What’s driving it, what’s next, and who should be concerned?
Consolidation is picking up again across the semiconductor industry, against a backdrop of looming interest rate hikes, geopolitical uncertainty, and the erosion of longstanding demarcations between markets.
In the past couple of weeks, Siemens signed a deal to buy Mentor Graphics for $4 billion, and Samsung purchased Harman, a Tier 1 automotive supplier for $8 billion. What’s different is that deals worth multiple billions of dollars—sometimes tens of billions—are no longer extraordinary. When Texas Instruments paid $7.6 billion for Burr-Brown in 2000, it was considered a jaw-dropping amount of money.
Still, the curtain may be closing on these kinds of deals. On the political front, there is talk by the incoming U.S. Administration that the United States will leave the Trans-Pacific Partnership in January, along with threats about tightening the North American Free Trade Agreement rules. Great Britain’s exit from the European Union has fueled similar kinds of discussions in Europe. And to add a sense of urgency to these acquisitions, interest rates are poised to rise over the next few months, signaling the end to a prolonged period of cheap capital that makes acquisitions far less risky.
“The industry has been in digestion mode for a while,” said Simon Segars, CEO of ARM. “We’re seeing another wave of acquisitions now, driven by economies of scale and more technological challenges.”
So why are companies willing to spend so much? Top executives from across the semiconductor industry almost always point to three main drivers for acquisitions:
1. Operational synergies. By merging operations, the idea is that redundant functions can be eliminated, such as general and administrative functions, allowing savings on one side and bigger investments in R&D. But the jury is still out on whether R&D will increase to equal or greater numbers after these acquisitions are completed.
2. Horizontal critical mass. The more volume and revenue a company has, the more convincing it is as a leader across multiple connected technologies, such as various verification technologies, software development, security, or power management. That makes it easier to wring out better pricing from suppliers and to garner a premium from customers. It also feeds the R&D machines of ever-larger companies, as long as the numbers continue to point up and to the right.
3. Vertical positioning. As the IoT begins to splinter markets into smaller and smaller pieces, it becomes harder to make an investment in those markets without enough mass, and one way to do that is by amassing more companies that have a share of those markets.
“For 30 to 40 years, semiconductor companies enjoyed a huge compound annual growth rate, which obfuscated some of the bad decisions,” said Jack Harding, president and CEO of eSilicon. “Now we’re seeing single-digit growth. The way to keep the investment community interested is by expanding margins and lopping off overlapping R&D and G&A. But the jury is still out on how much R&D was not overlapping and just what this means.”
Behind the scenes, though, numerous other changes are at play that are beginning to add new context to these acquisitions. As more devices are connected, longstanding perimeters around markets are changing, and in the scramble to stay current large companies are buying up technology companies rather than internally developing technology.
In fact, consolidation within the semiconductor industry and its various segments, such as EDA and IP, have been ongoing for nearly two decades. The size of acquisitions has increased recently, mainly because many of the semiconductor startups have disappeared and the only ones left are larger companies. Compounding that, there is not enough interest on the part of venture capitalists in funding new companies—or at least not at the same rate they are being acquired—because time to profitability is much longer than in software, AI, or other smart devices that can be built with mostly off-the-shelf parts.
“This consolidation has been going on for multiple years,” said Aart de Geus, chairman and co-CEO of Synopsys. “Along with the big ones, there have been quite a number of not-as-visible acquisitions. But consolidation can’t continue forever. There is a gradually decreasing tail.”
Along with these big acquisitions, though, there are other shifts occurring that are less obvious.
“The more they cut their R&D, the more we increase ours,” de Geus said. “A lot of companies are decreasing their internal EDA support. The migration of IP has been outsourced. If they drive their margin up from these acquisitions, they underinvest in R&D. And if they drive their margin down, they need to make sure they have enough investment.”
Typically, this kind of consolidation and reapportioning of work happens in mature industries, such as steel and mining, and until several years ago, in automobiles. Revenues flatten enough to warrant these kinds of efficiency moves. In contrast, semiconductors is a relative newcomer, and it continues to change and grow rapidly with the development of autonomous vehicles, robotics, machine learning and artificial intelligence.
“You can’t apply the same mold to the semiconductor industry that you can to other industries,” said Wally Rhines, chairman and CEO of Mentor Graphics. “We are still decreasing the cost per transistor by 30% every year, and people are still putting together pieces for dominant market share and spinning off pieces they don’t want. So the industry will continue to specialize, which is the same as other markets, but there also will be a lot of new developments.”
The slowdown in Moore’s Law is a contributing factor in these shifts, if not a central driver. The rapidly rising cost of developing chips at the most advanced nodes, coupled with the flattening of the mobility market, means that far fewer chipmakers can justify moving to the far reaches of the semiconductor roadmap.
Even Qualcomm, which has emerged over the past decade as the leading provider of smartphone chips, has signaled its intention to move beyond just mobility. Its pending acquisition of NXP opens up a number of new market opportunities in automotive—a result of NXP’s $12 billion Freescale acquisition, which closed in December 2015—and in smart payments, industry and other IoT-related markets.
“If you look at a chip five years ago, they had very regular structures,” said Steve Mollenkopf, Qualcomm’s CEO. “Now these are almost organic in nature. There are no regular structures. Plus, we’re developing 10 times the software we did 10 years ago, and there is a big shift going on in packaging. There a whole new discipline in packaging that combines physics, mechanical engineering and materials engineering.”
One of the reasons that consolidation is so difficult to follow is that it cuts across regions as well as industries, and most government agencies are opaque on this subject.
The U.S. Treasury Department’s Committee on Foreign Investment in the United States, for example, screens international acquisitions based upon national security and makes a determination without public review. It does not allow freedom of information requests for details about why it rejects certain acquisitions. The U.S. isn’t alone. The European Union and China don’t talk about regulatory issues publicly, either. And neither do the companies that are attempted takeover targets.
Still, the consensus among semiconductor executives is that China’s acquisition appetite is increasing. The country is running a trade deficit of between $150 billion to $200 billion a year, depending on estimates from different analysts, because it imports 90% of its semiconductors. At last count, the China IC Industry Investment Fund, a joint venture between the Chinese government and private investors, had between $120 billion and $150 billion in cash. (It’s difficult to get a handle on the individual players in those funds, because there are so many investment groups, sometimes using different names, and government agencies that span from the national to the municipal level.)
Nevertheless, top executives predict this will come in two main thrusts. One involves an effort to boost the companies that have the greatest chance for success, such as SMIC and XMC. Moody’s reported last January that SMIC had received a $2.9 billion (based on a November 2016 currency conversion rate) investment from the Shanghai Integrated Circuit Investment Fund.
Tsinghua Unigroup, a spinoff of Tsinghua University, came to the attention of many analysts after it bought Spreadtrum Communications and RDA Microelectronics. In July 2015, the company made an unsuccessful $23 billion bid for Micron.
Last July, Tsinghua acquired XMC, a state-owned memory chip maker (also known as Wuhan Xinxin Semiconductor Manufacturing Corp.), in a deal the Wall Street Journal reported was brokered by China’s National Integrated Circuit Industry Investment Fund and the Academy of Sciences. The combined company is now under the auspices of the Yangtze River Storage Technology Co.
The second thrust is to continue acquiring companies outside of China that can quickly add expertise and fill gaps in China’s domestic output. Tsinghua Unigroup’s Unisplendour subsidiary bid $3.8 billion bid for a 15% stake in Western Digital Corp. in Sept. 2015. Western Digital planned to use that money to finance a $19 billion acquisition of SanDisk, but when Unisplendour pulled out it put Western Digital in a bind and prompted a renegotiation with SanDisk.
“Over the past 24 months China has experienced some growing pains,” said eSilicon’s Harding. “But they will execute a cohesive and focused strategy to decrease reliance on imported semiconductors. The right people are now shepherding these funds and they have a well-defined strategy for what they can accomplish. We will continue to see regulatory issues between governments, but this is not a hobby for the Chinese government. It’s very real.”
But how and where China gains a foothold is likely to be as multi-faceted as its investment strategy. Lip-Bu Tan, chairman and CEO of Cadence, said China’s effort to do more buyouts is slowing down, in large part because of regulatory issues.
“There are a lot of good entrepreneurs coming up in areas like artificial intelligence, robots and drones, and there is a lot of good technology there,” Tan said. “In drones, Chinese companies are not following other companies anymore. They’re the leaders. There’s a lot of computer vision intelligence in there. These are very advanced systems compared to the other drones.”
Defining a successful merger
What distinguishes a good merger from a bad one isn’t always clear at the outset. In some cases, it may be years before the wisdom of a move is apparent.
The pending $8 billion acquisition of Harman by Samsung is a case in point. The assisted/self-driving market is defined more by the opportunity than an immediate uptick in sales volume. While Harman’s sales for fiscal 2016 were $7 billion, its order backlog as of June 30 was about $24 billion.
“The promise is that if you’re a semiconductor company, you may have had one automotive customer in 2012, whereas now you have 9 or 10,” said Charlie Janac, chairman and CEO of Arteris. “Self-driving cars are a new bonanza. So are drones, which also use V2x communications. This will go on for a long time, too. Hardware decisions are made 8 to 10 years ahead of time. This is still very early days.”
In contrast, Janac said Arteris had 23 potential mobility customers (suppliers) in 2012 . There are now 11. “Volumes have increased, revenues have increased, but suppliers are down by half.”
In light of this, chipmakers are in experimentation mode, and acquisitions play a big role in that experimentation because at the moment, it’s easy to justify an investment given the availability of cheap capital.
“There are two reasons for acquisitions,” said Vincent Roche, president and CEO of Analog Devices Inc. (ADI). “One is long-term value. The second is short-term value. My strategy is that I don’t buy fixer-uppers.”
ADI bought Linear Devices for $14.8 billion in July, which is currently pending approval. It bought Hittite Microwave in 2014 for $2.45 billion. The two acquisitions provide ADI with mixed signal technology, microwave communications and power semiconductors, which Roche said are critical for high-performance computing.
“One of the big changes in the last 10 years is that the systems industry is following analog hardware capabilities,” he said. “We need a more systems-solutions business and a bigger toolkit. If you look at the IoT, it has four elements–sense, measure, interpret and analyze. We have the sense and measure pieces, and we’re experimenting with a lot of technology and business models.”
That kind of experimentation will continue for a lot of companies, particularly with interest rates as low as they are right now. But as interest rates rise, it will be harder and take longer to reap the benefits from an acquisition. In addition, the current political climate in a number of countries appears to be headed away from open trade. While it’s too early to make a call, the fear is that could have a big impact on acquisitions across political borders, the ability to obtain work visas for qualified engineers, and the ability to get the same return on investment through international sales.
Consolidation Aftermath (Dec’15)
Why the latest round of acquisitions is causing angst in the semiconductor industry.
Consolidation Hits OSAT Biz
China, Taiwan and big foundries lock horns over packaging market.
More Consolidation Ahead
China and other big players will wait until the price of borrowed capital increases before making big moves.