Big Acquisitions, Big Changes

Rumors swirling about Intel’s romance with Altera—this has been off-again, on-again, and now apparently off-again, for the better part of a decade—coupled with Apple’s decision to shift A9 APU production to Samsung and away from TSMC, NXP’s pending acquisition of Freescale, along with the Chinese’ government’s massive semiconductor investment fund, all add up to some massive shifts under way in the global supply chain.

M&A typically happens for a couple of reasons. One is market weakness, where the strong buy up the weak at fire-sale prices for assets and to clear out any lingering competition. That clearly isn’t the case here, particularly with the prices being paid or offered. The second involves growth opportunity, where premiums are paid for companies based on the assumption that there are giant new market opportunities ahead. Google is famous for paying exorbitant sums of money for companies and research, and Facebook isn’t far behind, which have made them the darlings of banks, analysts, and business and engineering school graduates.

This kind of exuberance hasn’t been seen in semiconductors for many years. But times are changing, and they’re changing rather quickly and with massive implications everywhere.

What’s driving a lot of this is the IoT, IIoT, IoE, or whatever label people are applying to it these days. The ability to connect devices together everywhere, whether it’s just within a single market—cars to cars, for example—or whether it’s across markets, such as controlling a machine with a smart phone—is a bonanza for electronics companies. In fact, it may be the largest opportunity in history, because it cuts across political boundaries, across vertical markets, and it creates new opportunities and entire businesses that never existed before.

But it doesn’t create opportunities everywhere equally. If Intel buys an FPGA player—regardless of which of the big ones it ultimately grabs—it will have a pair of standard products that can support Moore’s Law and fund feature shrinks in its foundry business at least through the next decade. And with Apple’s business, it appears Samsung will do the same, with GlobalFoundries playing second source on that plus AMD’s business.

But it also could push the remaining foundry players—TSMC, UMC, ST and SMIC—in entirely different directions. It may cost less to invest in stacked die and packaging technologies than to invest heavily at the front end of Moore’s Law, with planar FD-SOI as the end goal. Or, it may be the reverse, propelling everyone to focus on Moore’s Law, particularly if EUV or some other lithography technique proves itself, rather than focus on new packaging options.

The acquisitions under discussion and the investment sums being discussed—in China’s case, it is estimated to be more than $50 billion over time, not just the current $19 billion fund—all have the power to reshape the entire semiconductor industry. That means processes, tools and equipment all the way out to who does business with whom and who the power brokers are within those businesses. We are witnessing the beginnings of what perhaps will the biggest upheaval in economics since the invention of the transistor, and so far no one has a clear picture of what that will look like when all the dust settles.