Vicor Corp (VICR) 2022 Q1 法說會逐字稿

Good day, and welcome, everyone, to the Vicor earnings results for the first quarter ended March 31, 2022, posted by James Schmidt, Chief Financial Officer. My name is Matt, and I'm the event manager.

(Operator Instructions) I would like to advise all parties that this conference is being recorded. And now I'd like to hand it over to James. You may proceed.

Thank you. Good afternoon, and welcome to Vicor Corporation's earnings call for the first quarter ended March 31, 2022. I'm Jim Schmidt, Chief Financial Officer; and I am in Andover with Patrizio Vinciarelli, Chief Executive Officer; and Phil Davies, Vice President of Global Sales and Marketing.

After the markets closed today, we issued a press release summarizing our financial results for the 3 months ending March 31. This press release has been posted on the Investor Relations page of our website, www.vicorpower.com. We also filed a Form 8-K today relating to the issuance of this press release.

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 statements regarding current and planned products, current and potential customers, potential market opportunities, expected events and announcements, and our capacity expansion as well as management's expectations for 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 2021 Form 10-K, which we filed with the SEC on March 1, 2022. The document is available via the EDGAR system on the SEC's website.

Please note the information provided during this conference call is accurate only as of today, Thursday, April 21, 2022. 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 May 6, 2022. The replay dial-in number is (888) 286-8010, followed by the passcode 19042-467. This dial-in and passcode are also set forth in today's press release. In addition, a webcast replay of today's call, along with a transcript will be available shortly on the Investor Relations page of our website.

I'll now turn to a review of our Q1 financial performance, after which Phil will review recent market developments and Patrizio, Phil and I will take your questions. In my remarks, I will focus mostly on the sequential quarterly change for P&L and balance sheet items and refer you to our press release or our upcoming Form 10-Q for year-over-year comparisons.

As stated in today's press release, Vicor recorded total revenue for the first quarter of $88.3 million, a 2.2% sequential decrease from $90.3 million in the fourth quarter of 2021. Advanced Products revenue increased 3% sequentially, while Brick Products revenue declined 9% from the prior quarter. Advanced Products revenue increased 54% from the same quarter a year ago. Shipments to stocking distributors increased 3.3% sequentially and decreased 1.2% year-over-year. Exports for the first quarter were relatively flat sequentially as a percentage of total revenue at approximately 72% from the prior quarter's 71.7%.

For Q1, Advanced Products share of total revenue increased to 59.9% compared to 56.9% in the fourth quarter of 2021, with Brick Products share correspondingly decreasing to 40.1% of revenue.

Turning to Q1 gross margin. We recorded a consolidated gross profit margin of 42.6%. Gross margin declined sequentially from 45.2% in the fourth quarter of 2021, primarily as a result of lower volume, a full quarter of higher cost of outsourced panel production, an increase in manufacturing spending associated with higher freight costs incurred to respond to supply chain disruptions as well as increased staffing and activity levels to support the start-up of our manufacturing expansion. In addition, tariffs continue to be a drag on gross margin at $2.2 million in Q1 and 2.5% of revenue. Our work to reduce tariffs by reducing imports from China continues.

I'll now turn to Q1 operating expenses. Total operating expense increased 2.8% from the fourth quarter of 2021, driven by increased compensation, engineering activity and professional services. The amounts of total equity-based compensation expense for Q1 included in cost of goods, SG&A and R&D was $251,000 $1,207,000 and $536,000, respectively, totaling approximately $2 million. For Q1, we recorded operating income of $4.8 million, representing an operating margin of 5.4%.

Income taxes for Q1 were a tax benefit of $48,000. Net income for the quarter totaled $5 million. GAAP diluted earnings per share was $0.11 based on a fully diluted share count of 44, 954,000 shares.

Before I review our financial position, just a brief update about COVID-19 and our workforce. As previously discussed, as a designated essential manufacturer, using masks and practicing social distancing from the onset of the pandemic, we have continuously operated 3 shifts at our Andover manufacturing facility. Cases in absenteeism due to COVID-19 are now negligible. Nevertheless, because much of the potential influence of the COVID-19 pandemic are associated with risks outside of our control, we cannot estimate the extent of such influence on our financial or operational performance or when such influence might occur. In particular, the zero-COVID policy adopted by China has caused disruptions in parts of our supply chain and the impact and timing of the effect on our results are unpredictable.

Turning to our cash flow and balance sheet. Cash, cash equivalents and short-term investments totaled $211 million at the end of Q1. Accounts receivable net of reserves totaled $52.7 million at quarter end, with DSOs for trade receivables at 37 days. All balances are current.

Inventories net of reserves increased 9.7% sequentially to $73.9 million and with annualized turns at 2.65. Operating cash flow totaled $4.6 million for the quarter. Capital expenditures for Q1 totaled $22.7 million. We ended the quarter with a total construction and progress balance of $50 million and approximately $42 million scheduled to be spent through the year, primarily for manufacturing equipment. Our factory expansion is proceeding on schedule and on budget.

I'll now address bookings and backlog. Q1 book-to-bill came in well above 1 and with 1-year backlog increasing sequentially by 22.6% from the fourth quarter of 2021.

Turning to the second quarter of 2022 and beyond. We expect to bring our integrated ChiP fabrication manufacturing expansion online starting in Q3. ChiP is our acronym for Converter housed in Package. Our automated chip foundry will establish the in-house process flow to assemble, mold, plate, test and finish, complete high-density power modules in high volume and at world-class quality levels. Until Vicor's chip foundry is online, existing capacity will limit our ability to increase our production and shipment rate in the short term.

During this quarter, we are working on installation of equipment in our chip foundry. Once installed, the production line will undergo full qualification and approving in period. As the line becomes fully operational in the coming months, we will be positioned to significantly increase capacity, efficiency and output, which we expect will result in improving operating results in the second half of the year.

Also during this quarter, we are finalizing and expect to complete our annual merit process. Our focus remains on our long-term success based on our pioneering approach to integrated high-density power module technology. We are working hard to combine world-class operating performance with the technical advantages of the Vicor power delivery products.

We've invested in a unique state-of-the-art U.S.-based automated manufacturing facility, which is in close proximity to our development engineering teams and allowance for co-development of next-generation products.

Our manufacturing strategy also de-risks supply chains, lays the foundation for cost efficiency and cost reduction and positions us to deliver the output needed in the highest volume, most demanding power delivery applications. And we are embarking on an initiative to drive operational excellence across Vicor.

At the heart of that initiative is our aim to become a customer-centric company. Our entire organization is engaged to make Vicor the power delivery supplier of choice for customers in a variety of end markets. We are committed to achieving our goal of operational excellence.

With that, Phil will provide an overview of recent market developments, and then Patrizio, Phil and I will take your questions. I ask that you limit yourselves to one question and a related follow-up, so that we can respond to as many of you as we can in the limited time available. If you have more than one topic to address, please get back in the queue. Phil?

Thank you, Jim. I'll keep my comments short and focused them primarily on our high-performance computing business, which has received a lot of attention from our investor community.

Our backlog growth is driven by our global high-performance computing customers and new supercomputer cloud server and AI accelerated card solutions being brought to market. The 48-volt bridging applications are now ramping with major hyperscalers as they deploy rack-based clustered GPU AI systems into their data centers in rapidly increasing numbers. Additionally, GPU and ASIC based 48-volt accelerated cards, which have varying power requirements are ramping with both 48-volt to 12-volt bridging solutions for lower power cards as well as direct 48-volt to point of load factorized power solutions for higher-powered GPUs and ASICs.

The technology gap separating Vicor from a multiplicity of aspiring competitors is expanding, not contracting. Specifically, current density and other key performance attributes of Vicor power system solutions are improving at a rate faster than achievable with intermediate bus architecture multiphase solutions. High-performance computing applications require a level of performance that only Vicor modules are capable of providing. Very soon, they will require Vicor proprietary power system architectures such as vertical and lateral vertical power delivery. And that is why our data center business opportunity is broadening and going up, not down.

To be clear, the high-performance computing industry is no different than any other high-volume industry seeking multiple sources to de-risk supply chains. However, wishing to accomplish a multiplicity of sources does not necessarily mean being able to do it with good enough performance. We remain confident that our performance advantages will continue to expand our opportunities and design wins with major customers.

With a few current multiplayer modules deployed laterally or preferably vertically, a Vicor power system can achieve higher density, lower noise, higher efficiency and higher performance and can be achieved with a multiphase solution consisting of over 100 large noisy and hot components. As discussed in previous quarterly calls, the demand for high-density power delivery solutions continues unabated across all of our target markets with power systems engineers turning to Vicor's modular power solutions to solve their toughest design challenges.

As our new vertically integrated capacity comes online in Q3, we will be better positioned to take full advantage of the increased number of high-growth opportunities with our current and new high-performance computing customers who are also eagerly anticipating this key event for Vicor.

As we have discussed in previous quarters, Vicor is gearing up for sustained and predictable long-term growth. A major part of our growth strategy will be success in the automotive powertrain market and having a vertically -- a full vertically integrated manufacturing process will also enable us to meet the exacting qualification standards demanded by our automotive customers. Our progress in developing this exciting new market and engagements with leading OEMs continues to be ahead of our initial plans.

That concludes my remarks, and Patrizio, Jim and I will now take your questions.

(Operator Instructions) And we already received our first question, and it is coming from the line of Quinn Bolton from Needham & Company, LLC.

One of your large customers, NVIDIA recently showed its next-generation H100 card powered by a multiphase voltage regulator architecture. Yet this card has power consumption of 700 watts. I'm wondering if you can talk to this, is this evidence that multiphase is viable above 400 watts and 1,000 amps of current? Or will Vicor participate in this next-generation architecture at that customer?

So as a matter of general policy, we will not comment on specific customers or specific applications. But in general, addressing your question, which is not surprising, I think, Phil, made some key points that are relevant to addressing your question.

First of all, the gap that separates Vicor technology from the intermediate bus architecture and multiphase solutions in the context of a 48-volt architecture, which would intermediate solution requires 2 stages of power processing. Our technological gap is advancing and advancing at a rapid pace that the intermediate bus architecture cannot keep up with.

To the point of your question, the intermediate bus architecture and multiphase continues to make advances. And certainly, in the last 3 years, they made some sides with respect to increasing the current density and capability of that solution. But as you know from past comments, it is a solution. It is an architecture and building blocks that are fundamentally handicapped in terms of what can be achieved. And that fundamentally adapts with industry trends that your point require for advanced applications, increasing current levels, increasing account density. By the way, not just in terms of primary well, but those in terms of auxiliary wells. And all these, we believe continues to support the choice of Vicor solutions.

Patrizio, maybe just a quick follow-up on that without speaking to a specific customer. If you look at an architecture that's multiple hundreds of watts of power consumption in, say, 1,000 amps, can you give us some sense how much more efficient or what kind of advantage in terms of power efficiency or power savings do you think the Factorized Power Architecture has over today's multiphase bus architecture approach?

So I'll give you a general example, or I should say an example that would generally apply to GPU or high current ASICs in the realm of -- in aggregate control between primary and secondary rails in the 1,000 amp range.

So we recently benchmarked our lateral vertical solution, which has been adopted by a notable company in the industry. And we benchmark it relative to our own lateral solution involving a multiplicity of modules or current multipliers deployed on the size of the ASIC. And with lateral vertical solution, we can increase efficiency by something close to 10 percentage points. And that's an efficiency advantage that sets us far apart from anything else that can be done, as you can imagine.

The next question is coming from the line of Jon Tanwanteng from CJS Securities.

My first one is just a follow-up to that. When a customer is using a 700-watt processor decides to go with the traditional power system, is that purely a capacity-related decision from their perspective? Or are there other factors that they're considering such as price or maybe something else that we're not thinking of?

Well, to your point, there is a multiplicity of factors at play. And generally speaking, within a large OEM, there are various consistencies with, as you can imagine, a different set of priorities. But at the end of the day, the solution not only has to work, work well enough. It has to, in effect, enable the compute density and the advance computing density that customers expect. And that's the pivotal question.

In an ideal world, nobody can dispute the proposition that an OEM would like the best of all worlds, a large multiplicity of suppliers, commodity solutions, unlimited capacity and the list goes on. In the real world, obviously, not all of those opportunities are available and recognizing the appraisals may be required to satisfy the essential objectives.

Okay. Fair enough. I think I understand. Having said that, your orders are still up, your backlog is strong. Is there anything at risk in your backlog at this point just because some of these lead times are getting so long and perhaps there is a risk of your customer second sourcing at this point?

Jon, this is Phil. The backlog is strong. I'm very confident in it. Don't see any risk in the backlog at this point in time because of the customers' demands are still really strong and they really want us to deliver that backlog as quickly as we possibly can, which is why the second half of the year is going to look a lot better than the first half for us.

The next question is coming from the line of Donald McKenna from DB McKenna & Co, Inc.

Just to follow-up a little bit on the previous question. Phil, can you just address how much of the increase in the back -- envelop number I've come up with somewhere in the range of $425 million -- is the result of the expanded lead times versus strong new demand?

And then my other question, Patrizio, I guess, would be to you was in the past, I've spent a number of years in manufacturing, and plating process back then was considered as much an art as science and good operators were highly valued. And I'm wondering what you see as the learning curve, how you've been preparing for it and how long you anticipate it will be before you achieve the same efficiency as your current subcontractor?

Let me take your question first. I expect that very, very quickly, we're going to achieve efficiencies far what we've been able to achieve working with the outside contractor. And that's because we have better equipment redundancy, a well-trained team. So I do expect that we're going to be able to ramp faster and more predictably than we've been able to accomplish with all the unpredictability in the way of progress, particularly over the last 9 months. Phil?

Yes. And with regards to the backlog, we're not really expanding lead times. We're working very closely with our customers to make sure that they are obviously placing orders at the right rate. Advanced Products are out 32 weeks or so, Brick Products in that 20 sort of week range. We haven't really touched those. It's demand, it's new programs coming on. As I mentioned in my prepared remarks, we've got the 48-volt bridging applications now starting to ramp. We've got new programs in the China market, in the European market, North American market with new customers. And we've also got continued strength in our existing customer base that we expect to go on for quite a while yet, as I've mentioned previously. So it's really demand-driven.

We do expect to bring in our lead times once we have demonstrated the additional capacity out of the vertically integrated facility. And we have brought about some other limiting factors because the issue, to be clear, over the last 9 months hasn't been simply one of factory capacity. It's also been one of component availability and supply chain disruption. And as you heard in the past that the issues with respect to, in particular, some of that component availabilities, they had to do with unique semiconductor components that we develop and fab for us and we've been able to secure a major step up in that capacity. But unfortunately, it's not just a matter of those proprietary semiconductors, even for garden variety, nickel type parts, right, that some of our products, particularly some of the big products are dependent on.

We've been, to some degree, exposed to the vagaries of the supply chain and of late, particularly the disruption that Jim characterized in his remarks relating to the Chinese Zero-COVID policy.

The next question is coming from the line of John Dillon from D&B Capital.

First of all, congratulations on the booking number. It's really phenomenal. It's a lot greater than I really expected. But today, I want to ask you about the technology and about your technology versus multiphase, which you addressed somewhat in your opening remarks. But I'm wondering, does your point load technology versus the latest multiphase still lower I/O pin count, lower the motherboard copper, lower the socket resistance, free up board space and reduce component counts?

Yes. So our technology with what are now soon becoming a past generation current multipliers and regulators is capable of delivering on the efficiency front with lateral vertical implementation, double-digit improvements in efficiency. And by the way, if 10% doesn't sound like much in one way of looking at it, in another very relevant way of looking at it, which is the heat, right? The heat is generated by the power system and the heat that's generated within the ASIC GPU itself within silicon. That 10% differential in efficiency is a large amount of heat that the solution doesn't need to thermally manage. And that actually translates into increased compute capability as these solutions that push the envelope as they need to with respect to elevating current requirements often run into because they will typically be limited by the ability to thermally manage the solution itself.

So in one way to look at it, the efficiency improvement, which translates into a reduction in heat is an enabler in terms of allowing greater compute density and a bigger value proposition for the OEM relative to its competitors, which we believe is the primary consideration, particularly if capacity is not a limiting factor.

Got you. And that kind of plays into my second question. So let's say you have a 700-watt GPU, and you have the Vicor lateral point load solution for versus the multiphase. It sounds like you've been saving about 70 watts of power loss due to the power distribution network and the efficiency. You're going to be about 70% less than the competitive solutions. Is that about right?

Well, so the specific example I was given was for the first implementation of what we call lateral vertical system, which uses the same, if you will, golden current multipliers, you've seen on the size of GPUs. In our vertical system, one of those current multipliers gets deployed, if you will, at the equator on the bottom side of the substrate. And with that vertical element, same building block just deployed in a preferential way from a power distribution architecture perspective, that current multiplier supports the active area of the silicon, particularly in its primary rail to deliver a much more uniform voltage.

The voltage distribution within the active area gets reduced by nearly a factor of 3, which results in more efficient recession of the silicon itself in its domains and results in better performance in the silicon itself in addition to the reduction in heat and again, the ability to, in effect, run the solution, overclock it or run it at a higher operating frequency and compute capability.

So not only are you saving electricity, you're enabling a higher-performance computer?

Yes. So fundamentally, the efficiency advances that include and leverage to some degree, the ability to deliver a much more, what we call equity potential surface across the active area or the primary rail, these advances enable more power to be delivered within the constraints of the thermally managed solution, which directly translates in either power savings or more compute capacity.

Got you. And then on top of that, it sounds like if you're talking about a data center or a supercomputer that has thousands of GPUs, you're talking hundreds of thousands of wasted watts by going with the computing solution.

There is a green element to this value proposition. Ultimately, I think it's all about compute capacity, which is obviously what drives the industry to continue to raise the bar.

The next question is coming from the line of Richard Shannon from Craig-Hallum Capital Group.

Patrizio, I might follow-up on the last few questions here. Don't mean to ask something that's kind of similar to them, but I just want to delve in a different way here. What you're describing here and certainly what I think most people's belief is you have a solution that's much better than others out there. Power savings and heat savings alone would be a dramatic improvement.

So are you expecting at the highest levels of power delivery in GPU and other ASIC accelerator systems that you're going to be a sole source? Because it seems like if you're not, then customers are using a deeply suboptimal solution. So can you state what your expectations are generally across existing and new customers at these kinds of power levels?

Let me take that in part, and then Phil will weigh in. So in general, as you can imagine, because of the proprietary nature of our solution as a Factorized Power solution and also as a solution I would say with unique power distribution architecture with or without Factorized Power being an element of it. This shows proprietorship. It tends to be a single source, right? And as discussed earlier, customers have to make a choice, which fundamentally comes down to getting more performance and in effect, a better product, a more capable product, but being single source with Vicor or having a multi-source solution with an intermediate bus architecture type of power system with these limitations and handicaps that have implications with respect to the performance of the solution itself.

Yes, just to add to that, I think that we're getting -- if you look at new customers that we have that we've brought on in the last, I don't know, 6 months, 9 months, I mean they have options, right, to use multiphase and best converters with multiphase if they want to do that. But they're not doing that. They're using FICO's lateral for now and also in the future, we're also working on a number of lateral vertical for higher current ASICs and network processes, for example, some of the stuff I talked about in the last call.

So they have choices, and they're choosing Vicor, which says very clearly what the advantage is, as we've spoken about many times of our solution is. So -- those are going to be single-sourced opportunities. And they're doing it because they need the performance, they compete on performance. They want it for their customers. And so that's the way they're choosing FICO over the multiphase.

But just to be clear, it's not that the Vicor solution is the only solution, the other solutions, which tend to be all grouped in the multiphase bus converter, intermediate bus architecture realm, they can be made to work at lower current densities with more noise, with more challenges from an implementation perspective with the growing multiplicity of phases that are harder to make work reliably. And certainly, customers and applications opt for those kinds of solutions if their requirements are low.

We're not suggesting by any stretch that we have a monopoly on the industry. To the contrary, it's a broader industry with a broad set of applications and requirements. And for less demanding applications, multiphase can be perfectly viable, particularly, again, if the current levels and the overall demand of the applications dynamically otherwise can make do with a multiphase application.

Yes. We sort of like that as well because we supply the 48 to 12 bus converter. We've got the highest density, highest performance device there, and we're increasing its power capability. And that's a good -- very good business for us as well. So we're happy to support the multiphase and those lower-power applications, too.

Okay. Guys, those thoughts, I appreciate it. My follow-on question here on your comments about the second half being better than the first half. It seems implicit that you have confidence in your manufacturing getting up to speed. So maybe if you can talk about the next steps here? And at what point do you have full 100% confidence and are going to be running at full rate or at least commensurate with demand. When does that happen? Is that sometime in the third quarter, fourth quarter, not easy to tell right now, but if you can just give us a little bit more color on why you say this about second half better than first.

So I have confidence in our operations team, particularly when they can be in control of their destiny is they're going to be very soon with vertical integration of important process steps. Needless to say, over the last 3 quarters, investors have been disappointed, I've been disappointed by the unpredictability of being able to step up unit rates and revenues as we had anticipated. That's been frustrating. But to be fair, there's been a combination of factors at play, not just the lack of vertical integration with respect to critical process steps, but also the supply chain bottlenecks and upsets that impacted many of the products, including some of the big products and gathering the way of having a predictable output.

I think as we get into July and the start of the third quarter, we're going to have -- we're going to have all the equipment in. The last critical piece of equipment is arriving this quarter. And at that point in time, our team will be in full control of its destiny. And I do expect that we're going to see our revenues catch-up with backlog that is obviously stretched way out in terms of customers having again to place orders with lead times that are longer than they and we would like them to be.

The next question is coming from the line of [Doug Camber from Potomac Capital].

Just a quick follow-up to the previous question. What would have to occur to delay the factory coming online in that July 1 time frame? And as far as components to stock the factory, as you said, the last 3 quarters, global supply chains are affected, it's impacted you guys deeply. How have you gone about solving those issues and getting the components that you will need internally to produce your products from start to finish and export them out the door to meet the customer demand?

Well, I think as we've discussed in the past, with respect to the availability of all our proprietary semiconductors and wafer outs, we have -- we made size in that front, both with respect to getting the output and also securing a long-term sourcing at rates that are expected to support our requirements.

With respect to getting variety parts that have been from time to time unavailable because of general supply chain issues, we're very focused on that. I think -- I'm beginning to see, and I don't want to be to definity with respect to this because it's still relatively preliminary. I'm beginning to see some easing and some increased availability coming online later this year. So obviously, we're watching that very carefully with respect to making sure that once we have all the process steps integrated, and we have our fab capacity in place that component availability is not going to get in the way of being able to use it.

Yes. That was where I was going with it. You have all the machines. You have all the people and then the assembly line standing around because you don't have the parts. That was more my concern because we've all heard about double and triple ordering from most of the layers of the semiconductor world. I was just wondering if you were aggressive in securing the parts you need so that when you do have everything up and running, you guys can start to tug the engine alone.

Well, we do like safety stocks when we can have it. It's been difficult to have it over the last couple of years, and we do have a plan to build some level of safety stock, I wouldn't call that double or people ordering. I would say, in a different perspective that component availability, particularly when it comes to a better balance of our products or components that we are going to use in a Skylake over the next several years, it shouldn't be a hand-to-mouth proposition, right? That's not the way to run the operation from a logistics perspective. We do offer inventory. But if you are call that, double, triple, I would acknowledge that we are looking to build some level of buffer stock where appropriate in order to ensure that our expanded capacity can run without an efficiency or disruption due to an availability of critical components.

Terrific. Yes. And I know that's not the normal way to run a business, but we are in very unusual times after COVID.

The next question is coming from the line of Quinn Bolton from Needham & Company, LLC.

I guess, first, just on the backlog. The backlog up near $425 million. If I've got my numbers right, it looks like you guys did record bookings in the first quarter. And wondering if you could confirm that the first quarter was indeed a record bookings quarter for the company?

No, it was a little bit short of 1 quarter, we had a little bit more Yes, just -- we came close. Okay. So it was just short of Q2 of last year, it sounds like then.

Got it. Second question, you mentioned the vertical power delivery. I know you've got a couple of initial wins on vertical power and some more specialized applications. Can you give us a sense when vertical power for GPUs or AI ASICs might ramp? Is that a 2022 ramp? Or is that out further, say, 2023 or beyond?

So we have 2 lead customers for fully vertical power delivery. And those are customers that as with Vicor, are, in effect, making a calculated choice to be on the very leading/bleeding-edge of technology with, in particular, a set of components for fully vertical power delivery that are progressing in terms of their general capability maturity but not yet at the level of maturity that would provide the level of scalability in the millions of units that is generally part in high-volume applications.

So I think to get a good sense of the landscape, it would be good to think of it in the following terms. There's a lateral power delivery, and you know what that means. So leading edge GPU applications have been using a lateral power delivery architecture with, in particular, current multipliers to the north and the south and in some cases, the west or the east of the ASIC.

And the current that these devices deliver as to traverse laterally right, under the active area of the GPU and ASIC to reach the domains of these devices, that's an architecture that works with a multiplicity of entry points of the north-south, east and west up to 500, 600 amps. Beyond 500, 600 amps, Ohm's law stands in the way of it remaining an efficient and scalable power delivery architecture.

Now vertical power delivery is at the other end of the spectrum. It's very advanced, still somewhat bleeding edge. But with vertical power delivery, those losses and voltage drops that occur as you try and carry the current from the size under the active area of the silicon essentially go away. There is an order of magnitude reduction in power delivery losses, if you can, in effect, supplant the multiplicity of lateral current multipliers with a vertical power delivery, we call it DC amp.

But again, the complexity associated with the full vertical power delivery system entered the lateral vertical architecture, which is innovative in essential respects. It retains the same modular components, the same golden current multipliers that are used in lateral power systems. But as suggested earlier, it places one of them as equator on the bottom side to provide a direct entry point in the central area of the GPU or ASIC. And that brings about a factor of 3 reduction losses and PDN loses. In factor of free reduction in the voltage differential that it would otherwise suffer or have to, in effect, live with within the primary rail of an ASIC.

So we see -- sorry for the lengthy answer to your question, I think it's important to have a good technical grasp of the evolution of the power delivery technology. From lateral and incentive caps, the high-volume applications are going to go into lateral vertical. That provides a 3 fold reduction in power distribution drops and a major improvement in efficiency within the ASIC.

Full vertical is being in effect, tested by some leading-edge applications. In my opinion, in terms of mass production in very large quantities, it's still 1.5, 2 years away. But it's going to be scaling up for certain applications that are PDA dependent on a VPD system later this year and into next year.

So the evolution is lateral, which is obviously a proven technology, but with certain PDN limitations getting the way of 1,000 amp solutions to, again, tracking it from the mass production perspective, lateral vertical, which is a major improvement, 3 fold, but with conventional building blocks, with the same building blocks that we've been making by the millions of units to full vertical, which is still in terms of mass production, 1.5, 2 years away.

The next question is coming from the line of Jon Tanwanteng from CJS Securities.

Patrizio, you mentioned some easing of the supply chain possible later this year. I was wondering if you could talk about the new constraints that you're facing actually in Q2. Are they getting worse or better in the near term? And if there are any levers maybe that you can pull, perhaps in pricing just to help you get better margin?

I think the environment right now, John, is relatively similar to the last quarter. We're dealing with some capacity constraints in the outsourced panel production. That's the same. As Patrizio mentioned, the semiconductor supply coming in. We do have a new wrinkle to contend with, which is zero-COVID policy in China. There's implications there that are tough to predict.

We're making our way through that. So I think we're saying that 2Q feels like Q1 and we're working very hard to clear overdue backlog and support customers. That's our focus this quarter. And most notably, I should say, job #1 is to get our factory expansion up and running. And the team is working extremely hard to make that happen.

Next question is coming from the line of Alan Hicks from Ainsley Capital Management.

I had a question on the front end products that you've been talking about it for a couple of years, and I know you had very high expectations, potentially bigger, bigger than point-of-load products. When was that going to move the needle in terms of revenues? And is it dependent on your 5G technology coming out?

Yes. I think if I understand your question, it has to do with so-called front-end products or AC to DC products. So we've made very good progress. We have 2 major customer engagements. And we've delivered initial sample quantities. One of these 2 engagements is going to be scaling up into substantial revenues in the second half of this year of the order of about $10 million. So the AC to DC component of our revenue mix is still at an early stage, but the products have a good deal of appeal for high-density applications. And this has been tested by 2 lead customers, and particularly one of them is about to get deployed in substantial quantities.

Okay. Can you say what kind of application that is?

Phil, can we say what the AC/DC applications are? Without any customers, I think…

Yes. We've got some data center opportunities, some edge computing opportunities and lighting opportunities. Those are the 3 sort of large buckets, if you like, that we're targeting. We also have a lot of interest in higher frequency versions of those AC products in the aerospace area, where they need 400 hertz operation here in the United States, 800 in Europe.

So we'll be doing sort of defense aerospace versions of those products, which actually is quite a large sum in Europe and North America, somewhere in that $200 million, $250 million range for an opportunity with very nice gross margins. So we're very interested in having defense aerospace versions of these front ends, too, which we are working hard to bring through to the market later this year to the broad market customers in, as I mentioned, data centers, edge computing, lighting, defense and aerospace. Those are the 4 areas.

So the first major revenue contribution is going to come from lighting in the category of several megawatts of lighting power in one location.

Okay. So that's supposed to ship in the second half.

Yes. We'll start building that market in the second half revenues next year and then obviously ramping up in '24, '25.

Is that dependent on your 5G technology? Or does that apply across the board to Advanced Products?

No, that's also using the same engines and control systems that we use for our other components. We leverage a lot of the technology, same packaging, yes, packaging the same. So think of us as a fab of chips as in migrated earlier, convert our housing package, whether it's a lighting application requiring 3-phase front ends or it's an automotive application requiring a 100 volt, to 500-volt bus conversion of 50, 100 or 150 kilowatts.

All this -- or a point of load application requiring the same current multipliers we were describing earlier. They're going to be going through the same fab, going through the same process steps and leveraging the same common denominator methodologies in terms of components, control systems and most of all, packaging technology in effect, wafer-like common denominator panels.

Okay. And then just real quick. How soon do you expect production revenues in the auto EV item?

End of '23 for the first customer, first OEM and then increasing '24, '25, '26, of course.

And I think with that, operator, could you please call an end to the call.

Of course. So everyone, that concludes your conference call for today. You may now disconnect. Thank you for joining and enjoy the rest of your day.