uploads/2016/02/Series-65-A51.png

Why Intel Is Slowing Down Development of Smaller Nodes

By

Updated

Tick-tock and what

In the earlier part of the series, we saw that Intel’s (INTC) technology development cycle has slowed from two years to two and a half years. With the delay in the 10-nm (nanometer) technology, this cycle is likely to extend to three years. While this would increase the equipment life, reduce depreciation, and reduce capital requirement, Intel may close the technology advantage it has over its competitors.

Article continues below advertisement

In the previous article, we saw that every new tick brings with it two architectures in two years. However, this has changed as the cycle has extended to two and a half years and three architectures are being developed for every node. The change from two architectures to three was first seen in the 22-nm node, followed by 14-nm and now 10-nm. However, if the company intends to maintain its tock cycle to two years and still maintain the three architecture count, it will have to introduce a new processor every ten months, which is unlikely.

Intel’s delays are not Intel’s denials

According to a news leak reported by The Motley Fool, Intel plans to launch three architectures—Cannonlake, Ice Lake, and Tiger Lake—for the 10-nm node. The three architectures are expected to be launched in 2H17, 2H18, and 2H19, respectively. This means a 7-nm node is unlikely before 2H20.

However, the company’s management stressed that it’s trying to get back onto the two-year cycle after the 10-nm, which means it plans to return to the two architectures per node model. If this holds true, Intel is expected to launch the 7-nm in 2H20 and the 5-nm in 2H22.

However, the return to the two-year cycle seems unlikely. If the 7-nm technology is delayed to 2020 by the industry’s dominant player and Taiwan Semiconductor Manufacturing (TSM) eyeing a 10-nm launch in 2H18, the latter’s technology would be interesting to see.

Article continues below advertisement

At the fiscal 4Q15 earnings call, Intel chief executive officer Brian Krzanich stated that having a node for a longer term would not impact the company. He said, “This is more about the amount of reuse, and the efficiency and speed at which we can do the conversion. And so, both of those numbers have improved.” The company aims to improve the reuse and transition speed, instead of only focusing on node shrink.

Delay could cost Intel technology advantage

At present, Intel enjoys a dominant position in the PC (personal computer) and server market. However, as the product cycle widens, more competitors would be able to compete with the behemoth. Thus, Intel would lose its technology advantage.

Advanced Micro Devices (AMD) plans to launch its Zen architecture produced on Samsung’s (SSNLF) 14-nm FinFET (fin-shaped field effect transistor) technology by fiscal 4Q16. It’s believed that it will compete with Intel’s processors as both would be 14-nm technology.

We’ll look at the competitive landscape in detail in the next part of the series. The PowerShares QQQ ETF (QQQ) has more than 8% exposure in the semiconductor market, with a 2.9% holding in INTC.

Advertisement

More From Market Realist