5 Takeaways From Semicon

What were the hot topics at the annual event?


As usual, the recent Semicon West trade show was a busy, if not an overwhelming, event.

The event, which took place in San Francisco in early July, featured presentations on the usual subjects in the semiconductor and IC-equipment sectors. There were sessions on 200mm, next-generation processes, transistors, lithography, MEMS and many others.

In no particular order, here are my five takeaways from Semicon:

Flat forecast and China
The semiconductor equipment market is expected to be relatively flat in 2016. In total, the equipment market is expected to reach $36.94 billion in 2016, compared to $36.52 billion in 2015, according to SEMI.

It’s a rather sober forecast. Capital spending in the DRAM sector is weak, while foundry and NAND are steady.

China is one of the few bright spots, however. Of the $36.94 billion figure, China’s equipment market is projected to hit $6.41 billion in 2016, up from $4.9 billion in 2015, according to SEMI. That compares to just $3.38 billion in 2013, according to SEMI.

In 2016, China will pass Japan, Korea and North America in the arena. China will become the world’s second largest equipment market, only behind Taiwan ($9.46 billion), according to the trade association.

It’s boom times for equipment vendors in China. A growing number of domestic and multinational companies have fabs, or are building plants, in China. For example, Samsung’s fab in China is making 3D NAND. The fab is at or near capacity. In addition, Intel is converting its fab in China from 65nm chipsets to 3D NAND.

On top of that, several foundries are also building fabs in China. And multinational OSATs continue to expand in that nation.

As previously reported, the China government is pouring billions of dollars into its semiconductor industry. China must import most of its chips and the government wants to close the trade gap. The worrisome part is that China’s economy is slowing. Perhaps the bubble will burst, if it hasn’t already.

M&A to heat up again?
In 2015, there was a frenetic amount of merger and acquisition activity in the semiconductor and equipment industries.

The M&A activity cooled off in the first half of 2016. But one deal that took place just before Semicon is a reminder that merger mania could heat up again.

In June, ASML Holding announced a plan to acquire Taiwan’s Hermes Microvision, the world’s largest supplier of e-beam inspection systems, for $3.1 billion.

And let’s not forget the big blockbuster in July, when Japan’s Softbank said it will acquire ARM for $32.34 billion.

To be sure, the chip and fab tool shakeout is far from over. In fact, there are a number of small cap companies in play in the IC-equipment sector, according to Stifel, Nicolaus & Co.

Semiconductor equipment remains a tough business. It takes vast R&D resources to develop new tools. But there are only a few chipmakers that buy tools at the leading edge.

It’s a somewhat brighter picture in the “More than Moore” segment. But even then, fab tool vendors also face stiff competition in LEDs, MEMS and RF. Advanced packaging is also a tough equipment market.

Older wafer sizes never die
Surprisingly, though, demand for 200mm fab capacity and associated tools remains relatively robust.

Not every chip requires a leading-edge process. In fact, it’s the opposite. Analog, MCUs, MEMS, RF and others don’t require the latest and greatest processes.

Besides 200mm, there is still a wave of activity for fabs at 6-inch and below. Much of the action involves the compound semiconductor market. For example, the vast majority of devices based on gallium nitride (GaN) are produced at 3- and 4-inch wafers. Qorvo is moving RF GaN to 6-inch wafers.

Chip scaling slows and speeds up
As in past events, Semicon had several sessions on chip scaling. The topic of conversion generally revolved around the same themes–What will happen at 5nm and beyond? What will the chip architectures look like in the CMOS and post-CMOS era? Will EUV ever happen? And so on.

Based on the events at Semicon and outside the event, the nodes are extending beyond the traditional two-year cadence. It’s becoming a 2.5-year cadence. Maybe longer.

Even so, the foundry vendors don’t appear to be slowing down. Samsung plans to move into 10nm production by year’s end, as previously announced.

TSMC continues to step on the gas. During a recent conference call, TSMC reconfirmed that it’s on track with 10nm and 7nm.

TSMC said it will start generating sales at 10nm starting in the first quarter of 2017. It will move into volume production at 7nm in early 2018. In addition, the company reiterated its plans to move into 5nm “risk production” in the first half of 2019, with mass production slated for 2020.

Dialing up 5G
One of the more significant events took place during Semicon week. It didn’t happen at Semicon, though.

The National Science Foundation (NSF) in the United States announced that it will invest more than $400 million over the next seven years to develop next-generation wireless technology.

Basically, that involves 5G, the follow-on to the current 4G cellular standard. 5G is a complex technology and involves a multitude of pieces.

5G is important for several reasons. First, today’s 4G technology doesn’t have enough bandwidth. Perhaps the next 4G standard, LTE Advanced Pro, can solve the problem.

5G promises more bandwidth at faster speeds. Another reason why 5G is important is clear. The semiconductor industry needs a new growth engine. The current engines for growth are PCs and smartphones. The PC market is on the decline, while the smartphone business is maturing and flat.

There are some bright spots like automotive and medical, but those markets are fragmented and remain relatively small in comparison.

It’s still a matter of debate, but 5G promises to jumpstart the smartphone industry. It will require a new class of phones with high-speed application processors, modems, RF chips and other devices. It will also require new chips for the infrastructure equipment.

To be sure, the industry is placing big bets on 5G. Intel, Qorvo, Qualcomm, Samsung and many others are racing each other in what will likely be a marathon.

Related Stories
Fab Investment Increases In China
What China Is Planning
200mm Equipment Shortfall
10nm Versus 7nm
Waiting For 5G Technology

Leave a Reply

(Note: This name will be displayed publicly)