Report from the 2015 Silicon Valley IP Users Conference

One of the most interesting events I attended last year was the 2014 Silicon Valley IP Users Conference, organized and presented by IPextreme and their Constellations program partners. It was a wonderfully well-organized day, with excellent speakers in the fun environment of San Jose’s Winchester Mystery House. On Tuesday of this week, I attended the 2015 version of the conference and once again was impressed by both the technical content and the networking opportunities.

This year we were nestled in the foothills of Los Gatos at the historic Testarossa Winery, coincidentally on the same day that Manresa Restaurant just down the street was awarded its third Michelin star. With a wine tasting after the presentations, we were all in a celebratory mood. I was most intrigued by the panels, so I’d like to devote today’s post to a summary of some of the more interesting points I heard and what they might mean for the semiconductor industry, the EDA industry, and Breker.

The morning panel discussed the “State of the Semiconductor IP Market: Where are we? Where are we going?” and was moderated by well-known journalist Ed Sperling of SemiconductorEngineering. The panelists were from camera/display IP vendor Apical, ARM, EDA promotions firm EdaBuzz, and eSilicon. They provided a wide perspective of creating IP, selling IP, evaluating IP, and using IP in chip designs.

The most provocative question that Ed asked the panelists was whether the IP industry is mature, or still an “awkward teenager.” The consensus tended toward the latter, with comments that this is in many ways a good thing. Lack of maturity also means lack of calcification. IP companies can still be flexible in their business models, there is plenty of room for innovation by small IP suppliers, and the IP industry is growing faster than traditional EDA.

There was some discussion about consolidation, which turned out to be a hot topic continuing informally during lunch. There has been some consolidation in the IP business itself, noting that Cadence as well as Synopsys has now made numerous acquisitions and major investments in this area. But consolidation among customers seemed to be the bigger concern. I have blogged before about consolidation in the semiconductor industry, and its impact on EDA. Any time that two competing projects are combined into one after a merger, there is potential loss of revenue for EDA.

Assuming that the single product sells about the same as two competing products, one could argue that the impact on IP royalties would be minimal. But of course up-front licensing costs could be reduced significantly, and even the royalty rate might be less given that a single vendor would be shipping higher volumes. Both panels hit on the themes of high quality IP and excellent customer services as keys to a successful executing of the IP business model despite pricing pressures.

The afternoon panel asked “IP: Product or Service?” and was moderated by Rich Goldman from Silicon Catalyst, a new incubator for semiconductor startups. The panelists represented Freescale, IDT, Lattice Semiconductor, IP vendor NGCodec, and Open-Silicon. The seemingly simple question posed by the panel title had many nuanced answers from the diverse group of panelists. It was clear that the IP model spans a continuum from pure consulting to fully standardized products.

Several of the vendors stated that customization was an essential component to their business, to the extent that some have never shipped exactly the same product twice. Tweaks for process, power, area, performance, and reliability are often requested by customers and may be the single biggest factor in choosing an IP supplier. This was said to sometimes be the case even when an in-house IP development team is competing for the design-in. The idea of an IP product with a small associated service component seems to be common.

Panelists were quick to point out that this business model is very different from a pure consulting services play, which is much harder to grow and manage. In fact, many IP companies started as services providers but transformed as they identified design blocks that could be sold as products to multiple customers. I asked how configurable IP,whether in the hands of the supplier or the customer, fit into the model. The response was that predefined configuration is a product feature and not the same as per-customer customization.

My final topic, and the one that most intrigued me last year, is how the Internet of Things (IoT) will (or won’t) change the semiconductor, IP, and EDA industries. At the 2014 event I heard the speakers say that there will likely be many new chip designs for IoT, but there seemed to be more restraint this year. A couple of the panelists mentioned that IoT applications will try to use off-the-shelf solutions whenever possible, with more custom chips developed over time. So I’m still bullish on IoT for Breker and EDA in general, but perhaps it will take some time for us to see the full effects.

Thank you to IPextreme for hosting a most enjoyable event and for arranging provocative speakers and panelists.

Tom A.

The truth is out there … sometimes it’s in a blog.

Tags: Accellera, Breker, DV, ESL, functional verification, IoT, IP, IPextreme, portable stimulus, SoC, SoC verification, uvm, VIP

This entry was posted on Thursday, October 22nd, 2015 at 5:15 pm. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.