2 Comments

Summary:

Applied Materials and Tokyo Electron are merging to become a $29 billion chip equipment maker as the semiconductor industry undergoes a radical shift. Here’s what’s behind the deal.

22nmwaferthumb

Semiconductor equipment giants Applied Materials and Tokyo Electron said they would merge to form a $29 billion company Tuesday morning. The all-stock deal has been unanimously approved by both boards. For those of you who might not know much about these companies, they provide the multi-million-dollar machines that make semiconductors.

Those machines range from etchers, which trace lines into a silicon wafer, to the chemical vapor deposition machines that deposit fine layers of materials onto the chip as part of the manufacturing process. The machines are big, they are ugly and they are really, really expensive.

The clean room at Applied Materials where chip-making machines are made.

The clean room at Applied Materials where chip-making machines are made.


They are also from an era when chip manufacturing was focused on cramming more transistors onto a chip — when performance and raw power was king. But as the lines etched on the chips get ever closer together, making chips that are both smaller in total size, but with more transistors densely packed in, is becoming incredibly difficult. And that difficulty makes manufacturing more expensive (billions of dollars and ten years of research went into creating Intel’s 3-D chips, which were an advance in getting more transistors on these shrinking surfaces).

Not only are the physics problems of keeping Moore’s Law going becoming more daunting, the industry has decided it may not need it anymore. That fundamental shift away from workhorse, powerful CPUs at the high end is probably the biggest reason for this merger. Another is that the number of companies that buy Applied and Tokyo Electron’s gear is shrinking.

The end of Moore’s Law

Moore’s Law, which states that the number of transistors on a chip will double every 18 months, has driven innovation in electronics for the last few decades. It ensures that the next generation of processors can do more and reduces the price of processors that came before. That’s why I can buy a terabyte drive for $100, giving me the storage that supercomputers once had.

A wafer-insertion tray as a piece of silicon becomes a wafer full of chips.

A wafer-insertion tray as a piece of silicon becomes a wafer full of chips.

But it’s becoming less relevant in an age where we’re spreading our computing into billions of mobile devices and sensors. Sure, there will always be a need for high-performance CPUs somewhere in the data center, but even in that market companies are exploring alternative architectures that don’t consume as much energy. Not only that, but there’s greater interest in all parts of the industry in systems on a chip, where multiple cores are integrated into a package that can handle computing and networking or computing and rendering graphics or any number of variations.

Additionally, with more and more connected sensors there’s a greater need for high-power radios that don’t suck battery, yet still can perform, which typically means using new materials instead of silicon as well as a greater use of microelectromechanical chips, or MEMS, which are basically any chip that senses something from the analog world like pressure, sound, light, etc.

Taken together, these four things are changing the needs of chip manufacturers:

  • a need for lower-power chips;
  • a need for systems on a chip instead of a single-purpose package;
  • a need for new materials for radios and high-performance chips at lower power; and
  • a boost in the use of MEMs.

And this means the equipment vendors must adapt. Companies and startups such as Suvolta, Soitec and ThinFilms Electronics have all come on the scene with improvements to the traditional manufacturing technology, while researchers continue to explore the use of materials such as graphene in making new types of chips. With a smaller customer base that has changing demands, plus a need to put a lot of dollars into R&D to meet those demands, both companies decided working together gave them the economic heft needed. From the release:

Today, the mobility trend is driving a new phase of industry growth and introducing dramatic and fundamental technology changes in the way devices are made. Materials innovation is the most significant lever for customers to drive cost-effective performance gains in mobile chips and displays. With the best and broadest capability in materials engineering, Applied Materials and Tokyo Electron believe this new company will be well-positioned to provide valuable, differentiated device performance and yield solutions that enable the new device architectures and cost-effective scaling that customers need to win.

That’s a fancy way of saying: “We got together because our industry needs some fundamentally different gear, and we can’t afford to do the hard physics research required to meet their needs without the investment from both of us.” On my first day at GigaOM almost six years ago, I wrote a story explaining how to make smaller and smaller chips you have to be big. That trend is now going beyond the chip manufacturers and on to their vendors.

The deal details

And now for the financials of the deal. It looks like Applied has taken out Tokyo Electron with Tokyo Electron shareholders set to receive 3.25 shares of the new company for every Tokyo Electron share held. Applied Materials shareholders will receive 1 share of the new company for every Applied Materials share held. When the deal is done, estimated to be some time in the middle or second half of 2014, Applied Materials shareholders will own approximately 68 percent of the new company and Tokyo Electron shareholders approximately 32 percent.

The combined company will have a new name, dual headquarters in Tokyo and Santa Clara, a dual listing on the Tokyo Stock Exchange and the NASDAQ, and will be incorporated in The Netherlands. We’ll have to wait and see what the new name is.

You’re subscribed! If you like, you can update your settings

  1. I’m pretty sure that Moore’s Law is not less relevant. It’s what’s behind the continued fall in chip prices and continual increase in performance. This merger is all about Moore’s Law.

  2. Maybe there is a new law emerging, driven by sensing capability, power consumption, and radio connectivity. Smart phones with cameras, cameras with bluetooth and GPS etc. We are in an era of using computing power for more specialized tasks, such as the cell phone, camera, medical devices, Automotive engine, transmission and braking control.

    In this new era it may be more valuable for small companies and emerging economies to have very low cost chip fabrication capability so that new devices we can’t even imagine today could become possible.

    If computational power was free or nearly so, what devices might come on the market?

Comments have been disabled for this post