Silicon IP Business is Reaching Maturity

Posted by 
in General
02 December 2009

Historically, most semiconductor companies were vertically integrated: they designed, fabricated, packaged and tested their semiconductors using internally developed software design tools, manufacturing processes and equipment.


However, as the cost and skills required for designing and manufacturing complex semiconductors increased, the semiconductor industry became disaggregated. Companies started to concentrate on one or more individual stages of the semiconductor development and production process. This disaggregation has fueled the growth of fabless semiconductor companies such as design tool and manufacturing equipment vendors, third-party foundries, packaging and testing companies and above all, Silicon Intellectual Property (SIP) companies that develop and license technology to others. Specialization has enabled greater development and manufacturing efficiency as well as an opportunity for IP companies to meet fundamental, industry-wide challenges. Throughout the semiconductor industry, IP companies and their technologies were adopted by supply chain companies that wanted to invest in technology infrastructure. This collaboration and investment cycle enabled semiconductor companies to bring new technologies to market faster and more cost-effectively, without making the investment themselves. In fact, the so-called semiconductor engineering productivity could not have kept up with the increasing silicon gate capacity without the IP, particularly complex IP that requires intensive investment. The IP sector became essential to the semiconductor supply chain as EDA tools did. System-on-a-Chip (SoC) designers started to rely on IP for building differentiated products in a timely fashion. As the semiconductor industry evolved, the IP business model has also gone through standard development cycles until it became an integral part of the silicon supply chain, like in any other industry.

Let’s go back in time to the year 2006-7 and read news topics and articles discussing the viability of the IP Business, especially those in EETimes.com. In articles such as "Opinion: Semi IP sector is a lost cause" and IP business model 'broken,' says exec , both writers and readers raised this question: did the IP model reach a dead end? Indeed, back in 2006, things probably looked that way. But rather than focusing on a single point in time, we should look into the long process to get a broad picture. Let’s use the product life cycle model to analyze the history of the IP industry. The life cycle model is a useful tool to determine the competitive force and changes that take place over time and analyze their effect on the evolution of the industry. The life cycle model is defined by five industry environment phases: Birth, Growth, Shakeout, Maturity and Decline as illustrated in the following graph.

 

The Birth phase, the beginning of the IP model, dates back to 1996 when DSPG and ARM started offering DSP core and RISC, respectively. (Who thought that you could simply give away your proprietary RTL code?) The whole business model was undefined: for example, it was unclear how long it would have taken to gain a significant amount of revenue beyond the standard license fee. The distribution channels had just started to develop and were also poorly designed. Availability of access to key technological know-how like the physical designs of RISC and DSP architectures created entry barriers. Competition at this phase was based not so much on price as on the ability to educate customers, and innovative companies gained the most opportunities to position themselves in the market.

Between  2000 and 2005, the IP industry enjoyed the Growth phase. Product demand began to take off which clearly showed characteristics generally seen in a growth industry. Demand expanded rapidly as many new customers entered the market, which became familiar with products using IP and the newly introduced IP licensing business model. The rapid growth in demand enabled the IP companies to expand their revenues and profits without taking the market share away from their competitors. It looked as if licensing was creating the next industry boom, and more vendors entered the business offering various IP, from RISC architectures including programmable solutions to physical IP libraries. For an IP company, an IPO became a reasonable and doable mission.
Following explosive growth, the IP industry had reached the point of excess capacity around 2006. At the same time, questions about the viability of the IP business had started to rise. The industry entered the Shakeout phase, and the growth rate slowed as demand saturated. Most of the remaining demand was limited to replacement orders because there were few potential first-time buyers left. Competition became intense and demand was no longer growing at historical rates.
To summarize, the consequences during the Shakeout phase were as follows: excess production capacity developed, demand growth slowed down and led to price wars as the industry matured, and creative sales ideas like royalty-free offering started flowing. Most of the inefficient companies were driven into bankruptcy, triggering market consolidations. Small IP vendors struggled and some of them were merged into larger IP vendors, EDA suppliers and Design Service companies.
Here is a (partial) list of the transactions that took place during this phase. StarCore was acquired by Agilent. 3DSP was completely wiped out. ZSP was acquired by Verisilicon and became part of their Design Service bundle (in the end, CEVA was the only pure DSP IP player left). Artisan was acquired by ARM, which was one of the surprising large-scale transactions. Impinj's non-volatile logic memory (NVM) intellectual property business and the CPU IP vendor ARC International were acquired by Virage Logic. Chip ID was bought by MIPS Technologies--who paid an enormous aggregate price of $147 million--and were eventually sold to Synopsys for $22 million (because of this transaction, Synopsys became the main force in the USB arena). Hantro, which developed video technologies for mobile phones, was purchased by a video compression software company, On2 Technologies.
Such consolidations created a debate whether an IP vendor must become a pure IP player or should bundle its IP with an EDA company which sells design service. In my opinion, both paths are legitimate; the decision strongly depends on the nature of the IP and its legacy. Indeed, at the later stage of market life cycle where we stand today, starting a pure IP player business with an ambition to grow it into a $100M company is too optimistic. Also, all of the previously mentioned symptoms are common parameters of a mature industry. The shakeout stage in the semiconductor industry is creating a new wave of opportunities for IP companies. The semiconductor crisis will accelerate disaggregation – where unsustainable indebtedness will drive the separation of the in-house IP development teams from their mother’s aprons. Virage Logic’s acquisition of NXP’s IP portfolio & R&D resource is an insight into the future of the new way to scale the IP business.

The Shakeout stage eventually gave way to the Mature stage in which the market was totally saturated, demand was limited to replacement orders and the growth rate was lower than or equal to zero. By the time the silicon industry had matured, the surviving companies were limited to those who had brand loyalty and low cost business operation. Both factors constituted a significant entry barrier that greatly diminished the threat from potential competitors. Following the shakeout phase, the IP industry consolidated and turned into oligopoly--which is what we are facing today.
Eventually every industry enters the Decline stage, in which growth rates become negative for a variety of reasons such as technical substitution. However, declining market environment is not an inevitable phase, the IP sector is an essential link in the semiconductor supply chin and strong, established IP vendors can and should continue their development effort which ultimately leads to a new growing cycle. That means developing new solutions and addressing emerging vertical markets
For example, in the hard IP sector, betting on the winning 40nm/28nm technology and developing vital IP will create a new wave of growth. Because complex technologies not only require large investment but also create a high barrier for new and small vendors to enter and help established vendors stay in better position. In the processor (CPU and DSP) sector, IP vendors should look into the 4G area that request a new set of processing power--a revolution as opposed to the gradual evolution that occurred when moving from 2G to 3.5G. In the consumer market sector, IP vendors should watch out for leading silicon suppliers in set-top boxes, video game consoles and digital TVs. The next-generation video game consoles and handheld devices will emerge in 2010-2012. Video game consoles will increasingly become more complex by having Internet connectivity and high-end computing, and will control a larger share of the digital living room traditionally served by the Digital TV (DTV). However, eventually the DTV will also introduce a new set of challenges by integrating a stronger CPU and connectivity to the internet.

Last modified on Thursday, 14 January 2010 13:33