By: George M. Lee
Delayed rollouts of FinFET-enabled chip production lines have seemingly become the norm. However, upcoming devices from major electronic device manufacturers suggest that this technology may finally be ready to step into prime time. Investors will want to keep a close eye on this promising technology.
First conceptualized by three UC Berkeley researchers as a double-gate transistor wrapped with a thin silicon conducting channel, FinFET represents a major technological advance. This fin-like gate structure allows better electronic control and helps reduce current leakage and other short-channel effects that become more pronounced as the size of transistors are miniaturized.
FinFET technology was heralded as the latest flag bearer in the unending march of Moore’s law, a widely publicized observation that the processing power of semiconductors doubles roughly every two years. Today, the term FinFET refers to any multigate fin-shaped field effect transistor. Unlike traditional planar transistors, these three-dimensional structures have multiple conducting channels, due to their vertical orientations, and can also be packed tighter together.
Intel (INTC, Financial) was the first to successfully commercialize this concept with its 22nm Tri-Gate transistors in 2011. Since then, other semiconductor fabrication plants have also tried to implement FinFET technology, but major obstacles have prevented industry-wide rollout efforts. To wit, in the first quarter of 2015, Samsung announced the start of mass production for its Exynos 7420 chip, which powers its recently released flagship Galaxy S6 smartphone. This 14nm FinFET-based chip represents the first time that a company other than Intel has successfully incorporated FinFET technology in a commercial product.
Even Intel, the undisputed industry leader, has encountered some difficulties since its initial foray at the 22nm production node. With much fanfare, its 14nm Broadwell-based 5th generation processors were first showcased in the third quarter of 2013 and were supposed to have started shipping by the end of that year. Yield issues, however, forced the company to delay its actual rollout for nearly a year. As of early 2015, Broadwell-based products are still conspicuously missing from many store shelves. Many major electronic device manufacturers have decided to skip this round of upgrades. With Intel’s highly anticipated 6th generation Skylake chips scheduled to be released in the second half of 2015, many customers are sticking with mature 4th generation Haswell processors for the time being.
At comparable yields, smaller production nodes typically improve price/performance ratios, which benefits all parties involved. Until an alternative technology is invented, FinFETs will likely be necessary at sub-20nm nodes. A steep learning curve in reaching acceptable yields, however, has led to industrywide uncertainty that has devastated companies that support the semiconductors industry, such as Ultratech (UTEK, Financial).
Recent developments, however, suggest greener pastures ahead, though competition will likely remain fierce. Taiwan Semiconductor (TSM) manufactures the bulk of Apple’s (AAPL, Financial) planar 20nm A8 chips, which power the iPhone 6/6 Plus. TSMC is also currently rolling out its 16nm FinFET-based manufacturing capabilities, albeit after a year-long delay. For its next-generation A9 chip, Apple has reportedly decided to rely on Samsung and Global Foundries in addition to TSMC. Meanwhile, Qualcomm (QCOM, Financial) also plans to dip its foot in the water with the fall 2015 release of its Snapdragon 820 high-end processor, which will also utilize Samsung’s 14nm FinFETs.
The battle at the 14nm/16nm node is underway and early results favor Samsung’s alliance with Global Foundries. TSMC, however, has ambitious plans to take back the performance crown with a new $16 billion fabrication plant under construction, and is currently eyeing 10nm for the end of 2016. With 7nm and 5nm nodes slated for 2018 and 2010 respectively, Moore’s law appears to be holding for the time being.
Featuring superior current leakage performance, 3D FinFET transistors will allow foundries to continue to shrink their die fabrication processes and push up against the physical limits of nature. FinFET adoption has been slower than expected, and many twists and turns have occurred along the way. The major foundries, however, have undoubtedly gained valuable lessons that should help enable smoother transitions in the future.
The semiconductors industry has had to deal with a difficult operating environment in recent years, and some stocks in this sector are trading at historically low valuations. Demand for advanced consumer electronics should grow at a healthy pace over the next 3- to 5- years, and ought to provide solid growth opportunities. Many estimates for the global semiconductors market suggest a compound annual growth rate of roughly 3% through 2020.
That said, concerns about a long-term slowdown in innovation have some merit. High technological and financial barriers to entry make it difficult for innovative new upstarts to compete with the handful of established players in this consolidated space. Also, as the hardware capabilities of electronic devices outstrip software needs, device manufacturers may find it tempting to stretch out the replacement cycles of older production processes.
For now, however, investors should not ignore FinFET’s long-anticipated entrance to the main stage. Indeed, near-term sentiments of many of the companies that deal with this technology have improved greatly in recent months. Barring global macroeconomic disruptions, these gains will likely be sustainable as FinFET matures over the next couple of generations. For more information about the near- and long-term prospects of the companies mentioned above, subscribers will want to consult our full-page reports in The Value Line Investment Survey.
