Industry Scorecard For 2014

First of two parts: It’s easy to make predictions, but the best in the industry have a track record of getting things spot on. How well did those 2014 predictions turn out and who was closest to the mark?


At the end of last year, Semiconductor Engineering asked the industry about the developments they expected to see in 2014. If you care to refresh your memory, they were categorized under markets, semiconductors and development tools. Now it is time to look back and see how accurate those predictions were and where they fell short.

The obvious trend, at the beginning of the year, was for continued development of portable devices. “The latest industry analyst reports bear out this trend, but there seems to be some slowing,” says Graham Bell, vice president of marketing at Real Intent. “As smartphones get more powerful and larger, they seem to be eating into the traditional tablet market in 2014.”


Several predictions centered on the incorporation of additional sensors and nine-axis inertial measurement units into these devices and sensor fusion. One application of that, predicted by Bernard Murphy, chief technology officer at Atrenta, was for the ability to navigate shopping malls. Murphy rates this as an A for technology and D- for adoption. “Apple introduced iBeacon using Bluetooth LE for local navigation. Adoption is more of a marketing issue than a technology limitation. The technology is already used in a number of major league baseball parks and by some stores. It will really take off when chains start to recognize advantages in steering customers and promoting products through this channel.”

As the mobile market starts to grow into new categories such as “wearables, Internet of Things (IoT) and other devices that all have an IP address, a sensor, transducer, memory, some processing power and a wireless connection, silicon usage will skyrocket,” predicted Mike Gianfagna, vice president of marketing at eSilicon. “The products are starting to show up,” he says, “but the general market is still nascent in terms of well-defined players, platforms and protocols. The movement is undeniable. It will just take a bit longer.”

We are, however, seeing the first signs of fab shortages, so it would appear as if Gianfagna was not far off the mark. That may grow worse in 2015. Wearables, while there were notable introductions, may not be the new market category given that the levels of excitement already seem to be dropping for many of them.

Another proponent for the IoT was Joseph Sawicki, vice president and general manager for the Design-to-Silicon division of Mentor Graphics. “The IoT continues to top the trend charts, but like all new technology directions it will take longer to become ubiquitous than the evangelists project,” he says. But there are areas that he points out are getting additional attention. “MEMS is getting hotter as more and more sensors are being integrated into designs for everything from cell phones to automobiles.”

Mentor says it is working with MEMS component designers, foundries and system manufacturers to make it easier and cheaper to integrate MEMS elements with supporting electronics.

And on the subject of automotive, several expected further convergence between this and the mobile market. “The automotive market continues to be a big driver for semiconductor products with the global market for OEM electronics growing by 12% in 2014,” says Aveek Sarkar, vice president of product engineering and support at Ansys-Apache. This is more than twice the growth rate for the industry as a whole. “As cars start to incorporate connectivity, the bill of materials impact from electronics parts will continue to become more dominant, but margin pressure on semiconductor providers will increase.”

Automotive may have turned into something larger than the predictions. “It’s not just the automotive systems themselves that are on the increase, but the maturity of the core technology that drives these systems,” points out David Kelf, vice president of marketing for OneSpin Solutions. “For example, the verification of safety critical designs, including their adherence to standards such as ISO 26262, has become a major topic in the industry, and this technology is central in many automotive electronic systems.”


2014 has been an interesting year for security. Many banks and retailers reported breaches, and the revelation of NSA actions has brought the topic to a heightened awareness. Murphy predicted that “we would see more focus on, and the maturing of, biometric identification technologies, driven by security needs for smart payment on phones, and an increase in smart-home applications.” While the smart-home market has yet to make much progress due to the fragmentation of solutions, biometrics has seen significant advances. “At a purely technology level, both Apple and Samsung introduced fingerprint scanning as a mechanism to authenticate access to phones and facial recognition is becoming available.” Ultimately, privacy is a policy issue more than a technology issue and so it appears that technology is no longer holding up progress.

Views were split as to the impact that these market trends would have on semiconductor companies in general, some predicting that there would be further consolidation while others believed we would see many new entrants coming into the market. “The mobile market has indeed converged on few big players,” says Sarkar. “Industry leaders continue to dominate the market especially for high-end smartphones with Qualcomm having a 68% share of the baseband market. The market for low-cost smartphones continues to be lucrative but increasingly taken by design companies in Asia, notably in Taiwan and in China with MediaTek (15%) and Spreadtrum (5%) leading the charge.”

Semiconductor predictions were clustered into three primary areas, the migration to new process nodes, adoption of FinFETs and increased utilization of 2.5D and 3D integration.

2014 was the year when people started to talk about 28nm being a popular node for a considerable period, especially given that smaller nodes appear to have higher costs per transistor. “28nm will remain a dominant node for chipsets going into cost-effective smartphones” said John Koeter, vice president of marketing for IP and prototyping at Synopsys. A year later he adds “There are many proof points, including what we know our customers are building. Probably the best proof point is that TSMC created a process – 28HPC – specifically designed to address this prediction.”

Murphy noted that “Wall Street has come to see all innovation in the semiconductor industry in terms of feature size, but Moore-stress and rapidly increasing cost with diminishing returns may bring a welcome shift to focus on content.” Murphy predicted that it would result in more design focus, a prediction for which he gave himself an E. “Tthere is not much obvious compensation through focus on more clever design techniques.” This is a baffling trend and if it persists will cause many markets to become commoditized.

Mentor’s Sawicki was undeterred at the beginning of the year saying, “Fabless customers are just beginning to implement their first test chip tape-outs for 16nm/14nm, and 2014 will see most of the 20nm early-adopter customers also preparing their first 16nm/14nm test chips.” Sawicki supports the claim saying “We’ve seen a big jump in foundry revenue at 20nm in the second half of 2014. For the foundry leaders, 20nm is now in the range of 10% of their total revenue, representing approximately a 10X increase year-over-year increase! Currently, there is some node-over-node increase in wafer cost for a design start at 16/14nm, but the percentage increase from 20nm to 16/14nm is reported to be less than that of 28nm to 20nm.” Sawicki also provided a status update on the major foundries saying “Half of the IC Insight’s top 20 will have had a 16/14nm tape out by the end of 2014.”

In the area of memory, Real intent’s Bell had expected strong expansion of the DRAM and NAND flash memory markets, duplicating the 35% and 28% respectively gains for 2013. However, “the forecast growth for DRAM and NAND Flash memory was too aggressive. Growth rates were more normative in the range of 16% to 22%.” Bell notes, “We did see the beginnings of the leap to 3D at IM Flash and SK Hynix according to IC Insights, and Samsung is already pushing from 24 to 32 layers in 2014.”


2.5D packaging technology will start to see the light of day in more applications,” predicted Gianfagna. He expected to see exciting innovations in interposer design. He is still a believer. “We are working on projects in this area, and so are a lot of other folks. It hasn’t reached the tipping point yet, which has some people frustrated.”

Nobody else dared to comment on the slow uptake of this technology. In a recent article titled “The Internet of Cores,” Murphy noted that “Interconnect crossing between die in 2.5D / 3D is still a research project in my view. It’s possible in principle, but you have to ask, ‘What is the economic motivation for splitting digital logic between multiple die?'”

Steve Schulz, president and CEO at Si2, points out, “One of the fundamental characteristics of the IoT is sensors, and these are not ideal to be manufactured in 10 or 14nm. Neither are the types of memory they will use. If it can be done on separate dies, from different foundries at the most appropriate nodes and processes, you will be able to do more cost shopping. And then it comes down to the cost of integrating the cheaper dies. That is where the economic clout will solve the problem.” Sawicki had said that “We’ll need 3D packaging implementations that are an order of magnitude cheaper than current offerings,” and we have yet to make much progress in that area. Maybe this will continue to be a prediction for 2015.

The recent International Electron Device Meeting (IEDM) allowed the fabs to show off recent advances with finFETs. TSMC is now talking about the second generation of its 16nm finFET process, boasting faster switching speeds and reduced power. Not to be outdone, Intel has its 14nm finFET following on from the 22nm transistor, which the company claims will maintain the 0.7 scaling.

Koeter noted something that he had not predicted, namely “the acceleration of 10nm finFET plans with major foundries significantly accelerating their original 10nm process plans by as much as one year in some cases.” We will have to wait and see if the rest of the supply chain is willing to invest to bring these plans forward.

Raik Brinkmann, president and CEO of OneSpin Solutions, had said, “It is worth watching the opening of the Intel facilities to other companies and the new platforms introduced as a result.” OneSpin’s Kelf responded that “the release of the new Altera Stratix 10 device based on Intel’s 14nm process is a significant step forward, but we thought this trend would move faster.”