The Big-bang Semiconductor Theory of 2015
25 December, 2015
Surprising medium-scale semiconductor acquisitions opened 2015. But when the really big deals were announced - it became clear that something serious is going on...
Surprising medium-scale semiconductor acquisitions opened 2015. But when the really big deals were announced – it became clear that something serious is going on…
The year 2015 (including the end of 2014) took everybody by surprise in the semiconductor industry. Even the most experienced observers remained speechless following the long list of gigantic mergers and acquisitions, that was announced almost every month.
To mention just few: Analog Devices for purchasing Hittite for $2.5 billion, Infineon acquire International Rectifier for $3 billion and Qualcomm bought CSR (previously acquired by the silicon) for $2.5 billion.
Than came the large-scale deals: NXP paid $11.8 billion to get Freescale Semiconductor. Intel has made a bold move and bought the top-tier FPGA producer Altera for $16.7 billion. Another pillar fell when Western Digital announced the acquisition of SanDisk for $19 billion, and in May 2015 Avago threw a bomb: acquisition of Broadcom for $37 billion – the largest semiconductor transaction ever.
|SEMICONDUCTORS MERGERS & ACQUISITIONS ($B)|
Gradually it becomes clear that something serious is going on the market. That the huge wave consolidations is changing the industry. This time it wasn’t impressive exits of innovative startup companies, but a game of giants pursuing to become mega-giants. Now it is clear that 2016 will be totally different than 2015.
The Problem with Moore’s Law
So here is the question: why the global chip industry is changing its strategy and focus on market economy instead of the traditional innovation driven economy? To be honest, the VC funds where the firsts to sense a problem. During the last 5 years they lost interest in semiconductor’s startups, and pushed many of them to a state of financial dehydration.
This effect is felt everywhere – from India, to Israel and even in the Silicon Valley. They understood that technology as an asset has lost its flavor. Later on, other signs came clear: The 450nm wafer size revolution lags more than three years after the original plans, and for now it seems as far as it was three years ago.
Than came the 28nm paradox. For 50 years the semiconductor industry has enjoyed from what is known as Moore’s Law: a continuous improvement in the production processes, which doubled every 18 months the number of transistors on a chip, and made each one cheaper. This was the foundation for the technology revolution. But when the industry passed from 28nm process to the more advanced 22nm, it became clear that the rules have changed: the price of each transistor actually increased!
Moreover, each new generation is harder to develop and less profitable. For example, although several chip manufacturers today produce chips with 11 nanometer process, they are limited mostly to memories and processors. Analog circuits are better produced in the former 65nm and 45nm processes. And today, only handfull are talking about the next step, of 5 nm.
In other words – Moore’s Law is cooling down. The research institute imec, for example, believes that we need new packages and die stacking technologies to maintain Moore’s Law. Nothing said about new process.
Maybe this is the new beginning: the semiconductor industry of 2016 will start without Moore’s Law. This is why the big players change their strategy to scale economy instead of innovation economy: In 2016 we will see a new kind of innovations – mainly in supply chain optimization, better yields and new sales and marketing schemes.