastructure.

During the quarter our non-GAAP SG&A spend was 26% of revenue, which as a percentage is lower than our long-term models.

We now have nearly 100 people supporting customers in our sales, marketing and technical support organizations.

We've also make key distance additions to our management team.

In late May we were pleased to announce that Christian Cabou had joined us as Senior Vice President and General Counsel.

Chris was most recently General Counsel of GE Global Research, the research arm of General Electric Company.

We are already leveraging his deep understanding of intellectual property issues in many areas of the Company.

In addition to Chris, we were pleased to announce the appointment of Dr. Jack Goldstein to our Board of Directors.

Dr. Goldstein was most recently President and Chief Operating Officer of Chiron Corporation until its acquisition by Novartis in April 2006.

Jack brings a wealth of experience in the diagnostics market, and will be instrumental in guiding us as we commercialize diagnostic products on our BeadExpress platform.

In summary, our revenue growth in the first half of the year was driven by supplying an incredibly compelling solution into one of the most dynamic markets that I have seen in my career.

We believe that the genotyping market is still in the early inning, and that the market will continue to grow as pharma and biotech companies adopts whole-genome genotyping in all aspects of their research and clinical development programs.

Our profitability and free cash flow was driven by efficient execution, and our ability to scale our manufacturing infrastructure to support the explosive growth in our markets.

In the second half we look forward to launching additional compelling products and continuing to work with our customers to assist them in pursuing their scientific goals.

We're now ready to open the call for questions.

Thank you.

(OPERATOR INSTRUCTIONS).

Edward Tenthoff with Piper Jaffray.

Thank you very much, and let me be the first of what I assume will be many people to congratulate you on just another phenomenal quarter.

I hate to shift focus from what you guys are doing to what your competitors are doing, but I just have to ask.

With the announcement after the close of Affymetrix in terms of slashing the prices of the two chip 500 K, can you make a comment in terms of does that have any impact for your guidance in the back half or your general business plan for the back half?

And can you specifically give us what your pricing is currently for the 550 and the 650?

No, that announcement doesn't affect our guidance going forward.

We feel very comfortable with the position of our products in the market and the pricing of those products.

We have continued to be able to command a premium price over the competitive products in the market because of the quality of the content and the ease-of-use of the assay.

And our customers just love the performance of our technology in the field.

We don't see any direct impact on that.

Our longer-term models of course do incorporate the idea that prices over time will go down in the marketplace, and we have been instrumental in driving those prices down over the last five years by something that is approximating a factor of 1,000, and have generated enormous profits even while doing that.

In terms of the pricing of our chips today, the list prices, depending upon which chip you're talking about, range from $800 up to over $1,500.

And then based on a volume discount schedule, we grant discounts off of that based on the number of chips that customers submit in a single purchase order.

Well, again, congratulations on the quarter.

I will get back in line.

Eric Schmidt with SG Cowen.

Congratulations also today.

I was just wondering if you'd give us some sort of sense in terms of your manufacturing capacity on the percent utilization you are running up against, and what are the -- any risk at all to your ability to [tie] the market to -- the market that has continued to ramp up at a feverish pace?

We have done I guess what we think is a pretty good job of managing our capacity right near 100%.

And we're able to do that because of the way we add capacity into manufacturing.

We add it in small increments.

We add it every week.

And we do that based on our visibility of how many chips customers want at certain time points in the future.

Because we ship to customers on a week by week basis, so if somebody puts in an order for 3,000 chips we would spread the deliveries of those chips over the timeframe they want to the project, and layer that into our backlog.

What it gives us is a nice snapshot 8 weeks out, 12 weeks out, 16 weeks out as to how many chips we need to deliver.

And then we can modulate our capacity by either increasing or slowing down the rate at which we add decoders into production.

Our goal is to keep that somewhere at around 90% plus utilization.

And we have been real successful doing it to date and don't see any risk going forward.

And then on the diagnostic side, you mentioned some -- a positive customer response.

Just wondering if you could provide a little bit more color as to who the initial target customers are going to be?

And a little bit more detail on your commercial strategy here, how you intend to rollout these products, what assays they might be running, things like that?

The BeadExpress system, as I mentioned, will be launched before the end of the year.

The initial customers will include our existing research customers.

And we will place this in many of those sites as a direct add-on to their existing systems to enable them to continue to use Illumina technology as their studies whittle down the number of markers, whether it is in an expression experiment or in a genotyping experiment.

[Seven], we expect to in a serious way push this into the clinical diagnostic market, initially into third-party laboratories that may be using it to do home brewed type applications, and then ultimately with proprietary markers sets that come from Illumina in conjunction with our partners.

Certainly in deals like what we have done with deCODE those products will contain proprietary content that we will develop and validate along with deCODE.

And it is likely that products that will be required to go through the FDA will be more in the 2008 type of timeframe at the earliest from us.

But we will be building the installed base in the clinical environment during 2007.

Nate Cornell with Pacific Growth Equities.

Congratulations guys.

It was a very impressive quarter.

What is really driving the growth -- the rapid growth in the whole genome genotyping business?

There are a number of factors.

Certainly there has been an enormous desire in the academic research groups to do large-scale disease association studies.

That is a key driver that has pushed this market forward so robustly in the last 18 months or so.

For the vendors in this market the good news there of course is that there is a lot of competing research groups that want to do these studies on the same disease.

There are lots of diseases to study.

And because genotyping is a statistical science, you need to deal with very large populations to get statistically significant results.

So all of that drives very large studies and a large market opportunity.

If we look forward, we think these disease association studies are going to take many years to get completed.

And so that part of the market by itself is going to be around we think for three to five years.

And then layering on top of that is going to be what we think will be rapid uptake in the pharmaceutical and biotech industry of genotyping in the clinical trial environment.

That market alone we think could ultimately be a $500 million market.

And we certainly see the beginnings of that adoption now.

We're dealing with quite a number of pharmaceutical companies.

We have announced a blanket contract that we have with GSK, and we're doing significant amount of genotyping work for them under that blanket contract.

We've also installed BeadStations at a number of other pharmaceutical companies.

In addition, every time a whole-genome study get completed there is a follow-on study that is a fine mapping study.

In other words, they find regions of the genome where they get hits, and then they need to dive -- take a deep dive into those regions of the genome looking at virtually SNP in that region.

And that is where products like our GoldenGate genotyping come in, as well as the new iSelect Infinium genotyping products that have the ability to put onto a chip a custom set of SNPs that the customers select.

As I mentioned in my remarks, ultimately we think genotyping is going to become a screening and diagnostic application; one that is done on every child that gets born.

And the question is whether that panel of SNP markers is going to contain a few hundred markers or a few hundred thousand markers.

No one knows yet.

But I think that evolution speaks to a very long-term robust market.

Is your technology or your platform being utilized right now in clinical trials -- ongoing trials?

It is, yes.

Just to follow-up on the BeadExpress.

Have you identified the assay that you would submit in conjunction with the instrumentation to the FDA?

Is that something that will come from deCODE?

The assay method will be our genotyping assay method.

We will deploy the product initially with low multiplex and mid multiplex genotyping applications, as well as some demonstration applications that will include things like proteins analysis -- protein marker analysis.

So it is a very versatile platform, and over time we will add lots of assay capability to it.

And customers in addition will have an ability to develop their own assays on the platform.

In terms of the menu, is it safe to assume that we are going to see the deCODE markers first?

I would say that is probably not the case, because that one is a proprietary step that is going to require validation and ultimately a path through the FDA.

That one probably is a bit further out, or that set of markers in those three diseases are a bit further out.

I think what you'll first see are our other tests that we will be putting into the market in 2007, either by dealing with partners who already have proprietary intellectual property around certain markers, or in the case of some of the more commercial labs, them developing home brewed assays on a platform on their own.

May-Kin Ho with Goldman Sachs.

Obviously the growth of the market is very rapid.

But previously I think the estimate is that for the genotyping market is about a $250 million market, growing at 30, 40% a year, and expression market is 800 million growing at about 5 to 10% a year.

What is your latest estimation now?

I think the expression market, I think the numbers are, if anything, a little on the high end.

So the growth rate may be a little slower in expression than 5 to 8%.

In the genotyping market, in the segment that we participate in, we think the growth rates are higher than 30 to 40%.

There is a part of that that today we don't play in, which is the very low multiplex genotyping market that is addressed by systems like Packman as an example.

And over time as we launch our BeadExpress platform that market will become available to us as well.

But the part that we play in today, the custom genotyping, and the whole-genome genotyping is probably growing in excess of 50% per year.

Can you talk a bit about the long-term view in terms of diagnostics, because that obviously can be a very large market.

Do you intend to eventually set up your own sales force there?

How do you envision the business growing in the next five years or so?

It is likely to be a hybrid model, and we will reveal more details about that as we get closer to market.

But the elements of that model will certainly include dealing with some partners who may take the instrument to market themselves.

And in that case, they may put their assays on our platform and OEM the platform from us, and sell it themselves.

In other cases where we have Illumina proprietary content, we may work through partnerships or may do that directly.

And we're in the process of making some final evaluations on how to do that, and what the best model is for Illumina going forward, so I would ask you to stay tuned on that one.

For the instrumentation that you have now, do you think that you have to do a lot of improvements in order to have them used in diagnostic labs?

No, it has been designed from the ground up to be used in diagnostic labs.

John Sullivan with Leerink Swann.

A couple of quick questions.

First of all, with your current array architecture, how many SNPs do you believe that you reliably place on a single array today?

We're shipping a product that has 650,000 SNPs on them today with very high redundancy.

That is commercially the product that we have on the market.

But as we have consistently said, we believe the technology has lots of more headroom.

We're nowhere near the ultimate capacity of this technology.

As you consider the possibility for expanding the number of SNPs on an array in future products are you constrained by the lack of relevant biological content by any chance?

At some level of course you are.

If you put all 10 million SNPs on a chip, then you are pretty much done.

We're not quite at the point where that is a limitation today.

What in some sense is a limitation is that if you look at the kind of coverage we get with our existing chips -- let's take the 550 chip as an example -- we get 90% of all the genomic coverage from all the HapMap SNPs that were analyzed in a Caucasian population on that chip.

No matter how many additional SNPs we added, the additional coverage would only be 10%.

That begs the question about how do you expand the productline.

The logical evolution we took of course was to add a set of additional SNPs that turned into the 650Y product that addressed in particular the African, African-American population that have greater diversity.

In those populations the coverage on the 550 was significantly less than 90%, and so we added additional content to increase the coverage in those populations.

I think what you'll see us do in the future is continue to look at mixing and matching different kinds of collections of SNPs to address evolving market needs and different kinds of customer needs.

And the flexibility of the BeadChip allows us to do that very easily and very deftly because of the ability to use these stripes in different ways and to gasket the chip in different ways for either multi sample or single sample applications.

One more question.

As you think about growing the business with academic researchers, do you perceive that the NIH as an institution is becoming more positive on funding large-scale genotyping studies?

Yes, we think they have been all along.

The limit has been that whole-genome genotyping didn't really exist before 18 months ago.

And so there was no funding in that segment of the market because there wasn't the solution.

They certainly have been active for a longer time in doing custom genotyping.

And they have been a significant customer of ours for many years.

I think as you look forward, because of the power of the science now, that there is a dramatic increase in interest and in the funds that would be porting to do the large-scale studies going forward.

There are constituencies within the NIH that have wondered whether large-scale genotyping is really translational research.

Do you perceive that that battle is being won as new information comes out on the relevance of individual SNPs?

I think so.

There are certainly various opinions about how to best deploy this technology, and how to best utilize these different types of studies, and there will continue to be some divergence of opinion.

But certainly we think that the kind of productline that we offer can address any of the types of applications that those researchers would like to address.

Quintin Lai with Robert W. Baird.

Congratulations on a fantastic quarter.

As we take a look at the guidance raised for the back half of the year, how much of that comes from the new contracts like deCODE and CHOP, and how much of that is coming from new business that has just come through the door from when you gave your guidance from last quarter?

It is certainly a combination.

In the case of deCODE and CHOP, both pieces of the instrumentation system will hit -- the CHOP system will hit in the second quarter.

The deCODE system will hit in the third quarter.

Those are large instrumentation placements that weren't specifically in our guidance last quarter.

We layered in their chip requirements over the next 18 months to 24 months.

So through the end of this year they will be only of a portion of the consumables revenue from those accounts.

And so the remainder of the increase in our guidance comes from general robustness in the overall business and what we expect to happen in some of our new products.

From that new business that is coming in, any kind of color that you can give versus -- is a coming from commercial customers or from academic and nonprofit customers?

It seems to be coming from both.

But certainly we are seeing, and we have talked about for the last couple of quarters, a dramatic increase in interest from the pharmaceutical and biotech sectors to participate now in using this technology in the clinical setting.

We absolutely believe that is going to happen, and that it is on its way.

We are seeing our technology now start to be used actively in clinical trials.

We think that trend is going to continue and increase dramatically over the next few years.

In those types of studies, in clinical trials, do you imagine it is going to be a combination of both whole-genome genotyping and fine mapping, or is it going to go straight to fine mapping?

It will be a combination.

For some companies who have spent years trying to analyze SNP data already, they may already have a marker set that they think is a good set.

And J&J is a good example of that.

They came to us on the iSelect platform.

And what we are going to be shipping them are 12 sample chips that have a customized set of SNPs that they already believe are important in their clinical setting.

Other companies may do whole-genome first and then follow that with custom genotyping.

You'll see both approaches.

And then with respect to your services model, are you still saying people coming in and asking you to do the services, or are you now just seeing people saying, it is easy to do and we would like to just go ahead and get our own instrument, and so services -- your services revenues might not grow as best as your actual instrument placement and chip placements direct to the customer?

We continue to see very robust growth in our services order rates.

Our view is that that business is going to continue to grow.

Whether it grows faster or slower than the other part of the business is a little hard to tell and to project.

But we priced our relative products in those areas so that we are relatively indifferent about whether we get a service order, or the customer takes the equipment in-house.

Now for purposes of just ongoing business, in general we prefer to place the equipment and let the customer run the assays.

That is certainly the trend we're beginning to see in the pharmaceutical setting.

Whereas previously big pharma largely outsourced genotyping because the technologies were moving quickly, the SNPs were not.

It was not very clear exactly which SNPs to use to do the analysis.

I think with the advent of whole-genome technology the ability to put the very high percentage of SNPs that give great genomic coverage on a single chip, that technology question is being resolved.

And many more companies are looking at bringing their technology in-house.

Alastair Mackay with GARP Research and Securities.

I had another question about the BeadExpress.

You had mentioned the timeframe of 2008 for saying the BeadExpress introduction to the clinical diagnostics market.

Can you comment a little bit about the regulatory pathway that Illumina is envisioning for the machine?

In other words, are you talking about perhaps filing a PMA by the end of this year?

And would an application for the machine be matched to one of the three genes that you have talked about from deCODE, such as leukotriene A4 hydrolase?

To be clear on what I said earlier, we will begin selling the system into the clinical environment in 2007.

Our intent is to get the factory that produces the system, and this is our operation, we call it Illumina East Downing Connecticut, and the system itself through the regulatory process by the end of the year.

That is the instrument part of it.

By the end of 2006?

End of 2006.

That's right.

That may slop over a bit into the first quarter, but it will be roughly around the end of the year.

What will then happen is that we will have the ability to sell that instrument into CLEO labs.

And they will take the instrument and develop their own tests on it.

And so it will be used in a clinical environment, but without our proprietary content.

This will be third-party content.

We will also have the ability to sell it to partners who could use it to develop their own content and then market that system on their own to their customers.

What will happen in 2008 is that we will begin to deploy our own proprietary content on the platform in '08.

That requires us in '07 to go through all of the analytical and diagnostic validation steps of the marker sets.

And certainly the deCODE partnership will provide a foundation for that effort, and we will be working through the three sets of markers there in some sequence that we have not yet discussed.

And then our hope is to be able to submit that in a PMA form in 2008.

Got it.

That sounds good.

Can you mention in a clinical setting what you foresee as the upper limit of multiplexing for BeadExpress?

In other words, how many SNPs might you be able to analyze on a BeadExpress in the clinical setting once things are up and running?

The technology itself, and one of the reasons we have such interest in this technology is that it is able to multiplex to a very, very high level, theoretically into the many, many thousands, if not higher.

The point where it crosses over economically with our other platform is roughly in the range of 384 markers.

Our general view is that we will put this system into the marketplace that would address complexities from a few markers up to 384.

If someone were interested in marker sets bigger than 384 they would move to something like the Array Matrix as the superior platform.

But that isn't where the technology tops out, it is just where the economics crest.

In terms of the manufacturer of the cylinders or the beads for the BeadExpress, well that be done in the San Diego facility or are there are plans to be actually manufacturing the actual beads themselves in Connecticut?

That is already up and running in our Connecticut operation now.

Edward Tenthoff with Piper Jaffray.

Just two quick questions.

Kind of looking out to '07 the margins have really held in nicely.

Where do you think looking two, three years out margins can -- gross margins can ultimately get to?

And then also, as we look at the tax rate is this phenomenal performance that you guys are putting up accelerating what we might see as a tax rate for next year?

And can you give us any first thoughts on what you think that tax rate might be next year?

I will handle the margin question, and I will let Christian tackle the tax rate question.

On gross margins we have, of course, some competing factors there.

We continue to believe that we have the ability to reduce our manufacturing costs, fundamentally because our technology processes are quite young, and so we have a lot of headroom to continue to improve how we manufacture the arrays, and in fact the reagents.

We also have the improving possibility of improving [porcelain] function, if you will, that consumables over time will become a higher percentage of our business.

Offsetting those factors, of course, is the competitive environment in the marketplace, which effects pricing.

Our models as we combine all those factors, indicate that our gross margin should stay sort of in the high 60s range, and that if we got into the low 70s that would be a terrific performance.

From the tax side, as you probably know, we have over 100 million in NOLs that we are burning through, and we are burning through them faster than we originally had modeled.

For the remainder this year we're thinking our tax rate will be around the 8% level, given the level of foreign income that we have got.

Then looking into next year I suspect it will probably be in that 8 to 10% range.

And then looking into '08, it is difficult to say.

It depends on what our profits are, but I think it will be, as we burn through those NOLs, we will stay in that kind of 8 to say 10 to 12% range and then move to a more fully taxed rate.

And when we get to that point I will give you guys guidance on what that might look like.

I look forward to that.

Yes, we look forward to burning through NOLs.

(OPERATOR INSTRUCTIONS).

Chris Montano with Global Crown Capital.

Could you talk a little bit more about the BeadExpress for just a moment?

For example, what will be the competitive landscape for the BeadExpress product, and what type of market size are we looking at?

The target market for the system will be both research applications and clinical applications.

In the research sector first, we will sell it into our existing installed base of academic and commercial users.

And as they do whole-genome work and then focus research and reduce the marker sets down to where they are dealing now with a few hundred rather than going off into other platforms, which is what happens today in sort of scientist competitive space.

Those platforms being things like [Packman, close icommiters].

In some cases mass spec systems get used in that marketplace.

They will now use our BeadExpress system instead.

And so that is how it will be used in the research market.

And we will have a continuous offering.

We will have two different scanners, but the software will be fully integrated.

The essay methods will be integrated.

And it will all come from a single supplier now, from whole-genome genotyping all the way down to, if they wanted to go to single plex genotyping.

We think that has a strong competitive advantage.

In the more clinical markets, of course, we will be selling to the third-party laboratories and through partners.

And the competitive space there is highly competitive I should say.

They'll be many players in the molecular diagnostics area, and many big players, in fact.

We think the platform has significant advantages in terms of the multiplex level that it is able to achieve, the kinds of throughput and sensitivity.

And in the long run we think that a key advantage that Illumina will bring here is that we offer the full pipeline of capability from the instrument, the software, the reagents, the essay methods, and ultimately the proprietary content.

And it is that proprietary content over the long run that we think we'll really drive great economics for Illumina.

What about the market size that you are addressing with this product?

In the clinical markets it is certainly some numbers of billions of dollars is the projections for molecular diagnostics in a relatively short order.

If you look at some segments of that market already that are doing what is called molecular diagnostics, there's some segments like blood bank testing that are quite large already, ranging from 4 to $500 million.

So the markets are evolving fast, and we think the opportunity that we could address is certainly going to be in the multiple billions of dollars.

Like I said, it is going to be a very competitive space, so we have no qualms about that.

There are going to be lots of players.

Thank you very much.

Congratulations on a great quarter.

Tycho Peterson with JP Morgan.

Along the lines of the last question, would you ever go after the point-of-care market?

Is that an interest for you?

You talked about going after the laboratories on the clinical side.

Maybe in the long run.

Certainly we have some technology ideas about how to potentially address that.

We have nothing that we currently talk about there.

But certainly it is going to be a large market if you look out say five to ten years out, and so it would be something that would be on our radar screen to take a look at.

I am wondering if you could give some color on the Invitrogen collaboration is going and any next steps?

Sure.

The collaboration is fully up and running at this point.

We had some good growth quarter to quarter -- year-over-year quarter -- Christian mentioned in his remarks.

I think we're probably not quite where we want to be in terms of the overall order levels, and so we're working with Invitrogen cooperatively to try to figure out ways to boost that.

And they're making lots of additions on their side, doing such things as adding a greater number of specialist selling oligos to improve the order rate.

So we're cautiously optimistic about how they're going to do over the next few quarters.

But as we mentioned, we are continuing to see good growth there, and it is a very profitable business for us.

And then finally on the diagnostic business, I know in the past you talked about some of your other collaborations like [Jennison].

Is it safe to assume the initial products will come from the deCODE collaboration?

How do we think about those other partnerships?

Yes, it is certainly more likely that they would come from deCODE rather than those other few partnerships that we had signed, because those other partnerships are still in the discovery phase.

So they are not even quite to the point where deCODE is in terms of discovering the markers.

We get access once they discover something relevant.

That is still an unknown.

It certainly is possible, however, that we could cut a deal with someone who already has some products in the market using an existing proven marker set and just wants to change platforms.

We're potentially talking to a number of OEMs about that as well.

Okay, great.

Thanks and --.

Proprietary content, deCODE is the most likely one to be first.

Okay.

Congratulations.

Zarak Khurshid with Caris & Co.

The numbers speak for themselves.

Great quarter.

I would imagine the range of consumable pull through per BeadStation has remained relatively constant.

Can you comment on the average utilization per BeadStation and if that has gone up?

It has gone up significantly over the last few quarters.

In Christian's remarks he had mentioned that in the quarter just completed that we are over $500,000 of consumable pull through per BeadStation per year.

The reason that the numbers have increased so dramatically is that in many of these large studies we're now specing out the equipment that the customer buys to essentially run almost 24 7.

And so we have full automation and robotics supporting feeding BeadChips into the scanners.

What will typically happen is the customer will say, I want to do an 8,000 sample project.

I want to get it done in a particular timeframe.

And then we will spec out how many much equipment they need to get it done running 24 hours a day.

And so what that does is on those installations the utilization approximates 100%.

Obviously, we don't quite get to 100, but we get close.

And that is bringing up the averages dramatically.

That is why we have gone from something that was maybe 2 or 300,000 a year ago to 500,000 today.

With the launch of the 650 K chip what does that do to demand for the 550?

And any chance that CHOP and deCODE would opt for potentially a greater content chip?

The 650 will be, we think, used by research groups that have particular interest in studying African or African-American populations, because we boosted the coverage in those populations on that chip.

In the case of deCODE, it is most likely that they will stay with the 300 because they're dealing mostly in Western -- Western European type populations.

And the coverage of the 300 K is terrific in those populations, and they get the best economics there.

We have seen, since we launched the 550 K, a conversion of quite a number of customers who were originally using our 300 K up to using the 550 K. And the way we have designed this productline, of course, if they switch to the 550 K in the middle of a project, they can then go back and use the 240S product to fill in, to get all the samples up to the 550 K level.

It is just a real great degree of flexibility that we offer the customers in that regard.

In terms of people there are already customers we expect some will bump up to the 650.

But I think that will be a smaller percentage than the number that bumps up from the 300 to 550.

Going back to the Invitrogen collaboration, did you break out what that contribution was?

No, we didn't.

A lot of people asked us about how the collaboration is going, so I thought it would be important this quarter to give everyone at least some metrics that it is growing.

It is included in product revenue, and we're going to continue to keep it as a component of product revenue.

At this time this now concludes the question-and-answer session.

Ladies and gentlemen, thank you for your participation in today's conference.

This concludes the presentation.

You may all disconnect, and have a good day.