Vicor Corporation (VICR) Q2 2017 Earnings Report, Transcript and Summary

Good day, ladies and gentlemen, and welcome to the Vicor Earnings Result for the Second Quarter ended June 30, 2017 Conference Call. My name is Tracy, and I will be your operator for today. [Operator instructions] I'd now like to turn the conference over to your host for today, Mr. Jamie Simms, CFO. Please proceed.

Thank you, Tracy. Good afternoon, everyone, and welcome to Vicor's second quarter 2017 earnings call. I'm Jamie Simms, Chief Financial Officer and with me here in Andover are Patrizio Vinciarelli, Chief Executive Officer; and Dick Nagel, our Chief Accounting Officer. Today, we issued a press release summarizing our financial results for the period ended June 30. This press release is available on the Investor Relations page of our website, www.vicorpower.com. We also filed a form 8-K earlier today with the SEC related to the issuance of this press release. As always, I remind listeners that this conference call is being recorded and is the copyrighted property of Vicor Corporation. I also remind you various remarks we may make during this call may constitute forward-looking statements for the purposes of the safe harbor provisions under the Private Securities Litigation Reform Act of 1995. Except for historical information contained in the call, the matters discussed on this call, including the statements regarding customers, opportunities, 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 in or implied by any of our remarks today. The risks and uncertainties we face are discussed in item 1A of our 2016 Form 10-K, which we filed with the SEC on March 7, 2017. Please note, the information provided during this conference call is accurate only as of today, Thursday, July 27, 2017. 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 the call will be available beginning at midnight tonight through August 11, 2017. The replay dial-in number is (888) 286-8010, followed by the passcode 98163065. In addition, a webcast replay of today's call will be available shortly on the Investor Relations page of our website. I'll start this afternoon's discussion with a review of our financial performance for the second quarter, and Patrizio will follow with his comments, after which, we will take your questions regarding our business. Beginning with consolidated results. Vicor recorded, as stated in this afternoon's press release, a net loss for the second quarter of $459,000 or $0.01 per share compared to the first quarter loss of $974,000 or $0.02 per share. Consolidated revenues for the first quarter increased to $57.7 million from $54.5 million for the prior quarter, representing a sequential increase of 6%. Gross margin dollars rose 9.6% sequentially. Despite challenges in ramping new products, gross profit margin as a percentage of revenue rose to 44.9% from the prior quarter's 43.4%. For the second quarter, we recorded an operating loss of $538,000 about $840,000 less than the prior quarter's $1.4 million operating loss. I'll now break down our revenue for the second quarter by operating segment. The Brick Business Unit, consisting of our Andover-based Brick product lines, Vicor Custom Power and the Brick-based activities of our Japanese subsidiary, VJCL, recorded a 4.1% increase in revenue, largely made up of increased shipments from our Vicor Custom Power operations. For Q2, our custom revenue from three subsidiaries recovered to the level recorded prior to the consolidation of our six locations into three, which occurred at the end of 2015. For the third consecutive quarter, VI Chip recorded the highest dollar revenue increase of our segments, $2.1 million sequentially, representing a 17% increase. Picor actually recorded a decline of $409,000 in sequential revenue, after recording a 39% increase for the prior quarter, but again, reflecting the idiosyncrasies of shipping schedules for Picor's SiP PRM portion of our VR 13.0 solution. International revenue, which we identify by the ship-to address, increased 9.7%, reflecting increased shipments of VI Chip products to Asian contract manufacturers. Brick exports to Asia were steady, with further recovery seen from China. Turns volume, that is orders received and shipped within the quarter was 36% of Q2 revenue. To conclude on consolidated revenue. Recognized stocking distribution revenue rose 7.3% quarter-to-quarter, totaling just under $5 million. Listeners may be aware of our pending adoption of revised revenue recognition guidance under ASC 606, revenue from contracts with customers, on January 1, 2018. With the adoption of this new accounting standard, we will no longer defer revenue recognition for stocking distributors, but will recognize the value of the products when the shipment to the distributor leaves our facility, rather than when the distributor sells the products to the end customer. And we also will record a reserve based on our estimates of possible returns and price allowances. We do not anticipate this change in reporting will have a material impact on our reported results. For an update on our implementation of ASC 606, please refer to footnote 11 to our consolidated financial statements in the upcoming Form 10-Q, which we expect to file with the SEC on or about July 31. Turning to product level profitability, as stated a moment ago, gross profit -- consolidated gross profit margin as a percentage of sales increased to 44.9% from 43.4% for the first quarter. Gross margin dollars rose almost $2.3 million or 9.6%. The primary contributor to higher gross margin was the improved overhead absorption of fixed manufacturing cost resulting from increased production volumes, despite inefficiencies caused by ramping new products. This was particularly the case for VI Chip, which shipped approximately 18% more units quarter-to-quarter. Volumes of VI -- of ChiP VTMs for VR 13.0 Data Center Solutions were essentially flat sequentially, after the first quarter's ramp, as some June shipments were delayed by contract manufacturers coping with their own supply chain issues, notably, extended lead times for DDR4 memory components. However, gross margins improved for VI ChiP VTMs. We did experience a meaningful increase in first-generation VI ChiP BCMs, which now have a gross profit margin well above 50%. Both VI ChiP -- excuse me, ChiP VTM and first-generation BCM margins reflect the impact of higher production volumes on the standard costs of our materials and labor as well as improved yields. The increase in VI Chip units volume and dollar revenue was primarily driven by a surge in shipments of our latest Factorized Power Architecture solution for powering servers, consisting of preproduction units of our breakthrough Power-on-Package 48-volt processor board level components, complemented by preproduction units of our new three-phase AC to 48-volt DC front end. During last quarter's earnings call, we discussed how we were preparing for these scheduled deliveries for a supercomputing application. Patrizio will comment on these important new products in his remarks, but I highlight here the impact of these shipments on our second quarter performance. To meet customer requirements, our organizational responsiveness and manufacturing flexibility enabled us to deliver high volumes of the new Power-on-Package modules. Because we had not manufactured meaningful volumes of these products prior to Q2, the learning curve was steep and, in a word, costly. Some of these expenses were included of cost of goods lowering the consolidated gross profit margin, but the majority of these learning curve costs were recorded as operating expenses in research and development. While we experienced a similar increase in prototyping expense for the first quarter, as we prepared for the second quarter deliveries, we believe we have completed our learning curve for these products. I cannot use terms such as nonrecurring in association with these expenses, but management believes the likelihood of another quarter of such extensive early stage prototyping expenses is remote. We have fulfilled commitments in support to the first application of our first Power-on-Package customer and believe we are in a position to sustain attractive gross profit margins with high-volume production of these new products. Consolidated operating expenses as a whole sequentially rose 5.7%, but, as just mentioned, the increase in total spending was largely due to an increase in prototype spending within R&D. Marketing and sales expenses rose with the seasonal increase in travel and G&A expenses were steady quarter-to-quarter. Nonoperating income consisted of lease income from of our California facility as well as favorable foreign currency exchange rates. Our consolidated income tax provision for the quarter reflected no unusual or nonrecurring activity, but it did reflect higher taxes for both state income and foreign income. Turning to cash flow and our cash position. Operations generated cash flow of $644,000, up from the prior quarter's deficit of $1.3 million. After capital expenditures of $3 million and proceeds from stock option exercise of $1 million, quarter end cash and equivalents balance declined $1.4 million to $51.4 million. As we have stated before, we believe we have the liquidity to fund anticipated growth, including the expansion of net working capital and further capital expenditures for expanding manufacturing capabilities. The quality of our receivables portfolio was excellent with second quarter days sales rising slightly to 45 days from the first quarter's 44 days. Our aging schedule is also in excellent shape with no meaningful past dues. Annualized turnover of consolidated inventories continues at the desired level, but did decline slightly from 4.2x to 4.0x reflecting, as has been the recent trend, increases in raw materials helped by VI Chip and Picor in anticipation of scheduled demand. No meaningful activity occurred in our inventory reserve accounts. Employee headcount at quarter end was 988, of which 964 were full-time employees. Total headcount declined by 10 quarter-to-quarter, despite implementing weekend manufacturing shifts for the last 6 weeks of the second quarter to support production volumes. Turning to new orders. Second quarter bookings increased 2.6% to $59.4 million from the prior quarters $57.9 million, and 1-year backlog rose 2.8% quarter-to-quarter to $53.2 million. New order activity for the quarter was characterized by encouraging increases in bookings from BBU and Vicor Custom customers, representing approximately $6 million of new orders. Picor's bookings rose reflecting demand from contract manufacturers for the SiP PRM element of our VR 13.0 solution. This combined robust activity was offset partially by a $4.5 million sequential decline in VI Chip orders. Recall that Power-on-Package products boosted first quarter's booking, whereas Q2 bookings of motherboard direct to processor solutions were depressed by contract manufacturers coping with the extended lead times for DDR4 memory components. These supply chain issues are being resolved and we expect new orders this quarter for these products from our customers. Also, we continue to book orders for our VR 13.0 data center solution as well as new ChiP and SiP-based solutions in aerospace, computing, electric vehicles and autonomous vehicles, reflecting our continued focus on diversifying our customer base. Turning to our outlook. We anticipate net profitability in Q3, and a sequential increase in bookings. I must remind listeners, as I regularly do, of the difficulty of accurately forecasting sales cycles for disruptive innovative technologies. Similarly, the decisions made by early adopting customers and their contract manufacturers and the timing of those decisions are subject to change brought about by many factors out of our control. And the consequence being sudden and anticipated changes in operating and financial forecasts. So, with that, I'll turn the call over to Patrizio.

Thank you, Jamie. As Jamie noted, our second quarter was characterized by accomplishments across the organization from engineering to manufacturing. I'm pleased with how we responded to the challenges, especially with regard to volume demands in a quarter in which we started using Power-on-Package modules. Notably, VI Chip recorded its highest monthly gross margin percentage, which exceeded 50% in June. This is encouraging, given the prospect of increasing demand over the coming quarters. Our VR 13.0 business in Q2 was negatively impacted by DDR4 memory shortages, causing customers to have to delay planned infrastructure buildup, but we're receiving orders from expanding range of customer for second half of 2017 and first half 2018 deliveries. As mentioned, we are also receiving orders for other advanced products from customers across a range of industries and applications, as we pursue expansion of our customer base. During our last two earnings calls as well as during my remarks at the recent Annual Meeting of Shareholders, I referred to our proprietary Power-on-Package technology for powering CPUs, GPUs and ASICs, in short XPUs, using high-performance applications. We already have three substantial engagements for these high-power density modules. As Jamie indicated, we've learned to manufacture them efficiently at scale. Just as with our 48-volt direct-to-processor solution, we established new performance benchmarks for powering servers in large-scale data centers. With our Power-on-Package solution, we're setting a high bar for powering our performance applications. Vicor's 48-volt direct-to-processor modular system implemented with our PRM and VTM, allow server designers to locate the current multiplier VTM stage next to the processor, while allowing the PRM regression stage to be located further away from the processor. This Factorized Power approach has delivered superior system efficiency, power density and to our cost of ownership. However, ever-increasing processing requirements placed on XPUs by demanding applications such as artificial intelligence, machine learning and big data mining are driving XPU operating voltages well below 1-volt, while operating currents are escalating to several hundred amperes, further contributing to power delivery challenges and power distribution losses. To minimize power distribution losses, power to the XPU should be delivered within the XPU package, at a voltage, at least -- in order of magnitude, greater than 1 volt. This has been an elusive goal, because of the high-power conversion density required to fit power stages providing high current multiplication within XPU packages. However, within the last six months, Vicor's power conversion and power stage partitioning technology has enabled early adopters to co-package 64x or a factor of 64 modular current multipliers, so-called MCMs, within XPU packages, essentially wiping out power distribution interconnect losses. Our Power-on-Package solution has accomplished this elusive goal by eliminating what we refer to as the last inch, the distance between conventional lower per-voltage power stages outside of the XPU package and the XPU itself. For high-current applications, this last inch is a major source of voltage perturbations and power delivery loss. Specifically, with high-current XPU, and last inch of the systems is comprising motherboard phrases and socket interconnect pins. The XPU may suffer a double-digit percentage power delivery loss in the last inch alone and large transient voltage perturbations that stand in the way of XPU performance. By copackaging MCMs within XPUs, Vicor's Power-on-Package solution eliminates transient voltage perturbations that affect the XPUs and power loss in the last inch, while importantly, freeing up XPU interconnect pins that are traditionally consumed by inefficient delivery of the low voltage. Reclaiming XPU pins to expand the XPU connectivity is in and of itself a high valuable proposition for XPUs. In summary, our Power-on-Package, corporately referred to as PoP, solution is based upon Vicor's Factorized Power Architecture, which has enabled 48-volt direct-to-CPU solution. The PoP system introduces two new building blocks, so-called modular current multipliers or MCMs and the MCM driver, so called MCD. The MCD is powered from a 48-volt bus and drives MCMs with a factorized 48-volt bus. The MCD is placed on the motherboard and the MCMs, which are very small and fast proprietary current multipliers, are mounted directly onto the XPU substrate. Because MCMs will now reside within the XPU substrate, high XPU currents do not have to traverse the last inch, and power from the MCD to the MCM can be efficiently routed to the MCM of the factorized 48-volt bus, with only a few XPU package pins on the powerpack. Early adopters of our Power-on-Package modular solution seek to leverage it with advanced CPUs and GPUs for supercomputing and artificial intelligence applications. I'm sure listeners have many questions, so I will open the call. Operator?

[Operator instructions] Your first question comes from the line of [Jim Bartlett]. Your line is now open.

Following on that, PV, you had said that you had three significant engagements. Can you give us some idea of when we would start seeing that in terms of bookings revenues? I believe one of them is a supercomputer engagement which you had mentioned before and I believe the -- you already have some, but ramping up 4x or something like that in the third quarter. But could you give us some other ideas of when all three of these and maybe more start having a significant impact?

So, to your point, the first one has originated a couple of million dollars of revenues in Q2. We had bookings for about 4x, that amount to be shipped in the third, fourth quarter, potentially, first quarter of 2018. The other two customers are still in the product validation phase. And I do not expect meaningful revenues until sometime in 2018 and the time frame for that is still somewhat cloudy. But we do expect, as 2018 progresses, to see at least two additional customers turning in volume to the MCM, MCD system with more advanced power delivery and overall system efficiency advances. There is initial interest from other customers that are not as far along as the first three. We do see this as being a high leveraged capability that delivers great competitive advantages to solutions that can benefit from Power-on-Package power delivery.

Also, if -- could you give us an idea of the -- the R&D ramp up of $900,000, which you said was related to the startup cost, is R&D now back off and go back down to approximately the level that it was before? Or slightly near that?

Yes. As you may imagine, because of the rapid ramp with products that were still in the preproduction phase, we had inefficiencies, and we had to do what it takes to satisfy the customer requirements in a very short time frame. That was not an efficient way of operating. It took quite an effort on the part of engineering as well as our operations infrastructure. And the priority was to address the customer need more than maximizing margins for the short term. And that's an issue that we believe is behind us. So, we learned what we needed to learn. We manufactured over 40,000 of these modules and are looking forward to making a lot more in the future in a more orderly and cost-effective fashion.

Also, could you address capacity a little bit? You mentioned going to -- number of employees were down, but you are going to weekend schedules, but you're -- what is your capacity? Is it both in the BBU segment and in the -- well, in the chip and in your chips that you're producing? What kind of capacity levels do you - are you setting it sort of right now and then over the next 6 months and with newer equipments needed? And then thinking beyond that, when you would anticipate another plant?

So, some of the weekend shifts, because a fraction of them were again due to sorting things that need to be done on a short-time bases to address this first Power-on-Package customer requirements. With respect to capacity, we believe we are well positioned with capital equipment investments that have been made. I think, as Jamie noted, just within Q2, we made $3 million additional investments in capital equipment. There is more of that coming to further expand capacity. We stand by earlier projections with respect to overall capacity of the facility in Andover. I think as noted during the Annual Shareholders' Meeting, we're also actively looking for incremental space nearby, not far away from the Andover facility, to further expand capacity in about one-year time frame.

Your next question comes from the line of John Dillon.

Patrizio, I'm just wondering if you've got any of the IBM Power9 booking charters. Are those going to be coming in the next Q or 2?

So, we are told that that program should go into production early in 2018. We're about to get a first order of the order of sub-$1 million for the last dry-down as it were prior to production, which should begin in early 2018. To some extent it's a function of availability of the P9 processors themselves. That may be a gating element with respect to this.

Right. So, will the bookings come before the first quarter of 2018 for those?

Yes, yes. Yes, so the lead time would typically be for something like this pretty close to 3 months.

Okay. And do you see substantial revenue from the Power9 coming?

Yes. We do.

Okay, excellent. And do you have any supercomputer wins with -- along with the Power9? Or is it more a servers-type thing?

I think at this point in time, it is servers. We do have supercomputer wins with other kinds of processors, other primary processors or Intel processors. But as far as I know, no Power9 yet.

Okay, so they're most for the data center-type server applications?

Yes.

Okay. In the Annual Meeting, I'm not sure but I thought you heard that your -- you actually have some 12-volt wins. Did I hear correctly?

We have -- well. So, I think we may need to get a little bit more specific because 12-volt amp is in a variety of ways within our product capability.

Right, I'm talking more about the new stuff, the new ChiPs and SiPs.

Okay. So, we do have -- in fact, we just recently received a first production order for delivering 12-volt outputs from high-voltage bus. That's for computing application, and that should scale up for next year. We're involved in providing 12-volt, also for 48-volt buses for a variety of applications. So, there's a lot of 12-volt activity. As you may know, 12-volt is still the primary bus. It is becoming a legacy bus, but it is not going to die away overnight. So, for quite some time, there's going to be a 12-volt infrastructure. We view that as an opportunity, because we have some very good 12-volt products, more in terms of delivering 12 volts from a variety of power sources than taking the 12 volt to the point of load. We don't bother with that, because frankly, that's a commodity play that over time will fade way. But while there's still an opportunity at 12 volt, we can efficiently address it with modules that deliver 12 volt for further processing to the point of load by the commodity products. So, in that respect, we have some very exciting products in our portfolio that we used more and more by a variety of customers in the server space, in computing and other spaces. And we're coming up actually with some additional products that will further raise the bar with respect to delivering 12 volt, even though again, a 12-volt bus is handicapped by its inherent limitations and we don't see that as a long-term play.

Right. I'm with you. Got you. And, again, that's more -- it sounds like more the front-end type stuff with the 12 volts.

Yes, it's not front-end, both front-end in the sense of addressing solutions at the point of load that go through a 12-volt intermediate bus, starting at 48 volt, to facilitate conversion to 48 volt. That conversion is generally beneficial to our capabilities and that we have said customers are still asking us, in some instances, to make the jump all the way to Factorized Power solution, if they want to take an interim step of taking advantage of the 48-volt infrastructure while still processing power to the point of load by use of traditional legacy, commodity POL regulators, so be it, we're happy to capture the -- what you call the front-end piece of that by delivering the 12 volt. Again, with this protection, the longer term, that part of the business will convert to Factorized Power solution. Either with PRMs or VTMs, or in more advanced applications with the MCD, MCM system that we recently enabled.

Got you. Okay. And before, you talked about a $75 million -- I think it was a bookings for the fourth quarter. Is it still on track?

We see that coming, let's wait and see when we get a little closer to it. We're still making plans for a step up, an official step up in the third quarter that appears to be within reach. We'll update our forecast as we get towards the end of third quarter, we have more visibility to the fourth quarter. But as far the eye can see, we see significant upticks in the bookings and in the top line as the next several quarters progress.

Okay. So, you do see a significant uptick for both bookings and revenue for the next quarter?

For the next few quarters, for the next several quarters. We see more and more of the seeds we've been planting over, in some cases, a few years, or, in other cases, more with more recent developments. We see those seeds coming to fruition. In some instances, it's taken a long time, but they're finally coming to fruition. As an example, the high-voltage to 12-volt modules for which we recently got a significant -- first significant order, that opportunity has been percolating for literally two or three years. And that being set back a few times in terms of the production ramps slipping. But they're finally coming to fruition. So, we see more of that happening; and in some instances, things are moving faster, particularly with our more advanced products.

Your next question comes from the line of [Don McKenna].

Patrizio, I want to follow up on that last question of John's too because that was the one that I had. I was a little bit surprised with the second quarter only being the 2.6% increase in the bookings. When you say significant increases for the quarters going forward, what kind of magnitude relative to that 2.6%? Are we talking something in the 5% to 10% increases per quarter?

We're talking about a larger percentage. As mentioned in the introductory remarks, Q2 bookings were somewhat depressed by availability of DDR4 memory.

Okay. And going back to your startup costs of ramping up the new line and everything else. If you were operating at a normalized level, would the quarter -- as far as costs go, would the quarter have been profitable?

Yes, it would. We really had to throw a lot of resources at making it happen for our first MCM-MCD customer in the past quarter. That was quite a challenge for the organization. And frankly, we did what it took to make it happen and we did it. But as Jamie mentioned, it was a costly proposition.

Yes. Very understandable. And, Jamie, you had mentioned that future capacity add-ons could be financed through cash flow, internal cash flow. Does that include the possibility of the new facility?

Yes, it does. We do not anticipate needing to go outside for certainly, a year to 18 months before I have to really start paying attention to the financing of anything significant.

As an order of magnitude, the outside facility with equipment would be in the ballpark of $10 million, $15 million. So not a huge investment.

Okay. Great. Now -- but you had mentioned too at the annual meeting that for a licensee to outfit their operation would be quite expensive. Is that, that same $10 million to $15 million that you just referenced?

Well, so the $10 million to $15 million would not be for a complete buildup of a complete capability soup to nuts. So, it would be for a initial baseline capability that would vertically integrate the process steps that are currently being performed outside of our facilities at significant cost. So, we are aiming to became totally vertically integrated with respect to all the process steps involved in chip making as we see demand requiring us to have that profitability.

Your next question comes from the line of Hasnain Karim.

Just a couple of questions. Now that Intel's new Xeon chips are generally available or will be at least in the next couple of months, could you talk about, I guess, how order trends will progress over that generation and the next 18 months? If you're seeing more customers start to fiddle with the -- your VR 13.0 VTM products, because those processors are generally available? And then finally, you'd mentioned a front-end order a couple of quarters ago and I was wondering if that's started to ship already. Or is that something that's going to come about in the next quarter or 2?

Okay, so on the VR 13.0 front, we see an expanding range of opportunities with some exciting new customers. In one instance in particular, the opportunities by way of an Asian supply of the second of building blocks and develop a complete solution, which is being shopped to various customers. But whether they are clearing here, actually we see an expanding range of opportunities and that's, obviously, particularly true for Intel sockets, which are still, as to say, dominant in the marketplace. But what we see also are great opportunities is in applications that don't involve intra-processor work. There is significant power demands that go in the direction of sub 1-volt, in some instances, 600, 700, 800 amp peak card rows with GPUs, ASICs that are outside of the Intel sea of dominance. It's for these applications for which I think there is a growing opportunity, where the leverage of our technology is greatest because the benefits of a current multiplier at the point of load, particularly one within the XPU package or if not that, next to the XPU, they are significantly more beneficial if the voltage is delivered is 0.7, 0.8 volt as compared to the Intel standard voltage which is around 1.8 volt. In other words, for the same power delivery, tall power delivery a 0.8 volt for the current is more than double. And in some instances, for other reasons, the current is 3x, maybe 4x as much at the peak. And it's these very current-hungry and current-challenging applications that represent the greater opportunity going forward. That's not to say we're not going to continue to pursue Intel sockets. We are developing a MCD-MCM system, that's going to be extremely cost-effective and very high performance for Intel sockets, recognizing that that's still going to be a dominant portion of the marketplace. But we're not overlooking the -- what is going on in the rest of the industry that is aiming to compete in one form or another with what Intel has traditionally provided. Now going to the second part of your question regarding front ends, we're doing more and more with front ends. I think Jamie mentioned that as part of our package solutions, we also manufacture and deliver to the customer front-end solutions. So that's one order that we have received and we shipped in Q2. And I may have mentioned this before, I may be repeating myself, but it's worth noting, this is a case in which we provide a solution, you might call it soup to nuts or for 3-phase AC line all the way to a 0.8-volt CPU through the various stages including if you phase front-end, followed by MCDs, driving MCMs within the CPU, the proprietary CPU package. So that in and of itself, I think is quite an accomplishment. It's -- it underscores the leverage of having a modular building block and power capability that spans the gamut from front ends, i.e., the source, all the way to the point of load. It's that tall power capability that in many ways sets us even further apart from the rest of the industry.

I guess that sounds like a great opportunity. But it also seems to be one that requires a lot more effort because you have to win at every stage of design on the board. But let's -- I guess let me follow up with another question.

If I can interrupt you for a moment just to touch on that point. I would view it differently. I think the people in sales and marketing (inaudible) Vicor view it differently. It's not that we have to sell the entire solution to penetrate the account. More often than not, we can get into the account and the particular application with a portion of the system. This is not an all or nothing proposition. It's -- but it's a proposition where if the customer sees the benefit of using our point of load solutions, we may well serve the customer further by complementing that with state-of-the-art extremely high-powered density, efficiency front end and vice versa. So, we have customers that feel more comfortable getting involved with the Vicor power system at the front end, and then want to extend it to the point of load. So, it's not an impediment to sales, because we're happy to do business wherever the customer takes us or leaves us. But I think it's important to project to the customer, it's our capability that can provide even greater benefits in terms of the efficiency, density and cost-effectiveness of the complete solution. And please go ahead if I interrupted.

Yes, okay. Not a problem. I guess if you were to look out 18 months, would the majority of your business still be tied into -- with regards to 48 volt, would that be tied into Intel processors or non-Intel processors? Looking out 18 months.

Well, I think we have some very exciting applications that don't involve Intel processors. But I think within 18 months, Intel sockets for point of load power delivery business opportunities for Vicor will still be the likely majority. But if we look further out, I think what is going on in terms of a variety of CPU, GPU, ASIC initiatives to enable again artificial intelligence, capabilities, autonomous driving, you name it, the path could shift shipped to where a significant component comes from sockets other than Intel.

Okay. And then finally for me, I think there's been a lot of questions regarding capacity with Vicor. And really, I mean, from the answers that you've given, the issue really comes about when you're -- at least from a revenue basis, when you're over $400 million, $450 million revenues. Are a lot of these questions being driven by customers or investors more?

Well, customers do ask questions about capacity, that's natural. And we're able to address them. I think generally speaking, customers that come to our facility, go away with a great deal of confidence in our capabilities. They formally remark we have state-of-the-art facilities. These are customers that are very savvy, they typically visited Asian manufacturing sites with a variety of capabilities. And almost universally, they comment about the level of capability, proficiency and expertise that we have in our manufacturing processes, in our packages, in our manufacturing infrastructure. They, having seen our facilities, get persuaded of the scalability of our capacity, because the capacity itself is built out of modules that can be copied and replicated and scaled up. Obviously, space is going to become an issue, that's why we're beginning to take some steps to install capacity at places other than our Federal Street facility. But you referenced $400 million, $450 million, we think we can do more than that out of Federal Street. But we're not going to be doing all of it out of Federal Street before we start doing some of it elsewhere. And I think that will also give some customers more comfort with respect to capacity. We've been able to address it, and we obviously have our customers that do tens of millions of dollars per year worth of business with us as a single-source supplier. They're very comfortable with us, because they've built up that confidence over many years. They have seen us perform and deal with their sometimes rapidly changing demands. We're very adaptive. We're very flexible, and we're very careful.

You have a follow-up question from John Dillon.

Patrizio, did you say you had a P&P design for an Intel application?

What are you referring to?

A P&P?

Yes, power and package. You threw around a lot of acronyms and (inaudible)...

The Power-on-Package applications as of now, as of now, they are all non-Intel.

Right, did you say you were going to get one though for an Intel or...

I didn't say that. What I said, and I'm measuring my words carefully, is that as of now, all of the Power-on-Package applications are for non-Intel sockets.

Okay, sorry, I misunderstood. And you talked in the pre-remarks about the DDR4 shortage and also, it sounds like there's some inventory buildup at some of your ODMs. Have those 2 issues been cleared now?

Well, so, I wouldn't say they're clear yet. But I'm told that as you have been seeing in the industry, particularly the semiconductor industry, capacity demands and shortages get addressed. It's an unfortunate trait of this industry that it progresses from famine to feast and then it does it all over again with cycle said in one way of looking at it, it's somewhat harder to forecast and somewhat irrational. So, I'm told that there is a lot of capacity coming, and what's inevitably going to happen when the capacity gets delivered is that we will turn it to our capacity and the vendors will be dealing with a different set of issues. But there appears to still be capacity constraint on that front.

Mostly on the DDR4 then?

Yes.

And at least, current issue.

Yes, current issue. Okay. But do you still see your bookings and revenue picking up for this quarter?

Yes.

Good. Okay. And you mentioned, I think, 1 year for a new facility. Is that breaking ground? Or is that -- what does that mean 1 year?

So, we're looking -- we had the review earlier today. We're looking at various options and assessing the relative benefits of either purchasing, converting an existing facility or going at it from the ground up have not come to a conclusion yet. And it will probably still take a little bit of time to make a decision on this front.

Right. so, you have to make a decision then I would imagine to build something up would take another year after that one and there's a...

Yes, so I think realistically, we get to look at capacity installed towards the other part of 2018.

Your last question comes from the line of Alan Hicks.

I had a question on -- did I hear that BBU orders rose by $6 million last quarter?

They did, yes.

And what's driving that?

Well, so BBU now encompasses our front-end capabilities and so, as earlier discussed, it's not just the good old Bricks, the legacy products, which still sell quite well. And they're keeping up their act and they fluctuate down or up a little bit. They seem to be immortal in some respects, considering how long they've been around. But because of the fact that they've had such a good track record of reliability and adoption in certain pockets of the industry, they're still doing well. They're still holding their own. But we're now layering on to that new customers, new applications, generally speaking, the new applications, the new customers tend to be higher-volume customers, much more cost-centric customers. But leveraging up into the power system world into the front ends in particular, our ChiP technology, various engines to perform at that front-end power conversion, we're enabling new kinds of front-end applications that should add significant growth to the BBU parts of the business. So, we're looking at ...

Are the VS part of that?

The VS are part of that so we just got a significant part there for VS. There's one customer, just one customer that next year should do around $4 million in VS. So again, it takes time to plant the seeds and see them come to fruition. But it is happening. Also, our DCMs, the whole DCM product line is generating, those are Brick products, but they're based on ChiP technology. There are the in-VIA packages or they are chips that mount exactly on the customer board. Those products are doing quite well. We got record registrations of new projects and subject to the usual cycle times and lead times of the industry which are unfortunately quite long, those two are going to result, in time, in additional revenues for the BBU. So, it's going to be a race in terms of revenue contribution and growth over time between a front-end business, a point-to-load business. They each have their own set of opportunities. And again, some of these, as earlier commented, are very complementary. We're happy to get in one way or the other or with both.

So, some of your new product and packaging is planting the older legacy Bricks, is that what's happening?

It is. But the word cannibalization comes to mind. But it isn't really that, because, generally speaking, we're dealing with different customers and different applications. The traditional customer for a legacy Vicor Brick has got the following traits. It's very low volume, not price-sensitive, very excited on certain classic aspects of a Brick, higher reliability, typically different types of applications; whereas a new BBU customer is a customer that is using products in much higher volumes, much more cost sensitive. So different type of customer, different kinds of application. And because of that, they coexist in somewhat separate compartments.

Okay. So, you think you can begin to stabilize and grow the BBU revenues going forward?

Yes. We are very much focused on growing our system business to complement the point of load stuff. And as I've said now a few times, the race is on. And I think it could -- the winner in terms of relative contribution to total revenues could be either. There is plenty opportunities for both.

Okay. But that $6 million won't all come in one quarter, it'll be spaced out over several quarters?

Yes. I don't know what the schedule of those bookings are, but...

Alan, the thing to remember about the two businesses that Patrizio's addressed is that quite frequently, the size of the orders, the individual orders for the board level business, can be quite large. And if -- when you look at a $6 million increase sequentially for the BBU, that's an awful lot of individual orders. It speaks to the -- one of the concerns we've had is some of the older industries, the more mature industries that the BBU serves haven't enjoyed the same robust recovery that we've seen in other areas. And there's really a very strong statement that, Patrizio, what was the word you used, the immortal, or whatever it was.

Yes, the immortality of Bricks.

It's just it's a -- it was a very nice thing to see.

And I had a question on the VI Chips, I think you said fell by $4.5 million in orders?

Well again, that's a reflection of the lumpiness of the orders, is that when a large OEM comes in or it's a contract manufacturer, they'll place orders that will go out for quarters.

So, in the prior quarter, we had had substantial orders for MCMs, MCDs. Those did not recur this past quarter. And with the team's solutions, while we did receive some very substantial orders, we fell short of the forecast because of the DDR4 memory issue.

Okay. So, is that picking up during July?

I'm sorry?

Are orders picking up in July?

Orders during July were ahead of the prior quarter in terms of daily booking rates. And the forecast is positive.

So, you think you can surpass the second quarter VI Chip revenues?

Yes. Yes, yes.

Okay, so that's good. Then I had one more question. What do you think the gross margins are going to be on these Power-on-Package products once you really get them in volume?

I think they're going to be very good, because they have a very good cost structure. The MCM in particular is a very cost-effective device. It's very scalable, very minimalistic. It's a very thin package; it's only 2.7 millimeter thick or thin. As compared, for instance, to other chips, it's 1/3 the height, that's obviously a key attribute in order to enable it to fit under the lid of an XPU package. So, we had to develop that for the purpose of fitting it in. But in the process of enabling that capability, we also enabled the significant cost reduction, because remarkably, a smaller product is a lower-cost product. And so that's giving the MCM-MCD chipset a cost structure that is more attractive than PRMs and VTMs.

So well over 50%?

Well, we expect PRMs and VDMs to get well past 50% in margins. With MCMs and MCDs we have a similar opportunity. Our pricing strategy will be to pass along to customers, because these actions we are achieving with these products to drive volume and growth. So, I wouldn't handicap one versus the other. I think generally speaking, customers with MCMs and MCDs are going to enjoy both higher-performance solution and one is more cost-effective measured in terms per watt ancestral land. And that's going to drive faster and greater acceptance sooner.

Okay, one more question. Are you going to need to raise money to put in a new plant?

No. I think we addressed that earlier. So, in the short term, we're only looking at $10 million to $15 million.

There are no further questions in the queue.

Well, thank you, and we'll be talking to you in 3 months.