Executive Insight: Gideon Wertheizer

CEVA’s CEO talks about why the IoT is so important, the ongoing shift from fabless to OEM customers, and where the real competition is.


SE: From your standpoint, what’s the next big thing?

Wertheizer: The industry was driven in the past few years by the structure the smartphone created. It looks like this area is about to grow. What’s changing is the integration of the smartphone with other applications. The smartphone is now a hub of entertainment and productivity with many devices connecting directly or indirectly to it. So, I’m not too excited about the smart TV market. But the Internet of Things is very interesting from this connectivity perspective.

SE: Why is the IoT so interesting now versus proposals 20 years ago for similar things?

Wertheizer: The smartphone created the market for connected devices. Everything can be connected. Every place can be connected. What the smartphone created is commoditization in certain areas such as, connectivity, application processors and sensors. Commoditization is how the smartphone evolved. Through this commoditization, these technologies are now making their way into new and innovative devices and use cases. People are in a place where they can be more creative with technology. The value it created has generated funding for innovation to address a lot of specific technology problems.

SE: Where does CEVA fit into these new markets as the industry evolves?

Wertheizer: DSPs are a good way for us to tap new markets. One way is through verticalization. We are taking certain markets and applications and specializing in them. Previously, we did that in the baseband market and had great success. Now we are doing embedded vision. And also we are now doing connectivity. Connectivity, however, is not the generalized focus area of DSPs. We are specializing with the architecture and the tools to enter this market with very robust specialized IPs.

SE: What’s your vision of how this unfolds? What are the first markets you see taking off?

Wertheizer: The IoT is an umbrella of markets with many different segments and opportunities. Right now, we are looking at what we are going to do and which of these IoT segments we want to address. The first one, and right now the biggest, is communication. It’s the baseband processing we’re doing for handsets and infrastructure. Our strategy is to look for big markets that have anomalies. Handsets are a good example. Back in 2008 we thought there were some anomalies in this market. In the handset, it became clear to us that the DSP was a difficult technology to develop and maintain for baseband, together with all the software that is written for it. It made sense that as the market grew, OEMs would benefit from a multisource supplier strategy in baseband chipsets and that a common DSP used by these suppliers would benefit the OEMs. Similarly, now in the infrastructure space, 85% of the market is dominated by TI DSPs.

As the base-station market is evolving we see opportunities for LTE, LTE-Advanced, Wi-Fi and multi-radio. Another market is imaging and vision. Here, we’re taking a pioneering approach, where we are designing everything from the processor to the algorithms and the software. Imaging and vision technology is applicable to smartphones, tablets, and industrial markets and a lot more. As we’re addressing this market, we approach customers and explain our philosophy and our approach. Before we came into this market people did things with ARM or TI or GPUs. GPUs are not for vision. They are for graphics. That’s the reason we took a pioneering approach. There was a gap in the market for a dedicated imaging and vision processor solution that was not being addressed with existing technologies.

SE: What was your thinking behind the recent acquisition of RivieraWaves?

Wertheizer: RivieraWaves is the leader in Wi-Fi and Bluetooth IP in the industry. It is a self-contained unit that has great IP, a strong customer base and is addressing an area of the market that we believe has enormous potential. Connectivity is an incremental market for us and is set to be one of the biggest volume markets with the emergence of the IoT. Also, Wi-Fi and Bluetooth are complicated technologies. So there are high barriers to entry. Not everyone can come and do the connectivity.

SE: This is very closely tied to your communications business, right?

Wertheizer: Yes, but the approach is different. When you target base stations, you have to provide the DSP platform and your customers bring a lot of the expertise and develop the system and the software. However, when you address the Wi-Fi or Bluetooth markets, your customers expect you to provide a complete solution. The connectivity is a critical block to them, but it’s very standardized so there is not a lot of customization that can differentiate them. They want a block to provide the full functionality for this part of their chip design.

SE: But you can make each much more efficient by understanding both sides of that, correct?

Wertheizer: Yes.

SE: What other markets do you target?

Wertheizer: Audio, which is a huge market. Like connectivity, the audio market is highly commoditized with a focus on low power consumption. This is why DSPs were introduced. There are competitors like Tensilica and companies in China, and many in-house DSPs in use today. We are gaining a lot of success in the audio market these days, as low power for ‘always-on’ applications like voice activation requires very efficient DSPs, and we are addressing exactly that space.

SE: There’s a big transition under way now at the leading edge of design that will force a transition into 2.5D and 3D IC. What does that mean for CEVA?

Wertheizer: There are things we need to take into consideration. We are doing design work for connectivity to memory and whether you support this communication, but we are not exposed to the outside.

SE: In order to make the IoT work effectively, we need to improve connectivity across large distances. What needs to be done to make it work?

Wertheizer: We see improvements in traditionally ‘short-range’ connectivity technologies such like Wi-Fi and Bluetooth being key here. That was fundamental to our acquisition of RivieraWaves. IoT devices cannot afford the power overhead of LTE-type communications, and in most cases they really don’t need that type of connection bandwidth. Most important are improvements in power management here, especially to enable more advanced algorithms for things like video and audio processing, not necessarily extending the range or distance of connectivity.

SE: Who do you see as your biggest competitor?

Wertheizer: Really it is people doing their own DSPs in house, the biggest example being Qualcomm. We really don’t compete head to head too often with other IP suppliers in the communications space, and in fact have a very large percentage of the market for licensable DSP cores, especially in mobiles and in Asia specifically. In our other target markets—vision, audio and connectivity—there is some fragmented competition from IP suppliers. But we are seeing more and more that customers want a one-stop solution for those types of things that we are the only ones providing.

SE: Some large OEMs are starting to develop their own chips. How does that affect CEVA?

Wertheizer: The OEMs still need the same type of functionality and performance that the fabless guys needs. We are seeing more adoption of our technology by OEMs. When we talk about the customers that we license our technology to these days, probably now our market is 60% fabless, 40% OEM, and that’s changing in favor of the OEMs. In the last several months we have won quite a few deals with major OEMs for baseband, for infrastructure, vision and audio.

SE: You mention verticalization in markets, but you’re going in with horizontal technologies. This is not really a market such as automotive, though, right? It’s more of a narrow horizontal play.

Wertheizer: By verticalization we mean an application area, like vision, or imaging or audio. Those technologies apply horizontally across many markets such as mobile, automotive and industrial. Those markets may have specific needs but the general platform for a specific functionality is similar. So in that sense, yes, we have horizontal solutions that be used by many different product applications and use cases. But we are vertically focused in terms of the technologies we offer.

Leave a Reply

(Note: This name will be displayed publicly)