Nicholas Hawkins: Hans, this is Nick. In terms of your other question, which is about sort of concentration, the concentration of auto chips per vehicle, which is I think the other part which is that we think longer term is up to sort of 25 plus, that is unchanged. And really isn’t impacted as to whether the OEMs do their own designs or whether they look for somebody else to do their designs for them.
Hans Mosesmann: Okay. That’s helpful. And then one last question. If we look at outside of China, how is that, the non-China business, doing in terms of licensing engagements? Or are there some delays there as well?
Karel Janac: Obviously, people are looking at their roadmaps. That’s — I would say people are looking at what projects they will do. At the same time, what we’re seeing is that some of the larger companies are now sort of looking at, okay, do we want to build the next-generation system IP for our products or do we want to outsource that? And we’re having I would say a significant amount of success in being selected for some of these projects where before they would do everything internally. The other sort of issue is that there’s also a capital difficulty in the smaller U.S. companies, but the larger companies are unaffected. I would sort of note that the number of design starts in Q3, which was 2022, is basically the same as we had in second quarter. While there is some shifts in the non-China market, I think overall the effects are I would say neutral.
Nicholas Hawkins: Yes. Essentially, it’s a pivot away from the start-up community to the larger guys. As Charlie said, the sort a number of new customers is a lot less. We used to get most of our new customers actually from China. But the, as Charlie said, the number of design starts is the same. But one of the other dynamics that was kind of following on from your question, and is really driving the reduction in guide on ACV for the fourth quarter, is that customers are generally having a look at — they’re aligning their product roadmaps to their expenditure cycles. And in some cases, that’s pushing out design or engagements into Q1 from Q4. And same kind of thing as we saw at the end of ’22 if you remember. There’s a number of different vectors going on.
And in addition, there is a tendency among some of the majors, not many, but there’s a tendency among some to take longer to do their designs. And that actually that has a direct impact on ACV and revenue intensity. It just lasts for longer. It’s — we’re okay with that because it demonstrates the stickiness and the importance of our technology to the customer if they’re prepared to engage for longer. But in terms of ACV, obviously annual, the A word, means the TCV is divided over a longer period.
Operator: Your next question comes from Kevin Garrigan, WestPark Capital.
Kevin Garrigan: Thanks for taking my question. Your 22 design starts. I know you had noted that 12 were for AI technologies. Can you give us some color on what markets these design starts were in? And to Hans’ question, you had noted automotive was still robust. Any kind of markets that were weaker that surprised you?
Karel Janac: Outside of China, not really. The — there is a huge amount of investment in generative AI. As we discussed earlier, the query costs are quite high with the existing GPU technologies, and so people are investing extensively in essentially ASICs that are lower power and lower query costs for generative AI. That’s a big tailwind. I think one of those designs that we’ve talked about is the NeuReality design in Israel, which is a data center design. There’s a lot of generative AI inference work being done in the data center. But we eventually see generative AI being expanded to the edge and to the endpoints, even into things like portable devices like smartphones and things like this. Generative AI is a strong, I would say strong tailwind for new designs. And I think we said on the earnings call that about half of the designs of the 22 were at least this quarter were related to machine learning.
Nicholas Hawkins: A bit over half. A bit more than half. I mean, to your — if you looked at, I mean here’s the complexity, is obviously that the AI/ML is actually horizontal. We kind of get a bit mixed up in terms of horizontal or vertical sometimes. But if you look at the verticals where we saw the most intensity of design starts in Q3, the enterprise compute actually, which is obviously where a lot of these designs end up popping up in the vertical sense, that was the second highest quarter on record in terms of design starts. That’s a good indication. The second was actually automotive, but that’s fairly typical and there’s a lot of machine learning applications within automotive anyway.
Kevin Garrigan: Okay. Got it. That’s very helpful. And then just as a quick follow-up, can you, at a high level, can you kind of give us a sense of your royalty rates and then the opportunities that you kind of see to increase these rates in the future? And how do you kind of go about having those conversations to basically increase the rates?
Karel Janac: Yes, I’ll take that one, Kevin. Directionally, royalties are always upwards. We don’t take lower royalty deals. The growth is, any increase in royalty rates, can only really follow increased functionality. You don’t get a sort of increased royalty rate just because the customer likes you. We have to actually give them something in terms of functionality. And so over time, you’ll see more higher rate of royalties for the newer products. And if you remember, automotive is actually — automotive is the richest, or certainly a lot richer than say consumer. Or the old consumer communications like cellphones, smartphones, and automotive is where we’re having probably the biggest success and it now represents about 55%, 60% of the total royalty stream followed interestingly by communications and industrial. A lot of use in industrial.
