Penguin Solutions Inc (PENG) 2019 Q3 法說會逐字稿

Good day, ladies and gentlemen, and welcome to the SMART Global Holdings Third Quarter Fiscal 2019 Earnings Conference Call.

(Operator Instructions) As a reminder, this call may be recorded.

I would now like to turn the conference over to your host, Ms. Suzanne Schmidt, with Investor Relations.

Ma'am, you may begin.

Thank you, operator.

Good afternoon, and thank you for joining us on today's earnings conference call to discuss SMART Global Holdings' third quarter fiscal 2019 results.

Ajay Shah, Chairman and Chief Executive Officer, will begin the call with a discussion of the market and the business; followed by Jack Pacheco, Chief Operating and Financial Officer, who will review the financial results in more detail and provide the forward guidance.

After which, we will open the call to your questions.

As a reminder, our earnings press release and a replay of today's call can be accessed under the Investor Relations section of SMART's website at smartgh.com.

We encourage you to go to our website throughout the quarter for the most current information on the company, including information on the various financial conferences we will be attending.

Before we begin the call, I would like to note that today's remarks and the answers to questions may include forward-looking statements.

Any statement that refers to expectations, projections or other characterizations of future events, including financial projections and future market conditions, is a forward-looking statement.

Actual results may differ materially from those expressed in these forward-looking statements.

For more information, please refer to the risk factors discussed in the documents we file from time to time with the SEC, including our most recent Form 10-K and Form 10-Q.

We assume no obligation to update these forward-looking statements which speak as of today.

Additionally, during this call, our non-GAAP financial measures will be discussed.

Reconciliations for those directly comparable GAAP financial measures are included in today's earnings press release.

With that, I will now turn the call over to Chairman and CEO, Ajay Shah.

Thank you, Suzanne, and welcome to everyone on the call.

In the third fiscal quarter, our team has performed very well under difficult circumstances.

While the business environment was challenging, our focus on operational excellence and areas that are under our control led to non-GAAP earnings per share of $0.34, which is within the range of the guidance we provided last quarter.

Additionally, we were able to improve our cash generation from operations by almost 20% sequentially to reach $46 million in the last quarter alone as we reduced inventory significantly in each of our businesses to be able to take advantage of lower component prices.

In addition, we were able to maintain our gross margins, even while average selling prices dropped significantly.

We ended the quarter with cash and equivalents at $126 million, which positions us well to execute on our strategies around growth and acquisitions to build our growth.

Let me first comment on the main factors impacting our results in the third fiscal quarter.

It's been very well publicized that the memory industry continues to see pricing declines in both DRAM and flash over the past several months.

In addition, as we've heard from other technology companies, inventory corrections at our major OEM customers were a factor in our business results for our Specialty Memory and for our Brazil businesses.

The biggest impact quarter-to-quarter was a nearly 33% revenue decline in Brazil, which as many of you know, is the only part of our business where we sell memory products into high-volume end uses such as smartphones and PCs.

This decline was primarily due to declines in memory component prices and as a result, our declining average selling prices.

Brazil now makes up 43% of our total company sales compared with 69% in the year ago quarter, so we have made some progress in reducing that exposure.

I'll now review each of our 3 areas of business in more detail for the third quarter and then turn the call over to Jack for a review of the numbers as well as our guidance going forward.

Turning first to our Specialty Memory business, which represented about 42% of net sales in the quarter and totaled roughly $99 million.

Net revenues were lower than the prior quarter due to continued inventory corrections at a few customers.

On a year-to-date basis, net sales for the first 3 quarters of fiscal 2019 was still 12% higher than the same period in fiscal 2018, in spite of significant headwinds in the overall market and price reductions.

And we continue to believe that fiscal year 2019 overall will grow in the low double-digit percentage range over fiscal 2018.

We are seeing some signs that the majority of the inventory corrections are behind us, and that should lead to unit volumes recovering in the periods ahead.

We remain very optimistic about growth prospects in our Specialty Memory business.

We are introducing new product families, adding new customers and benefiting from our focus on long-life products.

In addition to our broad portfolio of products, and in the last quarter we shipped over 600 unique products in -- just in that period, we're also seeing many new opportunities and we've been expanding our product portfolio to address these prospects.

For example, earlier in the quarter, we announced 2 new additions to our SMART RUGGED product family.

The first is a small form factor 32-gigabyte industrial-grade SO-DIMM module.

There's a growing need for ruggedized mobile computing devices in the field, and this product is ideally suited for robust and resilient computing applications used in industrial, defense, transportation, public safety and other similar markets.

Another addition to this product family is the T5E, which is a NAND-based SATA SSD that has up to 4 terabytes of storage capacity.

The product is ideal for high-throughput applications such as flight data recorders and sensor data capture as well as high-reliability telemetry, surveillance and other mission-critical storage applications.

With a custom flash controller, the T5E is a very flexible configuration and has multiple optional security features.

Moving on, our Brazil line of business represented about 43% of total company net sales for the quarter, as I mentioned before, and was approximately $101 million, which is down substantially from $147 million in just the previous quarter.

Consistent with what we indicated on last quarter's call, we have been impacted by falling memory prices, and we did take actions during the quarter to reduce headcount and adjust expenses.

We continue to see memory prices declining in our fiscal Q4, but in Brazil, we are also seeing significant increases in unit volumes, which are helping to stabilize the performance of that business.

We will talk some more about the look of that business going forward.

Also recently, the Brazilian government agencies have published changes in the method of determining local manufacturing, also known as PPB, requirements going forward.

We have mentioned on a previous call that there were changes coming.

Our assessment of these changes is that they will not materially change our current business profile.

We're working with our major customers in Brazil to put together our joint go-forward plans to support their business and needs for both memory products as well as for battery products under the new rules.

Finally, moving on to our specialty computing and storage business, which represented 15% of our net sales in the quarter, approximately $36 million.

We saw continued strong interest in our solutions, particularly from the government and aerospace end markets.

But due to the timing of orders during the quarter, revenue was below our expectations in Q3.

However, we are entering fiscal Q4 with a strong backlog and a pipeline, particularly during this period for government customers, due in part to the federal fiscal year-end.

We have strong expectations for this business in Q4.

Separately, through our team's efforts, we've been able to improve gross margins and reduce inventories in this business.

Our focus on solutions targeted at the AI vertical is resulting in important new opportunities.

As the industry increasingly looks to AI to solve the most challenging technology and business problems, Penguin is offering the latest technology capabilities with customized services and support.

These offerings are finding a lot of interest in a variety of sectors as varied as aerospace and retail automation.

Earlier in this quarter, we announced that Penguin's Relion family of Linux-based servers is now available with the latest generation of Intel Xeon processor, enabling Penguin to optimize performance with data center, high-performance computing and AI customers.

And as of last week, customers can now take part in an early validation program to benchmark their applications.

We also announced a collaboration with WekaIO, a private company that provides some of the most scalable and fastest file storage for AI and computer applications.

Our new integrated solution in partnership with Weka has broken 8 records on the STAC-M3 Benchmark, the industry standard for testing solutions for high-speed analytics used in applications such as algorithmic trading and quantitative analysis workloads, these being common kind of workloads in the financial services industry.

We achieved this performance benchmark by running in distributed mode over the network versus a traditional block storage, which then provides the ability to simultaneously support other applications on the same platform, a big leap in application acceleration for financial services.

Also in the specialty computing area, particularly related to high-performance computing, Hewlett Packard or HPE announced an agreement to acquire Cray Computing in May.

As a result of this acquisition, our company, Penguin Computing, would become the second-largest dedicated HPC provider in North America.

It's been a very interesting quarter and a lot of highlights going positively into the future.

Now let me turn the call over to Jack for a review of our financials and our guidance going forward.

Jack?

All right.

Thank you, Ajay.

Overall, gross revenue for the third fiscal quarter was $442 million, while net sales were $236 million.

As a reminder, the difference between gross revenue and net sales is related to our supply chain services business, which is accounted for on an agency basis, meaning that we only recognize the net sales, net profit on a supply chain services transaction.

Our breakdown of net sales by end market for the third fiscal quarter was as follows: mobile and PCs, 39%; network and telecom, 22%; servers and storage, 13%; industrial aerospace, defense and other, 26%.

Now moving to the rest of the income statement.

Non-GAAP gross profit for the third quarter was $43.7 million compared with last quarter's $57.8 million, due primarily to lower sales in Brazil.

The non-GAAP operating expenses were $30.5 million, approximately in line with the previous quarter.

Non-GAAP net income to the third fiscal quarter was $7.9 million or $0.34 per diluted share, and adjusted EBITDA totaled $19.2 million in the third quarter.

Turning to working capital.

Our net accounts receivable decreased to $230.2 million from $326.5 million last quarter, and our days sales outstanding decreased to 47 days for this quarter compared with 51 days last quarter.

Inventory levels continued to decline and totaled $133 million at the end of the third fiscal quarter compared with $172 million at the end of the second fiscal quarter.

Inventory turns remain the same as the previous quarter at 12x.

Consistent with past practice, accounts receivable, days outstanding and inventory turnover are calculated on a gross sales and cost of goods sold basis, which were $442 million and $399 million, respectively, for the third fiscal quarter of 2019.

Cash and cash equivalents grew by $31 million to $126 million at the end of the third quarter compared with $95 million at the end of the prior quarter.

Third quarter cash flow from operations was $46 million compared with $39 million in the prior quarter.

We continue to make progress in the integration of Penguin, and the gross margins improved again this quarter.

We remain ahead of our plan for achieving operating synergies and expect Penguin to be accretive to EPS in the fourth quarter.

I'd like to provide an update on the inventory correction that we have been seeing based on what we are observing in the memory market overall.

We are seeing flash lead times now at 4 to 6 weeks versus the 30-week lead times we had seen over the past few years.

Memory price declines are continuing to pressure revenue in today's market.

Input from our customer gives us confidence that the majority of the inventory corrections will be behind us as we exit the fiscal year.

With that as a backdrop, let me now turn to our guidance for the fourth fiscal quarter.

We currently estimate that our fourth quarter fiscal '19 net sales will be in the range of $270 million to $280 million.

Gross margin for the quarter is estimated to be approximately 19% to 21%.

GAAP earnings per diluted share is expected to be between $0.33 to $0.43.

On a non-GAAP basis, excluding share-based compensation expense, acquisition-related expense and intangible asset amortization expense, we expect non-GAAP earnings per diluted share would be in the range of $0.55 to $0.65.

The guidance for the fourth fiscal quarter does not include any view on the foreign exchange gains or losses and includes an income tax provision expected to be in the range of 12% to 16%.

The number of shares used in computing earnings per diluted share for the fourth fiscal quarter is 23.5 million.

Capital expenditures for the fourth fiscal quarter are expected to be in the range of $1 million to $3 million.

Please refer to the non-GAAP financial information section and the reconciliation of non-GAAP financial measures to GAAP results and reconciliation of GAAP net income to adjusted EBITDA tables, earnings press release for further details.

Operator, we're now ready to take questions.

(Operator Instructions) Our first question comes from Kevin Cassidy of Stifel.

On the inventory correction, can you say which end markets maybe you're seeing better demand?

Or is it just across the board?

And do you think that -- is there an increase in demand also?

Or is it just the steady demand that has worked down the inventory?

We're seeing that inventory is across multiple end markets, but I guess we're just -- we're seeing steady demand as they work down their inventory.

Okay.

So you don't see any particular uptick in demand?

It's just that the inventory is out of the way?

Yes.

I mean it's hard for us to tell if demand will be coming, and we just assume it's kind of normal demand for them and are just working through this inventory they had built up due to the long lead times.

Okay.

And as far as Brazil goes, do you have a sense of how your shipments are compared to the local content laws?

Are we -- should we expect a very strong second half?

It seems as though the units are picking up.

Is that expected to carry through into the fourth calendar quarter also?

Kevin, this is Ajay.

Again, we are unclear as to what's going to happen to memory prices.

What we have done in terms of our strategy is minimize inventory and be able to move very quickly to take advantage of the lower prices.

As we mentioned in our remarks, we are seeing significant increases in unit volumes in the Q4 period.

And we assume that, that is because we have both end markets as well as the customers' understanding of how local content is going to work is now clear.

So that's why we believe we're seeing the units come up strongly.

And whether that continues into Q4 is hard to forecast -- I mean, Q4 calendar year, it's hard to forecast right now, but we don't see any reason why not at this point.

Okay.

And if I could just ask one other question.

The other part of your business, the specialty computing.

The HPE acquiring Cray, is that going to change the dynamics?

Or was Cray not a direct competitor?

No.

We do compete with both HPE and -- which is really SGI that was acquired by HPE and Cray to a degree.

And so the combination of HPE and Cray, assuming that happens, is essentially a consolidation, which means that many customers, particularly government customers that are looking for diversity of sources, will then be looking for additional sources.

So we feel that, that will certainly open up more opportunities to us in the future.

Our next question comes from Mark Lipacis of Jefferies.

A couple of questions.

First, on the Brazil volumes increasing, is this a seasonal element that's impacting this?

Or is it product cycle related or macroeconomic related?

Can you provide a little more color on that one?

Yes.

Remember, Mark, we have said in the past that they were ordering less products earlier in the year.

And as we got towards the later in our fiscal year, which is more in calendar year -- later of calendar year '19, they would start picking up the orders to meet the local manufacturing content they have to buy.

So I think we're seeing the orders picking up, as we have said before, because they have to go meet these regulations down in Brazil.

Okay.

Got you.

And then the -- and then, Ajay, you had mentioned the change in definition of local manufacturing.

Can you flesh that out a little bit, like how has it changed?

I understand that you're of the view that it doesn't impact -- materially impact your business, but it sounds like you're working with your customers to make sure you're meeting their needs.

But if you could provide a little more color on that, I think that'd be helpful.

Sure.

Sure, Mark.

We in very high-level terms and trying to explain the intricacies of Brazilian local manufacturing requirements in a short phone call are a dangerous proposition.

But essentially, in the past, the way local manufacturing requirements was set up was as a percentage by each commodity.

So memory for smartphones would have a certain percentage, right?

I think you've seen that from us in the past.

Going forward, they have essentially a collection of local manufacturing requirements as a total, not by commodity.

So there's a certain amount of local manufacturing requirement to be met.

And each customer, and our customers, get a certain number of points for each kind of local manufacturing that they acquire.

So there are different commodities with different levels of points.

It turns out memory has, by far, the highest number of points, and then the second highest is batteries.

And in our conversations with customers, that effectively -- the math would come back to effectively the same kind of level of businesses we do today.

That's -- I know it's a somewhat high-level explanation, but to take you any deeper would require spreadsheets and...

Okay.

That's actually very helpful.

I think I understand conceptually what's going on there.

I mean, to me, it sounds like...

But it clearly is -- I mean, I think it actually clarifies things.

There was a little bit of uncertainty, but this clarifies things.

It's apparently compliant with the WTO rules, blah, blah, blah.

So I think that we've come out in a good place.

Remember, we said earlier, we didn't think these new rules really would impact the way they were doing things.

And they published the rules what they said they -- the way they said they were, which means they're just keeping things pretty much the same way for most of the companies in Brazil, which is a positive statement.

Understand.

But I think I'll maybe go one step further and tell you that if we look at our Q4, we're not forecasting.

Just to be clear, we're not forecasting an increase in Brazil revenue.

If anything, we're forecasting because of price declines, still a small reduction in Brazil revenue from Q3 to Q4.

Our increases are coming in other areas.

In Brazil, what's really happening is that the revenue declines are muted by the fact that units are increasing significantly.

I think it's worth a little clarification.

Understand.

And then on the Penguin on the timing of orders, I think most people understand that a number of the hyperscale players had added capacity ahead of actual consumption rates in the second half of last year.

Is what you experienced, do you think it's related to that?

Or it's just some orders didn't come quite when you thought they were going to come and it's just as simple as that?

We don't think it's related to the hyperscaler data centers because we really don't ship to them.

Those are not our customers in this business.

In our Penguin business, our customers are primarily large, high-performance computing clusters or AI clusters and in any event, large clusters.

And our biggest customers are either government, including labs and government-funded university installations and departments as well as the intelligence community; and commercial: oil and gas, financial services, media and so on.

So the particular lateness was because this is actually the high season for government buying simply because you have the year-end budget flush for the fiscal year-end.

So we had expectations for when those orders would come in and how they would ship, and it's just running a little later.

Understand.

And then lastly, you mentioned the flash lead times in the 4- to 6-week range versus previous several years in the 30 weeks range.

So is that classically the range that you have seen between 4 and mid-30s?

Or have you seen it go lower than 4 to 6 weeks in the past?

4 to 6 weeks is an average kind of lead time that we've seen, and that's -- and we don't see it getting much lower than that because you've got to get product and build up.

But the 30 weeks is a very long lead time that really happened when flash was in an undersupply situation.

So we're just getting kind of back to normal.

But the issue is if people built up inventory space in these long lead times and lead times shortened, they're working on that inventory without having to order to replenish, right?

So they're dropping their inventories down as well in -- to get them more in line with the 4- to 6-week lead time kind of situation.

Our next question comes from Sidney Ho of Deutsche Bank.

You previously talked about the change in the local content requirement in Brazil has led lower memory procurement, especially for high-density products?

I think, today, you have kind of updated that you're seeing that uptick a little bit.

But you also talked about the average content for phones seems to have slowed down in Brazil, not really sure if that's tied to the local content requirement changes.

Can you give us an update on both areas here?

So I think on density in the phones, I mean, it's kind of where we thought, it's lower than what we've seen in other parts of the world.

I don't know if that's as much of any of the local content rules.

I mean I think the reais is still pretty weak.

It almost has been -- in the quarter since, somewhere between BRL 3.7 to BRL 4.10.

And so I think a weak reais makes phones expensive.

And so I think the phone manufacturers are trying to make their phones less costly to sell to the consumer and so they put less memory in the phone.

So I think it's more related to the Brazil economic situation than anything else.

And then...

Okay.

Maybe I can switch -- yes, go ahead.

No, no.

ASP front, I mean, we talked about ASPs dropped in Brazil and they dropped actually a little more than we had thought in the quarter, right?

We thought they were going to go 40% to 45% down.

They actually dropped almost 50% in the quarter.

So memory prices, as Ajay mentioned, continue to decline.

And we'd expect the decline to continue into Q4.

So we still expect price, especially in Brazil, to decline to our consumer base based on the cost we get from our suppliers.

So just to be clear, the mobile side of the ASP, what would that be in the last quarter?

I think we finished last quarter about $16.45.

We're...

Okay.

All right.

Great.

That's helpful.

My next question is you guys kind of talked about the Penguin is strong in the second half of the year.

You just talked about timing may be slipping from Q3 to Q4, and you talked about seasonal strength from government and whatnot.

How should we think about -- is most of the -- you talked about Brazil as probably not growing, maybe down a little bit.

Is that most of the increase is going to come from Penguin?

Or is it from Specialty Memory in the upcoming quarter?

In the upcoming quarter, Penguin is a particularly strong driver -- in this quarter, I should say, the Q4 period.

And we then continue on with both Specialty Memory as well as our Specialty Compute or Penguin business.

So we're seeing both strengthening.

This particular quarter is, as I was saying earlier, due to the fiscal year-end has very strong federal component to it.

Okay.

Maybe last question for me.

You guys discussed a couple of times on the changes to the local content rules.

When should we expect those changes to be finalized and put in place?

And is it -- when you say you don't expect material changes to the current business profile, I assume that that's the same as no major changes to your financials as well.

Is that right?

I'll answer the second question first.

Yes, we believe that the financials would be largely in line with where they've been right now.

And well -- but that's not taking into account memory price-related pressure.

We don't make memory, as you know.

We buy and we -- and pass it through.

So effectively, our revenue can be affected by memory prices.

So with that caveat, yes, we feel pretty comfortable that the business is more or less as it's been in the past.

The guidelines for the reviewed -- for the renewed local manufacturing requirements came out last week and are effective July 1, i.e., next week.

Our next question comes from Blayne Curtis of Barclays.

Just one more on this content requirement.

Just trying to think through it, just the amount of time you were looking for increases in the mobile requirement.

Under this new framework, is there any embedded increases in that overall -- you're [calling] overall bucket?

Is there any kind of embedded growth in that content?

And then I'm just trying to think through of all these components, I'm assuming the OEMs will maximize where they -- the components that make the most sense from a cost perspective.

So in that framework, I guess, I don't know all of the different components that are in there, but at least maybe can you address memory versus batteries.

And what would be high density versus low density?

I mean how do you see this market [faring] in this framework?

Blayne, I'm not sure we know how to answer that question yet.

There is an opportunity and we are working with customers to try and see if maybe they can make it, frankly, easier for themselves that they can essentially, rather than trying to manage local manufacturing across, I don't know, was it 25 different items that you could do that with, I think that was [somewhat less] that, that you could make it much easier from a management standpoint with 1 or 2 commodities.

And memory would be, since it's by far the biggest, the easiest.

And so we think that, that presents an opportunity.

We haven't modeled that.

Okay.

And then just the first part of my question, is there any increases embedded in this new program?

Increases.

Sorry, I didn't quite understand.

In this new program, the -- you hit so many points, I don't fully understand it.

But I'm assuming are they going to require greater points going down the road in the same way that you had increasing requirements, 50%...

Yes.

I guess -- thank you.

Yes, we don't know.

But as Ajay said, these -- they came out like last Friday, these rules, and they're effective like Monday.

So we're still digesting this 'and the customer base are still digesting these rules that came out.

So...

Got you.

And then maybe just finally for you, Jack, just gross margin, you are guiding it up.

I would think that the mix of Penguin may actually hurt that.

Can you maybe just walk through where you're seeing the improvement?

Yes.

No, we're actually seeing Penguin's gross margins growing.

I mean, we've been working hard to improve their gross margins.

And so I think if you look at Q4, I mean, Penguin's margins are improving.

Specialty margins have been fine throughout this endeavor.

And then Brazil, we've -- like Ajay said, we have some -- we've done some cost reductions there and stuff, so Brazil margins are getting a little bit better.

But it's kind of across the board, I mean.

But Penguin, we've done a lot of work on the gross margins there and we are improving that business on a quarterly basis.

Yes.

To continue on with what Jack was saying, Penguin gross margins have improved every quarter since we acquired the company meaningfully.

And as you noted, Blayne, I mean, we're guiding revenue up significantly from Q3 to Q4.

We're guiding gross margins up sequentially.

We're guiding EPS significantly up, almost 2x.

We're trying to -- having just earlier this week seen a pure play memory company's guidance, clearly we're going in a very different direction.

And we're trying to figure out exactly what the market was expecting.

Our next question comes from Rajvindra Gill of Needham & Company.

On the -- just to follow up on the new rules going from a local content percentage system to a point system, my understanding of researching into it is that in order for Samsung or LG or whoever it might be, to get the tax benefits, they would have to accumulate 35 points or so.

And that could be a combination of a lot of different things, but memory was almost 27 points of that.

So I just want to make sure that in order to get the 35 points and qualify for the tax benefit that memory is -- they're primarily going to continue to use wafer cutting and testing of memory integrated circuits, and they won't necessarily try to achieve the 35 points through a combination of other different components, whether it's LCD TV -- LCD units or R&D investments?

Right.

Right.

So I think that's a good question.

First of all, let me just tell you that the 35-point thing has changed.

Last week, it was revised.

So there's a whole new set of numbers.

But I won't try and belabor all of those here.

Trust me, we're happy to teach in on that, and I will be an avid student hearing that teaching.

The point, though, is that, as you noted, memory is a very large part of what local manufacturing requirements are going forward for our customers, but it does provide them some flexibility, so they could buy.

But there is -- it turns out there's no -- yes, there are points for LCD, but there are no local LCD manufacturers.

There are points for, [quips] I guess they're called, but there are no local manufacturers.

In fact, there are no local users either.

Then there's points for manufacturing the motherboard, which are well defined.

So when you look at it, in terms of actual things you could buy locally today, you pretty much see that memory becomes by far the dominant commodity.

I don't think this is a coincidence.

The government sees what they have in terms of local in-country capabilities.

And they see that for memory, there is the capability with 1, maybe 2 companies that are significant.

And meanwhile, they don't -- they're trying to encourage people to come in and manufacture LEDs or other types of commodities, but -- or you know, that's great.

But in terms of actually meeting local manufactured component requirements, memory and battery are the 2 dominant ones.

And so that's the place from which we're coming from when we say we don't think that it changes our business significantly in any way.

So if it doesn't necessarily change the business, I mean, how do we just think about the mechanics of it?

Because in the past, it was a very simple kind of math calculation, 50% multiplied by the number of mobile phones multiplied by some sort of ASP per unit.

And if it's kind of moving from a percentage qualification to a point system, how do we think about modeling that going forward?

Because I mean, it seems like the -- let me just also -- just want to make one point.

So are we basically saying that these new rules have satisfied the WTO requirements going forward and there should not be another adjustment in the future, that these are -- this point system now is satisfying WTO?

Is that also fair to say?

Yes.

We believe that -- and they believe more importantly, that this point system, this method of encouraging local manufacturing is in compliance with what the WTO wanted.

In fact, the big rush they've been in to implement it is because the WTO wanted it implemented by the year.

And this is again hearsay, this is what they tell us.

But to come back to your question about how do we model this?

I do agree that it's not as simple as saying, well, gee, it was 50%, now it's going to 60%, therefore.

So I would give you 2 answers.

One is we think we have a pretty good fix for what the demand is going to be, and we'll try and continue to communicate that.

And secondly, if you look as -- in this quarter, Q4, hopefully this will be looked upon positively.

Brazil is now going to, in our forecast, be sub-1/3 of the business.

And so it's not the driver of our business than it was a year back.

I understand that everyone was kind of fascinated with the local content requirements in Brazil and all of the fun samba stuff that comes from it.

But the -- all I'm trying to say is that it will be an important element of our business and we will continue to communicate that potential.

But at the same time, we'd like to make sure that we communicate that we have other businesses that are doing well.

That we're growing.

That we're, as a result, less dependent on memory, especially the commodity memory products.

Our Specialty Memory products have shown that they've held up through this incredibly turbulent period in the memory market.

And meanwhile, our Specialty Compute product is growing.

And we continue to have plans for further growth and broadening of our portfolios.

So sorry for the long and perhaps a little off-your-question answer, but I hope we're communicating what I think is important.

No.

I think it's very clear.

And just one follow-up question, Ajay, so the Specialty Memory, I think you had said that you still expect it to grow low double digits year-over-year in fiscal year '19.

That would require a pretty big snapback sequentially in August.

Is that -- could you clarify that?

No.

In the first 3 quarters, we are up 12% year-over-year revenue-wise.

Okay.

I might have misunderstood.

Did you say something for the fiscal year '19 what you had mentioned?

Or you're not giving any guidance for that for the full year?

No, no.

What I'm saying -- we are talking about fiscal year '19.

We're saying that through the first 3 periods of fiscal year '19, we're up 12% from the first 3 periods of fiscal year '18.

And we are also saying that for the full year fiscal '19, we expect to be up in double digit -- low double digits from fiscal year 2018.

Yes.

That would require a significant snapback sequentially in August versus May.

Not really.

I mean we're up 12% year-to-date already.

Okay.

I got it.

We can maybe take that off-line.

And just on the -- Jack, on the gross margins, the 20%, which is, I think, really, really good given kind of all the memory pricing, a lot of that is due to Penguin.

In the past, we've been at 23%, 23.5% gross margin.

How do we think about the margin profile going forward?

Is there more drivers in Penguin to drive it up further and help the overall margins?

Are we going to have to kind of wait to see a recovery in overall memory pricing to kind of get back to 23% range?

What's kind of like the big picture thing?

I think a driver for -- one driver for getting the overall margins back up is going to be Brazil and just the utilization of the factory.

As Ajay mentioned, more units are going through Brazil.

So as we produce more units in Brazil, we have a large fixed cost in our packing facilities.

So as we get more units through Brazil and manufacture more units down there, that will help with the gross margin.

Penguin, of course, helps the gross margins.

And specialty, we're seeing unit growth, as Ajay mentioned.

So if we get the unit growth in specialty and that business continues to grow, then that would also impact the gross margins, and you'll see the gross margins grow.

So I think we're on a path to get the gross margins back up to where they were with all 3 areas contributing to that.

(Operator Instructions) We have a question from Suji Desilva of Roth Capital.

(inaudible) to understand as a baseline for that and what the typical seasonality can be?

Suji, you're cutting in and out really bad.

Can you repeat that?

Sure.

Sure, Jack.

On specialty computing, can you talk about what the typical seasonality is?

Quantify that wherever strong quarters stand.

Well, we've always said that we would anticipate that our Q4 would be the strongest quarter for that business, right?

And then -- but Q1 is not -- Q4 is the strongest and we always said Q2 -- somewhat Q2 would be the weakest due to the way the government spends.

And Q1 always is a pretty good quarter as well.

So it's kind of playing out here over the last few quarters what we've been kind of telling people, the seasonality for that business.

Okay.

And then my next question is on the deal size that you're seeing for Specialty Compute.

As you've been progressing, are the deal sizes still relatively stable?

Are you seeing larger deals as you move forward?

Any color there would be helpful.

I mean this business always has had some large deals go with some small deals, right?

I mean so you might get a $15 million deal or you get small deals.

So it depends on the installation that somebody launched.

But I don't think we're seeing bigger deals than what we've seen.

But as we grow this business, we're getting more bigger deals.

I'm showing no further questions at this time.

I'd like to turn the conference over to CEO, Ajay Shah, for any closing remarks.

Thank you, operator.

In conclusion, I believe we executed well on the challenging circumstances in this third fiscal quarter.

And we importantly are positioned to resume growth organically as well as in the future through targeted and synergistic M&A.

We look forward to reporting on our business in the coming months as we move towards a more balanced mix of business overall.

Thank you again for joining us on this call this afternoon.

Thank you.

Ladies and gentlemen, this does conclude today's conference.

Thank you for your participation, and have a wonderful day.

You may all disconnect.