Vicor Corp (VICR) 2018 Q3 法說會逐字稿

A very good day, and welcome to the Vicor Earnings Results for the Third Quarter ended September 2018.

(Operator Instructions)

This conference is being recorded.

And now I would like to hand over to your host for today, James Simms, Chief Financial Officer.

Please proceed, sir.

Thank you, Mark.

Good afternoon, everyone, and welcome to Vicor Corporation's earnings call for the third quarter of 2018.

I'm Jamie Simms, CFO.

And with me here in Andover are Patrizio Vinciarelli, CEO; and Dick Nagel, Chief Accounting Officer.

After the markets closed on Tuesday, October 16, we issued a press release summarizing our financial results for the 3- and 9-month periods ended September 30.

This press release has been posted on the Investor Relations page of our website, vicorpower.com.

We also filed a Form 8-K on Tuesday related to the issuance of that press release.

As we had completed the process of closing the quarter's financial statements, we released results on Tuesday.

But because we had already scheduled this conference call, we left its timing unchanged.

As always, I remind listeners, this conference call is being recorded and is the copyrighted property of Vicor Corporation.

I also remind you, various remarks we make during this call may constitute forward-looking statements for purposes of the safe harbor provisions under the Private Securities Litigation Reform Act of 1995.

Except for historical information contained in this call, the matters discussed on this call, including any statements regarding current and planned products, current and potential customers, potential market opportunities, expected events and announcements as well as forecast sales growth, spending and profitability, are forward-looking statements, involving risks and uncertainties.

In light of these risks and uncertainties, we can offer no assurance that any forward-looking statement will, in fact, prove to be correct.

Actual results may differ materially from those explicitly set forth or implied by any of our remarks today.

The risks and uncertainties we face are discussed in Item 1A of our 2017 Form 10-K, which we filed with the SEC on March 9, 2018.

Please note, the information provided during this conference call is accurate only as of today, Thursday, October 18, 2018.

Vicor undertakes no obligation to update any statements, including forward-looking statements made during this call, and you should not rely upon such statements after the conclusion of this call.

A replay of today's call will be available beginning at midnight tonight through November 2. The replay dial-in number is (888) 286-8010, followed by the passcode 77605130.

In addition, a webcast replay of today's call will be available shortly on the Investor Relations page of our website.

I will start this afternoon's discussion with a review of our financial performance for the third quarter, and Patrizio will follow with a few comments and take your questions.

Beginning with consolidated results, as stated in Tuesday afternoon's press release, Vicor recorded total revenue for the third quarter of $78 million, which represented a sequential quarterly increase of 5.2% from the $74.2 million recorded for Q2 and an increase of 37.2% over revenue recorded for the third quarter of the prior year 2017.

On a year-to-date basis, total revenue for the first 9 months of this year was 28.7% higher than the level recorded for the first 9 months of 2017.

Quarterly international revenue increased 2% sequentially and represented 62% of total revenue.

Turns volume, that is orders received and shipped within the quarter, was approximately 18% of third quarter revenue.

Lower turns volume has been a reflection of extended lead times.

With this quarter's call, we will begin providing a breakdown of revenue and bookings by legacy and advanced products.

We've been working to increase the efficiency of our organization and are planning to begin reporting with our 2018 10-K, our activities as one business segment rather than the 3 businesses, BBU, VI Chip and Picor, we have reported to date.

Going forward, we will present results only on a consolidated basis and product, and marketing details will be provided pursuant to ASC 606, the new revenue recognition standard in our discussion of the sources and characteristics of our revenue through footnote disclosures.

For some time, in our filings, we have characterized our products as either legacy or advanced.

Legacy products are those associated with our Brick Business Unit, historically representing the majority of our revenue, while advanced products are more recently introduced products, reflecting advanced power conversion engines, advanced power distribution architectures, advanced control ASICs and advanced packaging technology.

For the third quarter, legacy product revenue grow -- excuse me, rose 5.5% sequentially and as a percentage of consolidated revenue, were 65%, the same level as the prior quarter.

Advanced product revenue increased 4.6% sequentially and on a relative basis, represented 35% of total revenue.

Consolidated bookings rose 4.1% for the quarter, exceeding $91 million and bringing total 1-year backlog to $116.1 million, a sequential increase of 12.6%.

Bookings for legacy products declined 6.1% sequentially.

In contrast, bookings for advanced products increased 20.6% sequentially, reflecting expansion of demand for Power-on-Package solutions, notably for AI acceleration and supercomputing applications.

We also saw incremental growth of demand for a variety of advanced chips across a range of other applications.

The shift in the mix of legacy and advanced product bookings over the last year highlights the impending transition in our business.

For the third quarter of 2017, a year ago, the percentages of total bookings for legacy and advanced products were 64% and 36% respectively.

But for the third quarter of 2018, these percentages were 56% and 44%, an indicator of further shift in revenue mix from legacy to advanced products for the coming quarters.

Also, listeners should keep in mind booking and delivery patterns can differ through legacy and advanced products.

Legacy products generally are high mix, low volume, serving a statistical customer base of nearly 10,000 customers.

Orders are generally smaller and scheduled over weeks and months, contributing to a smooth booking pattern.

In contrast, advanced products are, thus far, low mix, high volume, serving a more concentrated customer base.

Individual orders are generally much larger, and deliveries can be scheduled over quarters.

Since we're in the early stages of market penetration with many of our advanced products, particularly those ordered by OEMs and shipped to their contract manufacturers, advanced product booking patterns at any given time maybe less smooth, or as I have said before, lumpy.

I'll now turn to product profitability.

We achieved a milestone for Q3, in that our consolidated gross profit rose to 50% for the quarter, up from the second quarter's 48.4% and the Q3 2017 gross margin of 44.2%.

This is a reflection of the scalability of our business model and more specifically, the improving performance of the manufacturing process associated with our advanced chip components, which are expanding as a percentage of our total unit volume.

During the quarter, we were successful in meeting our needs for raw material inventories despite ongoing supply chain uncertainties and long lead times.

Overall, we believe our visibility has improved, but we continue to pay close attention to ensuring availability of components.

The recently implemented Section 301 tariffs on Chinese imports did not have a material impact on our cost during the quarter.

However, the cost going forward may not be inconsequential given the volume of components currently sourced from China.

We are seeking non-Chinese alternate vendors.

In addition, we have filed requests with the U.S. government for exclusions from tariffs on a limited number of components for which no alternative vendor exists.

As tariffs on Chinese imports are becoming a material percentage of our material cost, we will add a tariff surcharge to the selling price of our products until these tariffs are no longer an issue.

Turning to operating expenses.

Q3's total declined sequentially 4.6%, in part because of nonrecurring severance expenses incurred in Q2.

On a relative basis, operating expenses again declined sequentially as a percentage of revenue to seventh such quarterly decline, falling to 33.3% of revenue for Q3 from 36.7% for Q2 and 40.6% for Q1.

R&D expenses declined 6.2% sequentially, largely reflecting improved efficiency in the development of new products, and fell to 13.7% from 15.4% of revenue.

Sales and marketing expenses were essentially unchanged but declined to 13.6% from 14.3% of revenue.

G&A expenses declined 10.2%, largely related to lower stock compensation and personnel-related expenses and fell 6% from 7% of revenue.

Operating income rose to 16.7% of revenue for Q3, up from 11.2% of revenue for Q2.

These results are in line with the statements I made last quarter regarding spending trends.

As stated then, we expect operating expenses to continue their relative decline as a percentage of revenue, while expanding on an absolute basis at low single-digit percentages, largely driven by compensation costs.

Our long-term model is to reduce total operating expenses to 30% of revenues as we drive gross margins towards 60%.

I'll now turn to Dick Nagel for a quick overview of our tax position.

Dick?

For the third quarter, our effective tax rate was 1.7% and we recorded a net provision of $227,000.

During last quarter's conference call, we explained our perspective on the approximately $33 million valuation allowance we had against the value of our domestic deferred tax assets at year-end.

With one more quarter of positive results behind us and an outlook that remains positive, we have increasing support for the reduction or the release of this allowance.

However, management, pursuant to the requirements of ASC 740, concluded it was appropriate at quarter end to maintain the full allowance.

We will assess the release of the allowance at the end of the fourth quarter.

Also note, the company has been utilizing available net operating loss carryforwards and tax credits to offset taxes due on taxable income throughout the year and as of September 30, had consumed its federal NOL balance, leaving federal and state R&D tax credits, along with other tax credits, reserves and other accounts as the balance of our DTAs.

If and when we decide to release the then current valuation allowance, the amount of such release would be lower than the figure implied by our 12/31/17 balance of DTAs other than the NOLs.

However, at the present time, we cannot reasonably estimate what the balance of DTAs may be at the time of release, the amount of the allowances to be released or the timing of the potential release.

Nevertheless, as stated last quarter, we believe it is more likely than not we will release some portion, if not all, of the then current allowance within the next 3 quarters.

Jamie?

Thank you.

So back to the Q3 P&L.

We recorded net income after minority interest of $13 million, representing a 66% sequential increase in after-tax earnings.

Diluted EPS totaled $0.32, up from Q2's $0.19 and Q1's $0.10.

Our quarter end diluted share count was 41,124,000 shares.

Turning to the balance sheet.

Cash and cash equivalents sequentially increased $14.3 million for the third quarter and ended at $68.2 million.

This increase reflects operating cash flow of $14.3 million and $3.4 million of proceeds from the exercise of employee stock options during the quarter, offset by CapEx of $3.2 million.

On a year-to-date basis, cash has increased by $24 million.

Net trade receivables were little changed for the quarter, with DSOs actually declining to 41 days and no indications of portfolio risk.

Net inventories increased modestly, up $1.7 million or 4%, largely reflecting rising material and component purchases to shift our increasing backlog.

Annualized inventory turns grew slightly to 3.6.

Winding up my review of the third quarter.

Total employee headcount as of 9/30/18 declined to 1,018 from 1,024 at the prior quarter end, largely due to lower temporary staffing, a reflection of improved factory-loading for the quarter.

Total full-time employment was essentially unchanged, up a net 2 from 972 to 974.

I'll now provide an update on our capacity expansion.

The Q3 capital expenditure total of $3.3 million does not fully capture the level of investment activity underway.

We have approximately $15 million of production equipment on order, of which $4 million is scheduled to be put in service during the fourth quarter.

This additional equipment to be deployed within our existing 230,000 square-foot factory in Andover will increase its capacity to approximately $500 million in annual revenue.

The next increment of capacity from approximately this $500 million threshold to $750 million of revenue will be deployed within an approximately 90,000 square-foot extension of our factory to be built on land already owned by Vicor as early as Q1 2020.

In response to the request of a large customer, we also are pursuing the possibility of establishing a manufacturing facility in Asia, possibly with an Asian partner.

Turning to our outlook.

Given our increased backlog and visibility into customer requirements, we are forecasting a sequential quarterly increase in consolidated revenue.

This sequential increase may be relatively small, pending further increases in manufacturing capacity for advanced products.

Listeners should keep the following in mind.

First, despite our planning and safety stock methodology, we remain exposed to raw material and component availability risks.

Second, we do not yet completely know what impacts Section 301 import tariffs will have on near-term results.

Third, we have seen early indications that the Chinese markets may be cooling, an indirect reaction to the U.S.-China trade dispute.

And finally, our backlog, at a record high level, includes substantial deliveries in the first and second quarters of 2019, which reflects the increasing proportion of our backlog made up of orders for long lead time advanced products.

Accordingly, I must remind listeners, as I do each time I speak with you, our operating and financial forecasts are subject to unanticipated changes, many of which are caused by factors and influences outside of our control.

With that, I'll turn the call over to Patrizio.

As addressed by Jamie, the third quarter 2018 was characterized by improved financial performance, notably 50% gross margins and appreciably higher profitability.

These improvements reflect increasing productivity and broadening adoption of modular power system solutions, driven by the distinctly superior performance.

In Q3, we hired a Global Automotive Business Development Vice President, Patrick Wadden, to lead sales and marketing in the automotive segment.

Vicor is already supplying a high-density power system for the leading developer of Level 5 autonomous driving systems and developing power system solutions for other automotive applications.

With significant long-term potential for Vicor products in autonomous driving and more generally, the electrification of vehicles, we've hired Patrick to expand the use of Vicor safety in the automotive 48-volt power systems market, with a mix of product sales and technology license agreements.

As always, I'd like to limit my prepared remarks as I would rather answer your penetrating questions, so I would open the call.

Mark?

(Operator Instructions) Your first question comes from the line of Quinn Bolton.

I wanted to start first, just thank you for the color on this, both between legacy and advanced products.

Obviously, a lot of excitement growing around the 48-volt architectures.

Is the advanced products largely or entirely consist of your 48-volt solutions?

Or does it include other power and power delivery other than 48-volt?

So as of now, it's primarily 48 to the point of load.

But as you might have seen from a press release earlier this week, before too long, it will be a mix of power delivery to the point of load and power delivery to the 48-volt bus.

Fundamentally, our strategy is to provide connectivity from the power source, whatever that might be, high-voltage, DC, single-phase or 3-phase safety lines to 48-volt as a stepping stone to the point of load.

The emphasis up to this point has been for the 48 -- from 48 to the point of load, but we've had initial applications to leverage our advanced products, soup to nuts, from the power source to the load.

And there's going to be more of those kinds of applications going forward.

So I guess just a clarification there.

It sounded like most of the revenue today then is from 48-volt stepping down to either 12-volt or directly to the low voltage.

Or is it including some of the 3-phase AC to 48-volt converters that might sit sort of at the bottom of the rack?

Well, it does involve, let's put it this way, a few millions dollars of 3-phase to 48.

There's going to be a lot more of that in the future.

It does involve, primarily, to your earlier point, 48 to the point of load, Intel processor, 1.8-volt, or in AI, XPU at less than 1 volt and hundreds of amperes of Quadro at dead-low voltage provided directly from 48-volt.

We also make our -- selling in the initial plan to this to a growing list of customers, bus converters that take the 48-volt down to 12-volt for intermediate bus applications.

And those products themselves are far superior, sometimes more in size and more cost-effective than alternative solutions for intermediate bus conversions.

By the way, those products, in addition to being relevant into the sector space or in retrofitting 12-volt racks with advanced GPUs that are powered from 48-volt, they're also going to be seen, I think, before too long in other types of applications, including automotive applications, where the electrification infrastructure is changing.

As we all know, it's moving to 48-volt, but there are 12-volt legacy loads that are going to be around for quite some time, and those loads need to be fed at the existing operational voltages, ideally directly from a 48-volt infrastructure, ideally without involving the heavy and expensive copper wiring that a 12-volt infrastructure requires.

So part of our strategy, in summary, an essential part of our strategy is to provide connectivity from and to any intermediate voltage, 48, 12, it doesn't matter.

The advantages of the technology are very comprehensively applied to each relevant node.

Understanding that the step-down to 12-volt today just allows you to interface with some of the legacy architectures, it seems, longer term, a direct conversion from 48 down to the low voltage makes more sense.

Can you just give us some sense how Vicor is positioned against some of your leading competitors in this space, Analog Devices, Infineon, Maxim, Monolithic Power?

In terms of the approach, I think you're taking a direct 48-volt down to the low voltage, where some of your competitors may have to do a 2-phase or a 2-stage conversion with an intermediate voltage.

Can you just talk about the architecture, what advantages you have relative to competitors when you look at the 48-volt space?

Well, it could be a long answer.

I'll try to keep it as short as it can be and at the high level without getting into too much technical detail.

But to your point, the competition for a 48-volt power system is being done in a number of paths, including direct 48 to point of load.

Some of these companies that you referenced are on the second or third try, involving, first, the rack conversion, which did not work out.

And then, essentially, for the most part, going back to the starting point, with intermediate bus architecture, where, to your point, this is an intermediate step to 12-volt, in effect, as a stepping stone on the way down from 48-volt infrastructures to the point of load.

There are significant handicaps that come with the 2-step approach.

And some of them are, in effect, as fundamental as Ohm's law, very basic.

I would say that, and I obviously hold it in extremely high regard, many of the companies that you referenced have accomplished all great things in their past.

But frankly, particularly the semiconductor companies, they don't fully understand power.

They believe -- or some of them believe that a better switch, a better semiconductor switch like a GaN device, will provide a magic bullet to solve each and every problem, and that's really not the case, far from it.

And I would say that GaN, in particular, is irrelevant at the point of load.

It's irrelevant in terms of Power-on-Package solutions.

It really wouldn't make any difference.

And even in upfront converters that provide conversion from high-voltage buses to 48-volt, or from 48-volt to, say, to 12-volt intermediate bus, GaN, as it stands today, doesn't offer any advantage in -- it costs more money than silicon.

It's still not nearly as mature as silicon.

It's got a number of limitations.

But most of all, it doesn't really offer an appreciable or any efficiency advantage if, as we do, you have the right power conversion technology and the right power distribution architecture.

So to paraphrase, a -- therefore purchasing at one of our customers, asked by a person that joined the company recently in our presence, what is Vicor's competition?

The answer was Vicor has no competition.

And the reason why we have no competition is that we've been working at this for nearly 15 years.

We have addressed all the facets of a very complex problem that, again, involves a lot more than a better switch, which is an element in a converter, among many, with many other ones providing -- or presenting limitations conservatively more significant than the switch itself.

Let me come at it from the -- another angle.

As I look at one of our higher-voltage converters, where a relatively high-voltage device such as GaN could be used, and I look at the power that we dissipate in a silicon switch, that amounts to a small fraction of 1%.

So even if that device, the silicon device, where we are placed to buy a perfectly ideal, mature, cost-effective GaN device, the upside in terms of reduction and loss will be negligible.

But somehow, that's what a lot of the industry appears to be focused on, and I will submit a view of my experience in the field that this is a mispriced priority, there's a lot more to making high-density, high-efficiency, cost-effective power converters than replacing a silicon fab with a GaN fab.

Maybe just a quick one for Jamie.

Jamie, from your prepared comments, it sounds like the component availability, discrete ceramic capacitors, while still tight, may be incrementally getting better.

Am I reading those comments correctly?

That's correct.

I mean, we're far from out of it, but we're in good shape.

Your next question comes from the line of Alan Hicks.

But going back to the gallium nitride technology, has any of your competitors got any traction with that so far?

None, none in the power conversion field.

I think that there is some very minimal traction for some specialty applications.

But again, we keep gauging progress with respect to the GaN fabs.

We benchmark them.

We've done that essentially again.

And it is something that's been high frankly for at least 5 years and whose drive to the finish line keeps pushing out.

And even once it gets there, and I have no doubt that we'll get there, its impact on power conversion will be very negligible.

Okay.

And then the area that you've talked about in the past, some of the CPUs and XPUs are coming out with very fine line widths, 7 to 10 nanometers, what gives you the advantage there?

Well, so at 7 nanometers, the processors are fed by a voltage, specifically 0.6, 0.7 volts.

To efficiently deliver that voltage, of course, that in many applications now are reaching up and getting past the 1,000 amperes, what you need, and which is a unique attribute of our technology, is a current multiplier, not a device such as traditional back converters operating from 12-volt that take to the 12-volt and average it down to a lower voltage.

That works reasonably well when you average down from 12 to 1.8, which is the standard voltage the Intel processors operate from largely because within those processors, there is a further step-down that takes place within a converter structure that Intel has developed.

When you stretch that methodology all the way down to 0.6-volt, it gets that much harder.

The so-called duty cycles get narrower and narrower and the dynamic performance, and I don't want to get too technical, but let's put it this way.

These fundamental limitations to the proposition of averaging down a voltage, it's a little bit like trying to make a low -- water that is a little bit warmer than the cold water faucet, by mixing it in with the hot water feed, where, in effect, you're trying to reach a water temperature that is close to the cold water.

There is a much better methodology for this, which we patented.

It involves so-called current multipliers.

They, instead of averaging down the high-voltage, they divide it, and they can divide it by an arbitrary large factor.

So we have applications where we divide down by a factor of 48.

We have other applications where we divide down by a factor of 64.

We've got some new applications we're going to be dividing down by a factor of 72.

So you take 48-volt, a 54-volt, you divide it down by 72, and you're right in the sweet spot of what the 7-nanometer processor requires.

And you get there with all of the right attributes, very fast response to low transients, very low noise characteristics, such that with our Power-on-Package and our packaging technology, you can integrate it within the XPU package itself or right next to it.

At what point would there be a tipping point where this technology will be required to enable a majority of processors?

That, to your point, is taking place now.

We -- just within the last 6 or 7 months, we have been approached by half a dozen other major competitors vying for their share of the AI market opportunity.

And universally, they are all relying on Power-on-Package, our current multipliers.

There is no GaN anywhere close to that.

There is not a solution anywhere close to that.

So there is no competitor that's even close?

There is no competitor that is even close because we are, I could argue, 10 years ahead of the competition, and we have nearly 100 patents just standing in the way of the competition.

And we have been working diligently over a long time to planting a minefield for any -- as who of those competitors would want to try and chase our track, never mind the fact that there are fundamental technical challenges to doing that independently of the IP.

Okay.

And on the new RFM product, how long will it take for that to go into high-volume production?

Well, that's often an architectural change within the system.

So don't expect an immediate self-topping revenue.

We have had a lead customer for that product.

As you can see, from the recent announcement, we're now broadening the offering to involve other customers, other input voltage ranges.

The engagement with the first customer was for Japanese 200-volt AC mains.

We've since developed a broader input range capability.

And by the way, we also have a whole new family of maturation technology products, our fourth ASIC generation, that will take the RFM and all these other products to yet another level of efficiency and density and cost-effectiveness.

So when it comes to RFM applications, front-end applications, I expect we're going to have some level of penetration with RFM that's been recently announced.

We're going to have a much deeper penetration as we release the first 4G RFM products next year.

And then on your announcement with hiring the manager for the automobile industry, there was a mention of licensing.

Do you have any prospects for licensing there?

Yes, we've been approached by a number of parties with an interest in taking a license for certain aspects of our technology for automotive application.

Patrick was actually at a conference in Berlin just within the last couple of days.

He's reported back that our NBM or the 48 to 12 bus converter was the talk of the show at that conference because it represents a tremendous opportunity for eliminating wiring and simplifying and reducing the cost of automotive system.

So whether it's the NBM or other products such as the power strip, which is a problem we have developed for the complex power system of a Level 5 autonomous driving capability.

Our technology, when it comes to automotive applications, has a number of sweet spots.

So are you in discussions already for licensing?

Yes.

We expect that these opportunities are going to expand.

Obviously, hiring Patrick was a key step to a broader automotive strategy.

We're going to take some time to assess all of the opportunities and prioritize our pursuits both from a product development and from a licensing partnering opportunity perspective.

Okay.

That will be both -- the automobile companies or auto parts companies or both?

Both.

Your next question comes from the line of Don McKenna.

I wanted to ask you a couple of things.

Jamie, first of all, on the increased lead times that you were talking about with the current backlog, is that because of capacity constraints you have?

Or is it because some of the raw materials you've been having trouble getting?

And then the second question I wanted to ask was about the expansion.

It seems like it's a turnaround from where we were 3 months ago with the new facility.

Could you give a little more detail on that, what your partner in Asia might be?

Let's start with the -- his answer.

Okay.

So I think the first part of the answer had to do with the lead times and capacity constraints.

So we've had some capacity constraints within the third quarter for advanced products, no capacity constraint on legacy products.

And that's why we've been placing all of this and starting to install additional equipment to expand the capacity.

As suggested in Jamie's prepared remarks, there's going to be a significant expansion taking place in Q4, actually next month in November, and further expansion taking place in Q1.

This expansion in capacity for advanced products is needed to meet the forecast of demand for 2019.

So as suggested in the prepared remarks, we are in a transition from the majority of the revenues being legacy products to come 2019, the majority of the revenues being advanced products.

And we're working proactively to make sure that capacity does not stand in the way of that.

We do not let customers down and we are very good at scaling up capacity and deploying it in time to stay ahead of demand.

But the recent couple of months had been a little challenging.

It's a challenge that is going to get -- we've had to add a lot of overtime and that kind of thing to ensure that we brought about the capacity that was necessary to take care of consumer needs.

We're going to provide some relief with respect to that with the selection of equipment that is about to take place, and then more so come the first quarter.

In the second quarter, demand with...

Real estate.

Real estate.

So yes, the -- we have fine-tuned our strategy with respect to capacity expansion in terms of where to do it and whom to do it with.

And that's come about as a result of a number of revelations.

So one has to do with the timing for breaking out and installing capacity in a different location away -- further away from our existing manufacturing facility.

As it turns out, we have enough land to expand the existing facility, to expand it and bring about nearly 50% increase in -- to all our capacity.

And this very leverage, with respect to the advanced products because in terms of revenue per square foot, it's a much more favorable multiplier.

We've also got an input from a potentially very significant customer that they would very much like to have us have a presence in Asia for a variety of reasons.

So we started looking at that possibility.

And this is not a near-term opportunity, it's something that will take some time to fully sort itself out and be executed upon.

We feel comfortable with the near-term strategy of adding essentially 50% increment of revenue capacity as an appendix to our existing facility.

All right.

Is the idea, the additional capacity of 90,000 square feet being available for first quarter 2020, is that an indication that you will reach your 500,000 -- or $500 million, rather, capabilities at the end of 2019?

I wouldn't draw that conclusion.

I think it's an indication of the factors I was suggesting earlier.

We want to make sure, when it comes to expanding capacity, with respect to in particular brick-and-mortar, which has got longer lead times than procuring additional equipment, that we are way ahead of our needs so that, that doesn't become a bottleneck so to be clear, purchasing and deploying additional equipment within existing walls is essentially a 3-, 4-month proposition.

And obviously, it takes a little longer than that to get new facilities built from the ground up.

Your next question comes from the line of John Dillon.

So my first question is, has anything fundamentally changed with Vicor in the last 90 days?

Well, I think the evidence keeps building on the measure opportunity we have with respect to AI applications.

As suggested earlier, there's no competition.

The only way competitors looking to establish some competitive position powering 7-nanometer processors, the only way to have a competitive solution is with our technology.

So that fact has become abundant really within the last 90 days as more and more companies come to us, they come here or approach us to pursue solutions for them, working with other partners, to enable advanced XPU's sub 1-volt, from 100 to 1,000 amperes.

I think that the other significant take away in my mind within the last 3 months has been validation of our fourth-generation ASICs at both with the 4G PRM and the 4G VTM.

We're very far along with these devices.

We're making initial products with the PRM.

Very close to making VTM and car multiplier products using the VTM 4G controller.

These devices bring about a major advance in terms of performance, density.

They eliminate a lot of the component count of early generation control silicon.

They enable much more advanced products, including, among other things, more advanced versions of the RFM.

So I think, in my mind, those have been very significant milestones in terms of continue to advance the state of our art.

We are on track to -- and we've got our own version of a Moore's Law, right?

Moore's Law, which has been in effect for quite some time came to an end recently in terms of further advances with respect to processor technology.

Our own version of Moore's Law is continuing to increase the density and efficiency of our products by about 20% every couple of years.

And we keep being on track with that.

And as we do that, the technological gap between Vicor and the competition gets wider as opposed to getting narrower.

We're on that track and our 4G silicon and the Power-on-Package technology and further advances with respect to our chief technology at large are continuing to build that strong competitive capability.

So nothing negative fundamentally has changed.

All you see is positive fundamental changes, and the fact that you're growing your market, you're attracting newer customers and you're accelerating into the AI world, that's what I think I'm hearing.

Is that right?

Yes.

And let me mention one more thing, cost-effectiveness.

In many products, take the NBM as an example, we have, by far, the lowest cost factor than any competitive product.

Not only are we 1/3 the size, but perhaps not surprisingly, being 1/3 the size, we're not quite 1/3 the cost, our cost, but we're substantially lower cost than any competitive alternative.

That's just great to hear because I know, a long time in the past, you just always have the technology superiority but not always the lowest cost, so that's really good to hear.

Well, you might have heard me in the past saying that -- and it remains true, that superior technology is not just superior in terms of efficiency, density, fast response, low noise.

Last but not least, it's got to be superior in terms of cost.

So when competitors say, we are 2x or 3x the cost, they're looking at a Vicor from 10 years ago.

They are either oblivious or purposely ignoring the realities of today.

Exactly.

Exactly.

You mentioned in the pre-announcement, and you had a nice answer to that about the GA, converters and all.

But what I was wondering, has their claims with lower cost in GA built parts, has that affected your wins in the lower-power CPUs that you power?

For example, the Intel chips don't -- they like said, they're only 1.8 volts instead 0.7.

Are they starting to pick up traction in those wins?

Or is that still consistent with you?

Well, so they have all the traction in those wins up to the point in time in which we successfully penetrated some of that business.

To be clear, and I suggested this in answer to an earlier question, our competitive advantage, our 0.6-volt, that is for the AI processors, is in one way of looking at it, several times greater than if these are 1.8-volt.

So I don't expect that our competitive position at 1.8-volt, which is where the Intel processors have been and may continue to be, may or may not, time will tell, that's not our strong suit.

Our strong suit is in powering directly 7-nanometer nodes.

And in doing so, whether or not competitors take an intermediate step, but through 12-volt.

Right.

But you're not losing any of your Intel business?

No.

Okay.

So is there any reason in the world why you're company would be valued at 40% less than it was 90 days ago?

Can you think of anything?

I'm not going to answer that question.

I cannot think of a reason but...

Okay.

I can't either.

You didn't really talk about bookings...

Other than the general market malaise, right?

We're all mindful of been going on in general.

But there is no company-specific rationale whatsoever.

There's no fundamental reason.

There's no changes or anything that would fundamentally affect your company?

I think our strength is growing.

I can tell you that there isn't a recognized player in the either CPU space or GPU or general XPU space that is not talking to us or coming to visit us to pursue opportunities for Vicor to give them a competitive advantage.

Nobody wants to be handicapped because ultimately the issue from the perspective of these companies is, what's going to help them win against their competitors.

And they are very smart, right?

They wouldn't be wanting to handicap their competitive stance by following a crowd that has failed to provide the level of advances in power system density, efficiency, flexibility, Power-on-Package technology that these systems require in order to achieve the performance that gets them ahead of their competitors.

Right.

I mean, you enable a higher-performance computer.

So of course, they saw them, right?

We do.

And we enable a higher-performance GPU and we're going to be enabling a lot of higher-performance ASICs.

Yes.

And if you wanted to, you can convert your designs to GA also, correct, to GaN?

But you don't need to.

We have actually benchmarked that.

We periodically do that, and there is no benefit.

No benefit?

Got you.

There is a cost penalty.

Got you.

Cost penalty.

One last question then, bookings this quarter, are they still on track?

I mean, do you think you'll still have about the same bookings for -- that you saw last quarter?

I'm not going to stick my neck out with respect to that.

I think that we've obviously been building up the backlog.

The backlog as it stands, they gradually -- as suggested in Jamie's prepared remarks, we see the revenues picking up at a modest rate this quarter, and we'll have to wait and see with respect to the bookings.

But generally speaking, as I look...

I imagine some companies trying to pump in and gave you pretty good bookings upfront for...

Let me put it this way.

As we look -- you might recall me saying several quarters ago that we need to get past the stagnation of a long timeframe in which we're building up our technology or capability, but we have not transcended yet in design wins and revenue growth.

And I think I mentioned back then that the key milestone was to get past -- or to the $300 million level as a yearly run rate.

They would open up more opportunities.

It will signify the kind of traction and more widespread adoption that we're now seeing.

And I think that's coming through just as expected.

By the way, in the past, I think many companies had some level of hesitation with respect to doing business with Vicor, having to do with the fact that we were viewed as a supplier of specialty product, a niche player that would be sort of a supplier of last resort if you needed to have higher performance.

That old perception has changed, has continued to change.

Again, the calculus now is along the lines of, I outlined them a little while ago, Vicor is getting traction, Vicor is getting to critical mass.

I got a lot more to lose from not doing business with Vicor than the other way around.

Your next question comes from the line of [Peter Lowe].

So with one confusion, you could say, we saw your release related to the Nvidia V100 cards, which we understand are 12-volt and are plugging into 12-volt servers, but you are selling a 48-volt product, so how does that work?

Well, there is no claim whatsoever that we have taken all of the business of any 1 customer, right?

So without mentioning customers' names, you can imagine that any customer undergoing a transition from 12 to 48 may, for a variety of reasons, continue to develop some products at 12-volt.

This is not an all or none proposition.

It's a progressive series of steps, whether it's the customer you mentioned or any other customer.

I have no doubt whatsoever that before too long, it's all going to be 48-volt, it's all going to be 48-volt in the data center space, and it's going to be 48-volt in automotive.

Okay.

So if the card or the product is plugging into a 12-volt server, then it will not use your product.

But if it's going into a 48-volt server, then it would use your product.

Is that fair?

Not necessarily because what we also do with the NBM, and we're going to be doing with more products of that kind, these are variation converters that can convert 48 to 12 and convert 12 to 48.

So whether it's a data center requirement with a 12-volt infrastructure where a 48-volt GPU needs to be powered from 12-volt and the NBM can convert 12 to 48 to power the 48-volt load.

Or as I suggested earlier, in automotive application where there's an opportunity to use the 48-volt battery, get rid of the 12-volt battery.

If they'll distribute 12-volt with heavy gauge wire, distribute 48-volt and then power a legacy 12-volt load with NBM, we can go either way.

And fundamentally, we got the best of all solutions from 48-volt to the actual point of load.

We also have, with the NBM, by far, the best, the highest performance, highest density and lowest cost solution for either converting 48 to 12, for converting 12 to 48.

Okay.

Can you give us a little more color or feel for what your ASP is for a -- what would be a total solution to deliver power from the wall to the CPU, how much content do you have, either in number or relative versus a more traditional product?

Well, for competitive reasons, I'm not going to mention a specific number.

But then obviously, the number is so much dependent on the particular opportunities, volume, the revenue opportunity and the margin opportunity as a function of revenue level that we want to achieve.

But I go back to my earlier statement which is both in front-end conversion and in point-of-load conversion, we have the lowest cost curve.

If you measure the cost of our building blocks in terms of cents per watt or cents per amp, we have the lowest cost curve.

The technology enables the lowest cost solution is densest, and with that, it's got less of everything that factors into the cost of the product.

It's got less silicon, it's got less copper, it's got less PC board, it's got less of a packaging cost.

Now the delivery of this is, to some degree, dependent on volume, right?

Because we've had an infrastructure that we've been paying for which needs to amortize over larger volumes.

So our cost, and with the margins, get better as the volume goes up.

And as that happens, we obviously want to leverage the reduction in cost to offer a more competitive pricing to our customers to make it more attractive for them to adopt the solution.

I'm not saying that today we are selling in every application at the lowest cents per watt or the lowest cents per amp, we don't.

We sell, to some degree, on value.

But I'm saying that the cost structure of our engines and everything that goes into it is evidently the lowest cost.

Got you.

Earlier in the call, you talked about Chinese markets are cooling.

Can you provide more color on that?

What segment of the Chinese market is cooling?

Are you seeing cooling anywhere else in any of the geography, U.S., Europe?

Is it more industrial focused?

Server focused?

Data center focused?

When did the cooling start, also, please?

So there's been some cooling taking place over the last several months.

Let me answer it this way.

Getting power semiconductors fabs -- I'm not talking about GaN FETs now, I'm talking about good old silicon FETs.

Getting FETs months ago was a lot harder than it is today.

Getting ceramic capacitors several months ago was a lot harder than it is today.

And that's symptomatic of the fact that, that -- that the strains on the supply lines have eased, I think, considerably over the last several months.

I mean, frankly, from our perspective, this is a good thing.

We would not want to be constrained by availability of components.

There's plenty of opportunity for growth next year and the year after that.

We don't need the most robust Chinese economy or the most robust global economy for us to deliver growth.

So this environment is perfectly fine for us.

I think, frankly, we have some level of concern with respect to the [Broadwell] pipeline months ago.

And in some instances, we have to pay premium prices for some of these components.

So that's no longer the case or much less of a case today.

But are you seeing that cooling in your own business on -- probably more on the legacy side versus the advanced?

Well, so bookings last quarter were sequentially lower for Bricks than in the prior quarter.

And Jamie gave you some quantitative measure of that.

I think Bricks' bookings declined by what percentage?

What was it, 5%, 6%.

It was 5% or 6% down from the prior quarter.

That was also very, very strong for the first half.

Yes, but we see that it declined.

However, the advanced products bookings went up, I think, by slightly over 20%, right?

So -- and that's really indicative of the general point that I was focused on, which is when you look at the cooling Chinese economy or global demand for electronic products and you correlate it to a legacy product, which is a mature product, where there's not going to be a growth, right, we recognize therefore a long time that it's good for the Bricks to hold essentially level revenues in a normal environment.

Well, those products, guess what, they are going to be affected by a cooling Chinese economy or a cooling global economy.

But when it comes to the advanced products, the much bigger driver there is the traction with new customers and new applications.

That was the temperature of the economy, right?

Because that's the second order effect.

Of course.

Within that 6% decline in bookings, can you isolate it to an end market?

Is it server?

Is it auto?

Industrial?

So we sell Bricks into a diverse set of customers and applications from transportation to industrial application, to some communications applications.

I mean, these are products, as mentioned in the prepared remarks, that are sold to something of the order of 10,000 customers.

So there is a good barometer of the general state of demand.

And I think they are -- that 6% reduction from Q2 to Q3 is itself indicative independently of the other indicators having to do with the components we use, caps and fabs, of a cooling economy.

But again, that's got to be contrasted with the significant growth in advanced products, which has got really nothing to do with a warming or cooling economy.

It's got everything to do with traction with new customers and new applications.

And the next question is coming from the line of Jim Bartlett.

Could you give us an idea of when you see significant impact, and this I mean by the halves of various things?

First of all, on the NBM.

Second, with the RFM products.

And third, with -- and this is interrelated if -- your new 4G.

Okay.

So we see significant growth in revenues next year from AI applications.

I will put that at the top of the list that you just referenced.

So MCD, MCMs, point-of-load solutions, they're going to drive significantly on growth next year.

The NBM has got design wins, a large multi-business design win, but it was ranked below that, the level of opportunity with point-of-load, 48-volt, direct to point-of-load solutions in the near term.

Now in automotive, in longer term with legacy 12-volt buses, the NBM has got tremendous opportunity.

But because of the architectural changes that relate to that, I think that's going to take time to fully develop.

And a similar comment would apply to the RFM.

So 4G is going to start shipping in volume to customers in the second half of 2019.

It's going to drive much further advances in performance, in cost-effectiveness of all of these building blocks.

It's a control system that is universally applicable to high-voltage process.

It could be 1,000 volts, 800 volts or it can be 0.6 volts, it doesn't matter.

We've got in a control system with 4G capability that would result to decimal and do so with extremely high performance and cost-effectiveness.

But that -- if you were to ask specifically what is the revenue that is earmarked to 4G, that's going to start in the second half for next year.

And in terms of being a significant share of the business, it will have to be 2020.

And again, I was confused on the front-end part of it, with RFM products.

Where does that start to have a significant impact?

Well, so there's going to be significant RFM businesses but on a lesser scale than the other 2. We'll rank them again.

Point-of-load, well ahead of the other things.

NBM, below that in the short term.

RFM, below that in the short term.

But longer term, the RFM type of building block represents essentially half of the pie.

One way of looking at it is that these -- as many cents per watt go -- taking the watts from 3-phase AC to 48-volt, as there is taking it from different directions, going from 48 to the point-of-load.

But these are common and we either need to take power from the worldwide AC mains, single phase, 3-phase, to the 48/54-volt bus that is going to be at the heart of the entire point-of-load infrastructure.

And that's particularly the case in any stationary application in -- and to some extent, they can also be the case in other modular applications.

And where do you see your first automotive revenue?

Well, we've got some de minimis automotive revenue now but with Patrick, we have a target, an initial target of $100 million in automotive revenues.

I'm not going to mention the timeframe of that.

I think certainly, since he's come on board, he has commented about extreme excitement, extreme opportunity from the meetings he's had with some major companies and I think that the timeframe in which that initial $100 million target can be achieved, I believe, has moved in relative to his expectations before joining the company.

Speaking of timeframe, you've mentioned -- or Jamie mentioned that 30% operating cost at 60% gross margin, driving towards that.

Could you give some timeframe on that?

I'm not going to get pinned down on that.

I think that, obviously, we've had some baseline progression on the operating expense reduction front.

We're actually closer.

I set an internal target of 30% early in the year, and some of this cash, as we thought it would take longer, to invest that again to get full suit of the lower 38%.

I think it would be easy and faster to get to 30% of revenue expenses than to get up to 60% on margins.

To some extent, that's going to be a function of how aggressive we choose to go on driving business growth.

So we -- that's a lever we want to keep at our disposal with respect to driving long-term dominance in the marketplace.

Would a 5-year time frame be reasonable?

Well, so if you were to draw a -- some kind of linear extrapolation from the progression that's taken place in the last year, on an initial -- modest revenue step up from $200 million, $250 million level to $320 million run rate of the recent quarter, you would draw the conclusion that it could happen a lot quicker than 5 years.

I'm not going to go there because, again, to some degree, we want to retain all the flexibility we should have with respect to driving the tradeoff between revenue growth, economies of scale and short-term gross margin, short-term profitability.

I think it's safe to say that that's a good problem to have, and we're going to drive that bus in a way that may evolve over time, depending on a variety of factors.

Yes, it's only taken $400 million but -- and a lot of sweat, but we are, I think, in a much better place today.

Thank you.

And with that, we'll look forward to talking with you in a few months.

Have a good night.

Thank you very much.

Ladies and gentlemen, this concludes your conference call for today.

Thank you for joining.

You may now disconnect.