Good morning, and thank you, everybody, for joining us at the 28th Annual Needham Growth Conference. My name is Quinn Bolton. I'm the Semiconductor Analyst for Needham. It's my pleasure to host this fireside chat with Vicor. The company is a leading supplier of high-performance power components and systems based upon a portfolio of patented technologies. Vicor's easy-to-deploy modular power solutions for power delivery networks provide the highest density and power efficiency from source to point of load. Joining me from the company this morning is Chairman, President, and CEO Patrizio Vinciarelli, Phil Davies, Corporate VP of Global Sales and Marketing, and Jim Schmidt, Corporate VP, CFO, Treasurer, and Secretary. Patrizio, Phil, Jim, thank you for joining us.
Thank you. Good to see you.
I want to remind everybody listening to the fireside chat, if you would like to ask a question to management, please do so by submitting a question in the dialog box at the bottom of your screen. I'm going to hand it over to Jim now for the safe harbor language, and then we'll start with my questions. Jim, over to you.
Our remarks to follow, including answers to your questions, may contain statements that we believe to be forward-looking statements within the meaning of the Private Securities Litigation Reform Act. These forward-looking statements are subject to risks that could cause actual results to be materially different from those expressed or implied by such forward-looking statements. These risks include, among others, matters that we have described in our most recent Form 10-K and Form 10-Q, and other filings we make with the SEC. Further, Vicor has not yet disclosed our financial and operating results for the fourth quarter ended December 31, 2025. Therefore, we are unable to discuss or answer questions regarding our financial and operating results or developments for periods after our third quarter ended September 30, 2025.
Thanks, Jim. Guys, I wanted to start with the licensing business. In 2025, Vicor had a number of milestones, including securing a license agreement in the first quarter with a large hyperscaler, announcing a $45 million settlement in the second quarter, and signing a second license agreement with an OEM in the third quarter. As you look into 2026, just overall, what's your level of confidence that based on the ITC success, you'll be able to continue to expand your base of licensees?
We had the first ITC success. We came to fruition last year. Early this year, within the last week, we've started the second ITC case, and there'll be more. We have a lot of IP. There has been an exciting level of infringement with respect to that IP. The patents have been asserted thus far and have been focused on NBM technology, but there are other patents that are being infringed. We have just begun with respect to monetizing a big value proposition with our IP portfolio by way of a licensing business that runs parallel and in many ways is very synergistic with our module business.
Patrizio, I want to make sure I heard you correctly. I think you referenced a second ITC action. Could you just expand on that if I did hear that correctly?
No, you heard that correctly. That was filed early in January. I think it is at this point a matter of public record. I'm not going to make a lot of comments about it, but we filed district court cases in the Eastern and Western District of Texas. Those are for sure a matter of public record. We filed an ITC case that may not be visible yet, but it should be in a matter of a few weeks.
And can you say just broadly, does it pertain to your power module patent portfolio, or does it pertain to your Vertical Power Delivery patents?
This is still focused on our NBM technology, and it's focused on some initiatives to work around the findings of the earlier case.
Yeah, that is actually my next question was just for the original ITC case. Have you seen Delta submitting workaround modules to Customs? And has Customs declared any new power modules no longer infringe the original ITC patents?
So Customs has been helpful with respect to banning the importation of products that are infringing and not licensed. This, the ITC's ability to protect U.S. companies from foreign suppliers of infringing products, is a very effective tool with respect to protecting American innovation. So we're very satisfied with what we've been able to accomplish in that regard. Now, to your point, after the first exclusion order with respect to the initial set of patents that were asserted with respect to NBM technology a year and a half ago, some of the respondents submitted so-called workarounds. But in hindsight, those workarounds are really no workarounds. They still infringe our NBM technology IP, and that's becoming an issue with this new case.
Staying sort of on the legal front, what's the status of your appeal of the ITC ruling that the company had given away its patent rights to Foxconn through the boilerplate purchase order language? Has there been a ruling on that appeal? Is Foxconn currently subject to the ITC exclusion order?
So that's actually, in terms of the real world, had no impact whatsoever with respect to what OEMs, hyperscalers have been able to do in terms of accessing NBM technology. But in any case, our appeal is fully brief. We had the last word late last year. It's now ripe for proceedings at the Federal Circuit. The next step is going to be, in our argument, that it's going to be scheduled probably late this year. But that case, to be clear, doesn't play a significant role with respect to the next set of events on the general front of asserting IP and keeping infringing products from entering the U.S.
And you mentioned that you think there are others that are infringing your IP. I know you've got a pretty meaningful patent portfolio around Vertical Power Delivery. Do you think that you'll take action this year to enforce your rights under your Vertical Power Delivery patents?
We will take steps to protect all our IP. So we're taking it in a logical sequence with measured steps. As you know, Vertical Power Delivery has been somewhat elusive in terms of real-world implementation. It has been very challenging in terms of mechanical, thermal cost aspects. What's been implemented thus far is primarily with one large hyperscaler. It has relied upon what we call a first-generation VPD, which Vicor pioneered many years ago. As with every first-generation technology, it has significant limitations in terms of, again, its mechanical, thermal, and electrical attributes. We've been very much focused on our 2nd generation VPD. In general, the first-generation VPD hasn't gathered steam in terms of general adoption yet to a level where action on our part is warranted, but we're getting closer to that point in time.
Okay. And to the extent you were successful in protecting or securing licenses to your Vertical Power Delivery patents, just sort of at a high level, would that generate a similar royalty stream to what you currently receive for your power module patents? Would it be higher? Would it be lower? Just any sort of directional thoughts on how that could expand the licensing business if you're successful?
In general, the value of a license is commensurate to the value of the IP. Vertical Power Delivery is critical to leading-edge AI and separate apart from the bus converter technology that's been the primary subject of our enforcement strategy today. We expect that licenses that encompass VPD technology will carry substantially higher cost, which is still, though, cost-effective in terms of the value proposition to AI OEMs and hyperscalers that use that technology.
Great. You've mentioned in the past that licensing contracts incentivize OEMs and hyperscalers to source products directly from Vicor. Are you seeing the OEMs or hyperscalers take advantage of this incentive?
Yes, because it makes economic sense, and it makes sense in other ways for them to use Vicor as a supplier in addition to, in effect, using Vicor IP by way of a license. So that's been very effective. I think a licensing model is very well to have. It works. And we've already seen significant benefits in terms of increased demand for our products and increasing usage of our first fab that stem from the licensing program.
Do the licensees typically purchase just the NBMs that incorporate the patents, or do any Vicor product purchases sort of fall under those licensing agreements?
The licensee is motivated to use any Vicor module, but thus far, the usage associated with the licensees has been focused on NBMs. I expect in the future it will encompass 2nd-gen VPD.
Okay. Great. And then my last question, on the third quarter conference call, you mentioned the licensing and royalty business could approximately double over the next couple of years. Are you still sort of confident in that trajectory given everything that's gone on in calendar 2025 and year to date in 2026?
Yes. And that's partly related to new actions that we just recently and to this in the future will bring about. I think the ability for module makers, contract manufacturers, hyperscalers, and OEMs to use our technology while eluding our patents is the window of opportunity, is narrow. And we have a very comprehensive strategy. So on that basis, that we have confidence in building our licensing business, again, in synergy with our module business.
Does the doubling incorporate any assumptions around vertical power, or would you expect to be able to double even on the strength of the power module patents that you have today?
Power module patents include VPD, right? But I think if your question could be refined along the lines of, could we get to a double just on the NBM patents?
Yes, effectively. Yeah.
I think we could, but certainly the VPD technology has complements that in a way that gives us greater confidence with respect to having a very successful, very high-margin growth licensing business.
Excellent. Okay. Well, I wanted to move over to the product side and discuss the Gen 5 technology and Vertical Power Delivery. Maybe for folks, if they're online that are newer to Vicor, can you spend a minute, Patrizio, just talking about the benefits of Vertical Power Delivery and in particular, the advantages of your Gen 5 or 2nd-gen VPD technology that you're currently bringing to market?
Yeah. So this technology indeed is driven by escalating requirements in terms of quantity, in particular, by advanced AI processes. Current densities, in order to enable more compute and higher AI performance, are rising well above one amp per square millimeter. And that's the point beyond which, with lateral power delivery, given the low voltage, high current requirements of XPUs, GPUs, TPUs, you get to a point of diminishing returns where fundamentally you can't deliver the current without excessive voltage drops that fundamentally handicap efficient power delivery. So that's where being able to carry current power at low voltages into XPUs becomes critical. As suggested in earlier comments, we first enabled this kind of capability through what we refer to as a 1st-gen VPD.
That was done given the limitations in terms of current density a number of years ago, several years ago, by the stacking of our processing, filtering, redistribution layers, which make for solutions that are relatively thick, heavy, and thermally challenged because of how heat gets trapped within the stack. With our 2nd-gen VPD, we've advanced the current density to a large multiple of one amp per square millimeter. And that facilitates a 2nd-gen form of VPD where the power delivery, the filtering can all be done without having to stack layers. It can all be done fundamentally at one elevation, either at the very bottom of a substrate or within a package. And that greatly simplifies it, makes it more cost-effective, removes the thermal and mechanical challenges.
On the third quarter call, you mentioned that your Gen 5 VPD solution had met the target specifications for your lead customer and that you're progressing to a first quarter of 2026 production launch. Are deliveries to this lead customer still on track?
Yes. In fact, with respect to an earlier generation product, they are escalating at a rapid rate. We're also progressing with respect to the more advanced solution where we are getting to the 100% target and we're getting beyond that. And so we are satisfied with the progress being made. We got more opportunity coming the first half of this year to further raise the bar with respect to general performance. But we believe that our methodology, our FPA approach to VPD has got all the right attributes in terms of addressing future needs by GPUs, XPUs, TPUs in the AI space.
You also talked about on the third quarter call that it has started engagements with selected customers comprising a hyperscaler and OEMs who informed you that your 2nd-gen VPD is really the only solution that can meet their processor requirements. How are those engagements progressing? And would you expect to still ship product to those customers into the second half of 2026?
I'm going to turn that over to Phil.
Okay.
Yeah, so we're continuing to move along with those early conversations, as I discussed before. Very soon, we're going to be demonstrating. We have two demonstration systems of slightly different architectures that we're going to be demonstrating in Santa Clara, our offices there, coming very soon to a number of customers that will come in for a half a day with each of my applications engineering teams there. Yeah, so that's coming very soon, and that's progressing along pretty well.
Excellent. This may be a little premature knowing that the sampling is just sort of getting started. But Phil, have you been able to secure any new AI processor design wins since the third quarter call?
So the focus right now is on our lead customers we've talked about and making sure that they've got a very nice ramp going on with their existing system. And we're focused on bringing that up to another level of performance, as Patrizio spoke about. And then the next step for us is the Gen 5 rollout, as I mentioned, coming very soon to the general market, as it were. And so further designs will start to happen later this year.
Got it. And for the general that broader availability production ramps probably then in 2027, maybe some prototype deliveries in the second half of 2026, but volume design was probably targeting 2027 ramps.
Yeah, absolutely. That lines up with their chip development. If you look at what's going on in the market for different hyperscalers and OEMs, chip OEMs, it lines up pretty well with that, yes.
Excellent. The other opportunity that we've heard a lot about from investors is the 800-volt or 400± high-voltage DC rack opportunity. Can you talk a little bit about how Vicor is positioning for the 800-volt or the 400± rack? I know you guys have done 800- volts for a while. Maybe tell us what products you're targeting for that opportunity.
So once again, Vicor pioneered 800-volt high-density modular solutions. We have proprietary engines for accomplishing that and methodologies, which again are subject to comprehensive IP. That's probably going to, from an enforcement perspective, come later. To be clear, the 800-volt initiative, which was really launched in earnest within the last year, has got some time to come to fruition. And that's relevant both from the IP perspective in terms of the modular solutions that would best address those requirements. To your point, we've had and been in production with 800-volt fixed threshold bus converters for quite some time, years. Our power levels in the few kilowatts. We have a 10 kW module. So state of the art, very high power density in development approaching completion would be something that probably around mid-year.
It is an evolving requirement that represents opportunity because it's an additional stage of conversion in the total power delivery. It will represent incremental opportunity both in terms of module sales and intellectual property realization.
Patrizio, from a process technology perspective, I think there's some in the investment community that believe 800-volt may be exclusively the domain of silicon carbide or gallium nitride higher voltage solutions. I think some of the architectures we've seen proposed for 800-volt include ISOP or input series output parallel architectures. Can you play in this market with silicon MOSFETs or technology that Vicor has used for, I think, decades now?
Yeah. So some of the characteristics you referenced, as well as the sample converter structure of these solutions, again, fall within Vicor IP. But we have, in other words, novelties that work to advantage with respect to this solution by, in effect, the powertrain being able to utilize silicon FETs, lower voltage silicon FETs, which in terms of high-frequency switching can provide a higher figure of merit than wide-bandgap semiconductors. Because fundamentally, even though there's some level of stacking that the competitive alternatives have been pursuing, again, with some IP ramifications related to that, there is no substantial division of the 800-volt node to a level that enables very high-frequency switching. In other words, the wide-bandgap semiconductors, the GaN FETs, the silicon carbide FETs are, in some respects, ideal switches.
But because of the fact that their value proposition relates to switching a high voltage, and at high voltage, capacitance gets in the way of high-frequency power processing, they are somewhat limited with respect to how far they can operate in frequency. And that's the limitation that with our further advances in technology, we have overcome.
Excellent, and maybe for Patrizio or for Phil, as the 800-volt or 400± racks come to market, how much might that expand your TAM from the current TAM that's mostly 48 NBMs and then obviously the factorized power point of load for processor power?
Yeah. So I think if you look, just to take some baselines, if you look at the so-called megawatt rack that's being talked about, a megawatt rack would have probably somewhere between $10,000-$15,000 worth of power module content if it were to go power modules. So the question there is power modules versus silver box, power supplies built with discretes. So there are companies that we've talked to that say it's going to go discrete power supply silver boxes. There's some companies looking at power modules. That gives an opportunity for Vicor power modules if there's a density play. But that market could get incredibly commoditized very quickly. So from a power module, $10,000-$15,000, all the discrete guys are going to be fighting over what's there, and that's going to become a margin problem. Everybody competing for the same stuff.
It could end up going that way, Quinn. We're keeping an eye on it. But we obviously have, as Patrizio said, a power module play, high-density, high-performance power module play, and an IP play. So we'll be watching that carefully as it plays out.
Yeah. So the NBM is an IP play. It was commoditized as a product, but that did not deprive an opportunity to capture value. And a similar opportunity exists with 800-volt bus and bus conversion down to a lower voltage.
Okay. Got it. Got it. So it sounds like it starts maybe as a high, especially on the integrated higher-end integrated modules, but could over time turn into more of an IP play to the extent it does commoditize. Understood. Maybe shifting just to some of the other end markets, can you give us an update on the automotive opportunity? How are you feeling about OEM engagements and design wins in the automotive segment?
I think the exciting part of automotive for us is the expansion of 48 volts, not just in electric vehicle platforms, but through mild hybrid, through even ICE platforms as well. That is obviously a very nice opportunity for our NBM technology. In terms of you look at the North American market versus the China market, Japan, Europe, everybody's looking at different mixes of electric, pure electric, mild hybrid, plug-in hybrid, ICE. I mean, it's been a bit of a chaotic market over the last year to two years. However, having said that, we've continued to expand our collaborations in the last number of quarters with different Tier 1 companies around the world. We've got three nice ones just started in Japan, for example. Looking at 800 volts to 48 volts, 48 volts to 12 volts, and 12 to 48 for legacy systems.
So it's a mix of different bus converter and regulator, 48 to 12 regulated products for us as well in terms of working on some onboard charger applications as well. But we have two OEMs that are in production, and those programs will, as I've mentioned before, ramp through this year and in through 2027. But I don't think automotive is going to be a big contributor to Vicor revenue until the 2029, 2030 sort of time frame. That's sort of the time we're looking at.
That's kind of when some of those higher volume opportunities have started production dates.
That's right.
Okay. Perfect. And then just quickly on the outlook for the aerospace and defense and industrial businesses, how are you feeling about those end markets as we come into 2026?
Very, very good. I think especially in the industrial side of things with automation in terms of semiconductor manufacturing, automatic test equipment. We've got some great design wins there from a couple of years ago that are now starting to ramp through the 2026 time period because of AI, both on the processor side and also on the memory test side. So that's exciting for us, and aerospace and defense continues to grow. We've got a new chipset on the space side of things launching in the second half of this year. So both of those markets I've predicted will double: industrial in five years and aerospace in seven years. This is going back to a prediction a year ago. We're on track to do that, so yeah, they'll be continuing to be major contributors to our top and bottom line.
Excellent. And then maybe looking forward, Patrizio, I know you've completed the first ChiP fab in Andover. Obviously, we're looking to get better utilization out of that facility. But as you look forward over the next couple of years, can you give us any thoughts on timing or thoughts on a second ChiP fab? Do you have any current plans? And would that second fab be a Vicor fab, or would you look to potentially build something through partnership or maybe even through a license agreement?
So all of the above. So the first fab in Andover is getting filled relatively substantially this year. We've been actually within the last couple of months very focused on a second facility. And we may be securing the location very soon for that facility. We're also having discussion with respect to licensed alternate sourcing. We have customers that, given their rapid ramp and potential for other magnitude increases in volume requirements, they would like to have alternate sourcing either by way of a second fab or ideally by way of a licensed alternate source. So we are beginning to explore those opportunities as well. So in a way, it's a combination of all the things you referenced.
Okay. And would you sort of think about that second? It sounds like the second fab would support the high-volume AI data center HPC market. But as you look forward to the automotive ramp, do you think automotive would at any point require its own dedicated manufacturing, or is a second plant likely to serve all end markets?
I think it's probably too early to say. There's a lot of common denominator opportunity with respect to a fab being able to address requirements in different markets. Given a fab methodology, the counterpart to wafers in some of the adaptive fabs, in our case, become panels, right? So the ChiP panels is converted to in-package panels. They can be processed for AI applications and with the same process steps for automotive applications. I think in the automotive market, there may be opportunity down the line for, again, licensing to large Tier 1, Tier 2 type of players. So I think those kinds of opportunities are further out than the ones that are nearer term with respect to the growth in the AI space.
And sort of last question on a potential second factory. Just given the political landscape, the Trump administration push for onshore manufacturing, would you envision most likely a second plant being U.S.-based, or would you consider overseas locations for a second facility?
No, we're moving ahead as suggested earlier in the very near term with the second facility, not that far away from the first, but far enough so that it is not dependent on exactly the same electric power feeds, and so it can be considered fundamentally sufficiently remote and sufficiently independent and fault tolerant with respect to any limitations associated with the first.
Okay. Got it. Maybe for Phil, as we start looking just sort of at the business, the book-to-bill for the business has hovered around parity plus or minus for most of the past two years. Do you think the company is positioned to get back to a sustainably positive book-to-bill this year?
Yeah, I can take that one, Quinn. And so our book-to-bill ratio is now well above one.
Excellent. Okay. Well, we'll look for more details, obviously, when we report fourth quarter results here in a few weeks. Maybe another one for you, Jim. Just you've kind of mentioned in the near term the Andover fab is currently underutilized, and that's weighing on margins. The standard product margins are pretty good, but is there anything you can sort of give investors just how they should think about margins as the utilization rates in Andover begin to rise?
Yeah. So our standard product margins are among the best in the proprietary semiconductor industry space. So we're very excited about the standard margins that we're able to generate. And those margins overall will lift as the capacity utilization increases. And apart from the margins improvement that we're going to see, Quinn, as utilization improves, our licensing practice will really ratchet up our gross margins over time. And we believe now to a level that will exceed our margin model in our financial model. And so therefore, we're thinking that we would increase that from 65% to 70%, given the strength of the licensing business and the lift in product GM that we're going to get as the loading increases.
Excellent. Well, I know we're coming up to about our time limit, but I wanted to ask maybe just your thoughts. What would your highest priorities be for 2026?
Realizing the capacity in the first fab while laying the groundwork for a second fab, possibly as early as middle of 2027 or maybe third quarter of 2027.
Excellent. And then just any last thoughts or anything you'd like to pass along to investors who may be listening on the call? Anything you think I may have missed in my questions?
No, I don't think so. I think you've asked really good questions, and we're excited about this year. Very excited about it.
Yeah. No, it sounds like a lot of great things going on. We'll certainly look forward to your fourth quarter earnings in a few weeks and look forward to tracking your progress, and gentlemen, thank you very much for joining us at the Needham Conference again this year.
Thank you. Thank you. Thank you.
Thanks, everyone.
Bye-bye.