Rambus Inc. (RMBS) Earnings Call Transcript & Summary

Hello, everyone. Good morning, and good afternoon, and good evening. This is Vivek Arya from the Bank of America Semiconductor team. Really delighted to have Luc Seraphin, Chief Executive; and Rahul Mathur, the CFO, of Rambus join us today to share their views about the company and the industry. We will go through a Q&A session. But please, if you have any questions in between, please feel free to e-mail them to me through that console that you have, and I'll be sure to bring it in. But with that, a very warm welcome to Luc and Rahul, and maybe, Luc, let me turn it over to you to perhaps give an overview of Rambus to people in the audience, who might not be as familiar with the company.

Yes. Thank you, Vivek, and good morning, everyone. So let me give you a brief overview of Rambus. We started our company about 30 years ago as a memory interface company. We actually define the memory interface technology that is now utilized by all DRAM manufacturers and processor companies as processors have to interface to DRAMs. We started the company as a patent licensing business initially, and that remains the backbone of our business today. It's good because it's predictable, it's profitable and it's stable. And it's a very strong cash generator for our business. Several years ago, however, we evolved our business model. We invested the cash generated by patent licensing programs into developing IP Cores and products that are important and relevant to the road maps of our customers. We recognized the growing importance of semiconductor embedded security as well and we also invested in developing solutions to address this trend. So today, if you look at Rambus, we deliver these critical technologies of high-speed interface and embedded security in 3 forms. In addition to our patent licensing programs, we have an IP Core business and a semiconductor product business. For the IP Core business, we develop complex high-speed interface and security cores that we typically sell to semiconductor companies. These semiconductor companies then integrate our cores into their products and they sell their products to their customers. For the product business, typically the buffer chip business for us, we design and sell the semiconductor products ourselves. Last year, we further focused the company by redefining the perimeter of our company. We sold our payment and ticketing business, which was not central to what we do to Visa. We also bought 2 companies that strengthened our position in interfaces and security technologies. We bought Northwest Logic, a memory and interface controller company, which complements and strengthens our offering in high-speed interfaces, and we bought the inside secure portion of Verimatrix, which strengthened our position in embedded security. So based with this redefinition of our perimeter, our strategy continue to be based on 3 pillars: one is to continue to focus and strengthen our position in developing key interface and security technologies in our target markets of data center, communication, AI and automotive. The second is to continue to improve our operational efficiency by leveraging the synergies that exist between our different businesses. And the third one is to use our continued ability to generate very strong cash from our operations to support our M&A programs to accelerate growth. So in a nutshell, this is how we evolved in the company and where we're going.

Excellent. So maybe let's get into some of the Q&A, first, starting with some near-term issues, and then we will dig a little deeper into the longer-term trends. On the near-term side, maybe give us a sense for how has the COVID pandemic impacted your business, both from a supply and the demand side perspective? Because it seems like the kinds of things, Luc, that you went through, suggest that the business should have been perhaps more resilient to some of these trends. But I think it will be good to hear from you, how has the pandemic affected you positively or negatively so far?

Yes, great question. I'll start by saying that I've been impressed and humbled by our teams going through that transition. Before talking about the market, I think the employee base of Rambus did an outstanding job in moving to a work-from-home environment without disruption. So now being -- this being said, if you look at the structure of our business, as I said a few minutes ago, we have patent licensing, buffer chips and IP Cores. The patent licensing business is completely immune to the macro environment. So it was completely immune to the COVID-19. These are just legal agreements with partners. The buffer chip business has been very resilient for a couple of reasons. On the demand side, the work-from-home and learn-from-home environment has created demand for data center technologies, and it's right at the center of what buffer chips address. On the supply side, our supply chains are based in Taiwan and Korea. These 2 countries have not been really affected as much as other countries like the Philippines or Malaysia. So that business was quite resilient from a supply and demand standpoint. And on the IP Core side, remember IP Cores is we develop IP that we sell to semiconductor companies. But these semiconductor companies take time to integrate that IP and then ship to market. So by the time these products go to market, it can take several months. So our activities there has not suffered much as well. The only thing I would say is that it might have slowed down some of the engineering interactions with our customers. But overall, if you look across the 3 types of activities we've had, we've been very resilient to the COVID-19 situation.

Got it. And that, I think, has also been one of the more interesting aspects. Thankfully, that most of the semiconductor companies have adapted to this work-from-home. And from what we have heard from a number of companies, especially in the computing side, they have managed to release products on time. And I think, Luc, to your point, that workflow has not been disrupted, thankfully.

Correct.

The one other near-term aspects of the industry that it's on top of mind for investors is the whole U.S., China trade tension dynamic. Give us a sense for how it has impacted you, if at all, so far? And do you think that all the new restrictions that could potentially come from the Department of Commerce, will that -- will those restrict your ability to sell to any customer directly or indirectly in China?

Yes. That's a great question. And I think my answer will be similar, not completely identical, but similar to the answer of the COVID-19. If you look at the business through the lens of the 3 go-to-market activities. Patent licensing, again, is completely immune to the tension between the 2 countries. We made an announcement recently that we had signed a patent licensing agreement with CXMT, which is a DRAM manufacturer in China. But this is a pure patent licensing agreement. This means that they just pay the right to develop their own DRAMs. There's no technology exchange. There's no support activities whatsoever, as if and when CXMT ramps DRAM that they manufacture, then they will have to pay royalties to Rambus. So that business is immune from that situation. If you look at buffer chips, our buffer chip customers are the 3 DRAM manufacturers, 1 is U.S.-based, the other 2 are Korea based. Of course, a portion of their customers are based in China. But the very fact that 2 of these direct customers are Korean and 1 is the U.S. make the buffer chip business also kind of immune from that situation. And again, like for COVID-19, we may have here and there some issues with some IP Core engagements with Chinese partners. But remember, IP Cores is a small portion of our business and China is a small portion of our IP Core business. So the impact on Rambus is really minimal. And this is another area where our company has been quite resilient.

Got it. Maybe to follow-up to that. Is China a smaller customer because their state of technology is behind. So if in the future, as they try to get ahead, they would need more access to your technology? Or I assume, obviously, your big memory customers, for example, are outside of China. But over time, do you think there might be greater reliance on customers in China to achieve your growth objectives?

China will remain a demand creator once we've passed some of the challenges we see today on the COVID side and on the tension side. So we will continue to monitor that. China was on the path to develop their own semiconductor industry. DRAM was at the center of that. And any company that has the ambition to build DRAMs will have to use foundational technologies that we have invented. And therefore, we will probably have to go through more of these licensing agreements, patent licensing agreement for Chinese companies who have the ambition to develop DRAM for the Chinese market or for the global market. Now the other engagements we have with China through IP Core. There's a lot of start-up companies in China, and we continue to deal with them. And it's just been slowing down. But in the long run, China will continue to create demand for us.

The next thing, I got a question from an investor, which is, how does your business contrast with Synopsys or Cadence? Are they complementary? Are they overlapping in certain areas?

They are complementary, overlapping, all of the above. First of all, as we develop chips, Cadence and Synopsys provide tools for us. So we are a customer of Cadence and Synopsys. In some areas, we are complementary. For example, we do provide memory controllers and PHY controllers. When on their side, they provide the PHYs to the customers. So in that case, we are complementary and we partner. And there are some other areas where we compete with different customers. And that's not unusual in the semiconductor industry.

Got it. Okay. Next thing, I think you started, Luc, to go through some of the end market drivers. Maybe let's start with the data center. Obviously, there is a secular aspect to growth in cloud computing. But just maybe on a near-term basis, do you feel that there has been perhaps a pull-in of demand into the first half [Audio Gap] decelerate as we go into the second half?

What we've seen is, first of all, an increase of demand for data center applications because of the work-from-home and learn-from-home environment. And I think the way the industry has reacted to that is that they wanted to make sure that the supply chains to support that demand were not disrupted. So everyone in the supply chain took what I call an insurance policy. They ordered a little more than they typically did in the past. But I wouldn't say that they have overbooked unreasonably. I think they just took an insurance policy to make sure that the supply chains don't break any place as demand goes up. So we've seen that. We still remain very confident with the long-term trends of that business. It's just that there's a lack of visibility for the second half of the year. It's more of a lack of visibility than the risk to the business. But we remain confident in the long-term trends for this market.

All right. How are you looking at the overall supply-demand aspect of DRAM? Because I think that's another place where there is some concern about some digestion into the back half by the cloud customers, like how sensitive is your business to just the movement in DRAM pricing?

So our business for the buffer chip and for patent licensing is not highly correlated to pricing. Our licensing agreements are not -- patent licensing agreement are not correlated to pricing because the -- at a high level, they are fixed agreements. And on the buffer chip side, the buffer chip supply to our customers does not depend on the DRAM price of these customers to their customers. So we're not sensitive to the pricing of DRAMs in the industry. What we're sensitive to is the overall volume demand. As -- one thing that may happen is a shift from DRAM demand from consumer to data centers. With COVID-19 and the economic impact of COVID-19, maybe consumer demand will go down. And therefore, the demand for DRAM in the consumer space might go down, but we don't play in that space. On the other hand, DRAM demand for the data centers is probably going to go up, and that's where we play. So I think that's a positive trend for us.

Got it. Makes sense. And then I think you mentioned the CXMT agreement. When do you expect to generate some material revenue from that? And if you could just give us some sense of the structure of the agreement? And then just how it kind of places you within the Chinese market?

Hi, Vivek. It's Rahul. Let me jump in and see if I can help there. We were really pleased to announce the agreement with CXMT. This is actually a royalty-bearing agreement. And so we don't expect significant revenue contribution to this agreement in the near term simply because the production from CXMT isn't expected to be that significant in the near term. I think what's really exciting to me about the agreement is that, what it does is shows an independent third-party that says that our patents are going to be relevant, well beyond the existing renewal and extension time lines for the other big 3 licensees. Our patent licensing agreements are usually 5 to 7 years. And for those of us who follow us a little bit more closely, we have one of our big 3 licensees except for an extension at the end of this year, and we've been very confident that they'll extend. We have another one that's up for renewal in 2023 and then a third that's up for an extension in 2024. So assuming that we get the renewal at the end of this year, and recently you'd see extensions and renewals in '23 and '24, so a long runway. But if you look at the timing of the agreement with CXMT, that certainly extends beyond that renewal cycle. So it's an independent third-party who's really saying that our patent portfolio will continue to be relevant beyond those time frames. So again, not significant revenue estimated in the short term, but as CXMT ramps, and we expect they will, we'll start to see more coming in there. Does that help answer the question?

Yes, absolutely. I think if we stay with that royalty revenue recognition, Rahul, you mentioned the different customers that are up for renewal. Maybe if we just take a step back, what percentage of your revenue is being generated from kind of the long-term licensing and patent agreement. And then what goes into the renewal aspect of these discussion that you mentioned? I think the renewal with Micron, right, is up later this year. Just conceptually, without being customer-specific, what goes into a renewal? And just what percentage of your revenue is generated from these long-term licensing and patent agreements?

Sure. It's a great question, Vivek. And I think 1 thing that's a little unusual for Rambus is that, we've been impacted by the change in accounting standard to ASC 606 more than any other company I've seen. And honestly, it's a testament to exactly what Luc said earlier, which is, at first blush, these agreements are relatively fixed and determinable. And there's no performance obligation. So that means that there's hundreds of millions of dollars that we receive every year for which there is no performance obligation. So what's happened with ASC 606 is that if there's no performance obligation, then you take the entire value of a contract as revenue in the period in which you sign it. So if we had a $400 million contract, which was $100 million a year or $25 million a quarter, than under the previous revenue standard, we would recognize $25 million each quarter as we bill and collect from our partners. Under ASC 606, we would recognize almost the full $400 million the period in which we signed that agreement, then we put the balance, call it, $375 million on our balance sheet. And then each quarter, as we bill and collect, we wouldn't call it revenue, but we'd decrease that unbilled contract asset. Now if you have an opportunity to look at our investor presentation that's available on our investor site, and it has in our safe harbor, which references you to a lot of our filings, which describe this in much more detail, it gives a lot of those details. So specifically, for example, on our balance sheet as of the end of last quarter, I think we had $487 million associated with royalty agreements for which we have no performance obligation. Now to answer your question, what most of our investors and analysts do, and this is what we do internally as well, is that what we do is, we combine what we bill our partners and we refer to that as licensing billings, which is an operating metric, with what we report for revenue for our silicon IP partners, which is what we classify in contracts and other and then for chips, which is what we classify in product revenue. And what we do is, we combine those 3, and that's really a better indication of the overall billings of our company. And if you look at what we presented at our Analyst Day in September, I think what we said is the midpoint for licensing billings for 2020 was going to be between $220 million and $240 million. And as Luc mentioned, at first blush, that is immune to any changes because the majority, the vast majority of that billing isn't dependent on performance, on royalty or units. And if you look at where how many analysts have us for both the product as well as the silicon IP business, that also could be anywhere from, call it, $180 odd million. So I think 2020 is a really key year for Rambus because it's one of the first years where what we bill our partners for our licensing agreements could be equivalent to what we recognize as revenue from our product agreement. And I'll tell you Vivek, if 4 years ago, people would have said to you that Rambus is going to bill as much in patent licensing as they're going to have in product, I don't think many people would have believed you. So I think that gives you a bit of background, just in terms of a little bit of the challenge of accounting, but also how our product business has grown very nicely over the last several years. And as I mentioned, we have great visibility into the licensing billings, which, again, gives us a lot of comfort in any period of uncertainty here. And then what we will do is continue to add product revenue, both organically and inorganically. So long answer, but I hope that was helpful.

No. Absolutely. So is there a certain target mix that investors should think about your business in terms of product versus some of these more recurring licensing and royalty parts of the business? Is it like a 50-50 mix that you think is the right mix over the next 2 to 5 years?

Right now, it's close to 50-50 and that might be the case in 2021. But what I expect is that licensing billings will stay roughly flat, right? As we indicated earlier, most of those agreements are fixed and determinable. So there's not much variability. But it gives us hundreds of millions of dollars of cash, of cash flow, and that's something we should talk about. And then as we grow from a product perspective, then you'll see that product portion becoming larger and larger as we get bigger as a company. But as I mentioned, the fixed nature of that license agreement really helps us from cash flow and that's something we haven't talked about yet. And one of the things that's most exciting to me about Rambus is exactly that, our cash flow, and it's supported by those agreements. Last year in 2019, we generated $128.5 million of cash from operations. And in Q1, we generated $37.3 million, and that's a fantastic number, given our top line from a billings perspective. And we're relatively small from a CapEx perspective, where typically $10 to $12 million a year of CapEx. 2020 is higher because we were in the process of moving to a new headquarters facility. But as we show leverage in our model and continue to grow product revenue, I expect to see that cash from operations continue to grow.

Got it. No, I think that's a very good point. You mentioned that looking at cash flow is perhaps a better metric for judging the success of the business. Do you have any specific metrics around that, Rahul, in terms of -- is it a certain percentage of sales that investor should be tracking over the last 2, 3 years, it's been a very strong aspect of the business, but is there a certain metric that you're driving towards in terms of cash flow generation?

Yes. So we've grown cash flow very nicely. And if you look at our investor presentation, I think we have a slide that shows exactly that. I encourage you to go through. I think historically, under ASC 605, we had targeted somewhere between 37% and 40% of operating income. And so as we get incremental product revenue, what we expect is to have that or better from an operating income perspective and continue to grow the cash from operations there. But I think it's interesting to me because one of the things Luc mentioned in the earlier conversation is just some of our peers. We are partners and collaborators and competitors with peers who get much higher valuations in their cash flow than we do. So I think it's an opportunity really for us that with the confusion on accounting and as people start looking at cash to see how that's really kind of shining through for our company.

Got it. Maybe, Luc, I wanted to ask you, how important is the rollout of 5G to your business, which parts of your business are impacted? And are you going to see -- I imagine you will see more of it on the infrastructure side, but if there is a handset aspect of that business as well that affects you? Just give us a sense for how important 5G and the rollout there is in terms of impact to Rambus?

Yes. Great question. I'll start by saying that our main target segments continues to be data center and communication. And we continue to see increased demands for new applications, new architectures in that space. But we also see potential growth in other markets like 5G and others. In 5G, our main engagement, if you go back to the structure of our business with patent licensing, IP Cores and semiconductor products, our main impact to 5G is our ability to provide IP Cores to people who develop chips for 5G and -- for the 5G infrastructure. Why? Because in the communication links in a 5G infrastructure, you need high-speed SerDes, in particular, 32 gig SerDes, and we do have a 32 gig SerDes at the center of our offering for IP Cores. You also need embedded security. So we have customers that actually contract with us for either embedded security or SerDes technology to develop their own chip, that they will then sell to the 5G market. And as I said, the nice thing about IP Cores and security is that those technologies above and beyond the demand from data center and communication apply to other markets. 5G, we just talked about. AI is another interesting market. AI needs a lot of bandwidth, access to a lot of data. So people who develop chips for AI need high-speed interface, high bandwidth interface. They also need embedded semiconductor security to protect the data that they use for training and inference. Automotive is another market that is a nice market for the IP Core business. In an autonomous car, you have to read data from sensors in the car. So you need high speed, high bandwidth, real time, so you need high-speed interfaces, but we also need to make sure that the data you read from the sensors in your cars is authentic and protected just for safety reasons for the car. And that also calls on embedded security technologies that we provide. So the nice thing about having that portion of the business in the form of IP Cores is that it eventually addresses several markets above and beyond our core target markets of data center and communication.

Got it. Give us a sense of the competitive landscape, maybe in memory interfaces? And how do you differentiate in this environment?

Correct. So if your question relates to buffer chip as memory interfaces, there are 3 suppliers worldwide. We're one of them. You have Montage, a Chinese supplier and Renesas, Japanese. Renesas has entered that business last year by purchasing IDT. So they are the 3 players, and their 3 main customers, that's SK hynix, Samsung and Micron. So that's how the market looks like. Over time, we've gained market share in -- against Montage and IDT. There are several reasons for that. One is our semiconductor -- sorry, our design win footprint continues to improve as the ecosystem introduces new generation of DDR4 memories. So every time there's a new generation of DDR4, there's an opportunity to win share by winning designs. And every time, we've improved our footprint in terms of design wins, that translates into higher share. The second reason is that we've put a great emphasis on quality. And given the tightness of the supply chains, 3 suppliers, 3 customers and very, very important end customers, hyperscale companies, it's very important that the supply chains are very robust. And therefore, it's very important that the quality of products are flawless. So we've put a lot of emphasis on that. And then there's the macro environment. Again, we are the only U.S.-based company left in that space. We're investing in that space. So we're seen as a reliable, stable supplier. One of our competitors is from China. And the other one has moved from a Silicon Valley-based semiconductor company to being part of a larger Japanese conglomerate. So if you add all of these factors, this has improved -- it allows us to improve our competitive positioning in that space.

Right. No, I'm glad you brought that up. I was going to go there next at how you're positioning in the U.S. market and this greater need for depending more on domestic suppliers, how that impacts right to your business going forward. So I'm glad you touched on that. Maybe the next question. M&A has been an important aspect of your strategy. And Rahul also referred to the very strong cash flow generation. Talk to us about how you are evaluating the pipeline, do you see enough opportunities to consolidate the market further? Are there parts of the market where do you think it will be useful to expand via inorganic means?

Yes. That's a great question. So first of all, outside of any environment considerations, we have the ability to generate a lot of cash from our operations, as Rahul just explained, because of the structure of our business. So that's the first factor. The second factor is, we've been able to grow organically quite fast. If you look at our buffer chip business, that business has doubled in size 2 years in a row, approximately. So we're doing -- the team is doing quite a good job in growing organically. So we grow organically, and we have the ability to generate cash. So we can accelerate the growth through M&A, and that's what we want to continue to do. But we are very disciplined about the type of M&A we go after. We want to make sure that it makes strategic sense. We've just redefined the perimeter of our company. So we want to make sure that we address the markets and the customers, where we can have a leadership position, especially in embedded security and data center technologies. It has to make operational sense in the sense that we have to be able to integrate those companies fast and deliver to our customers fast. And we've proven with both Northwest Logic last year and Verimatrix that we've been able to do that. We've integrated the teams quite fast, and they're up and running and delivering as we speak. And financial -- and finally, has to make financial sense, of course. So if we look at that, those filters, then yes, we are willing to continue to accelerate our growth through M&A. Now if we look at the environment, I think COVID-19 is going to create an interesting environment. We talked earlier about the resilience of Rambus in that environment. Unfortunately, some companies are not going to be as resilient as we are. And that may create some opportunities for us to move fast and accelerate our M&A play.

Got it. And maybe just the final question to wrap things up. Is M&A the main kind of use of cash going forward for the company?

Hi, Vivek. Great question. And the way we look at it is that primary focus is continued organic investment. We continue to invest in our patent portfolio, which puts us in a very good position for the upcoming renewals and extensions. We also continue to invest in product programs. As we talked about over the course of the call, investing so that we're first at DDR5. I think that's a stair-step opportunity for us to really dramatically increase share, and also investing in programs in our silicon IP business as well. Then I think the second would be inorganic. As Luc mentioned, we certainly have the cash and the discipline. And I think the transactions that we did last year clearly identify us as a good buyer within our industry. So I think there certainly could be an opportunity for us to grow, both organically and inorganically. And those are the 2 primary uses of cash for us.

Great. With that, we are at the end of our time. Thank you so much, Luc and Rahul, for joining us this morning or afternoon, our time, and sharing your views about Rambus and the very interesting growth drivers. Thanks to the audience for joining us. And if you have any follow-up questions, please feel free to call. All right. But let's close the call there and thank Luc and Rahul one more time.

Thank you, Vivek. Thank you, everyone for your interest. Thank you.

Thank you for having us.

Okay. Take care.

Bye-bye.