Who Owns What And Why

Changes in how chips are designed, manufactured, packaged and tested will create some interesting power shifts.


Who’s calling the shots these days—and how long they’ll continue calling the shots—is turning out to be as much conjecture as playing the futures exchange. There are so many changes underway that even engineers are crossing boundaries no one ever expected and ending up in companies outside of IC design or moving from seemingly far afield into the design world.

Still, there are some changes lately that no one would have anticipated six months ago. Here are some of the key ones to watch:

1. Microsoft is now in the hardware business. You might have surmised something was up with the introduction of Microsoft stores near Apple stores, not to mention rival Oracle’s acquisition of Sun, but Microsoft has really only competed in the hardware space with the mouse and gaming hardware. Something fundamental has changed here, and it’s the ability to bridge the gap between hardware and software. Apparently Microsoft is tired of waiting for Intel to make inroads in the mobile market, which is why it has started working with ARM. But is this a one-off or a trend?

2. Foundries have begun hoarding and jealously protecting technology that will be required for 2.5D stacks—notably the interposers and microbumping. With EUV late, the cost of developing an advanced SoC hitting hundreds of millions of dollars and very few companies willing to put that kind of money into a chip, all signs point up. With TSMC’s recent investment in ASML, not to mention its previous stake in Global Unichip, things could get very interesting. And where exactly do the packaging and test houses fit into this picture?

3. EUV lithography is still not ready. In fact, it has missed the past several deadlines. You can call that water under the bridge, but the repercussions if it misses the 14nm process node could be enormous. Even though there will be fewer players at that node, those chips will play an important role in the most popular consumer electronics, not to mention becoming platforms for stacked die. If the IC industry is forced to consider alternatives—gridded design rules, directed self-assembly and a few other more exotic ideas—all of those could have disruptive repercussions throughout the ecosystem, from design to manufacturing.

4. Consolidation in EDA is continuing at an astounding pace. Witness Synopsys’ acquisition history of late. What’s intriguing about this is that companies at the bleeding edge are complaining that EDA companies aren’t investing enough in advanced nodes. Still, there are also startups emerging, notably in China, France, India, and even a handful in Silicon Valley, focused on the most advanced problems. Some of these startups may actually become industry names (nothing is a household name in design unless you actually work in design). When these startups emerge from stealth mode, keep an eye on whether they’re focused on growth or exit strategies.

Individually, these moves are interesting. Collectively, they’re like connecting the numbered dots on some children’s puzzle. The only question now is what the final picture looks like—and who’s drawing it.

—Ed Sperling