Moores law is not dead, it is moving.
It is becoming obvious that the semiconductor industry is facing challenges like never before in its history. Organic growth is stagnating and costs of pursuing Moores Law of ever-increasing performance gains, are skyrocketing.
To support the increasing costs of participating in the technology race, semiconductor companies are trying to acquire each other in the hope that scale of economies will improve the semiconductor business model. Unfortunately, the industry is obsessed with the short-term return at a time where a long-term perspective is needed to sustain the technological development.
Mergers are complex financial transactions and their success is based on assumptions on synergy effects in the future. As both the winning and the losing management team of the mergers are rewarded handsomely, it is difficult to understand if they make senseAcquiring revenue is not cheap and so far there is no evidence that mergers have enabled the top semiconductor companies to grow organically again.
Although the graph cannot tell if a merger is sensible or not, it does raise some questions of the logic of some of these mergers from a revenue perspective.
China has not been allowed to play the acquisition game
Recently the Obama administration blocked Fujian Gand Chip Investment Fund's attempted takeover of German chipmaker Aixtron on the grounds of national security. Earlier the government's interest in a deal between Tsinghua and Lattice made the deal fall apart, although it looks like Chinese-backed investors from Silicon Valley is going to make another attempt.
The decisions by the western governments to try and prevent China from gaining a larger share of the semiconductor market, are logical. For quite some time the Chinese government has known that their electronics manufacturing industry only captures a minute fraction of the value of the end product. For an iPhone, the labour costs are between 2% to 5% of the retail price of the phone.
To capture more manufacturing value, China needs semiconductors
China has a semiconductor plan
To capture the semiconductor value, The Chinese government has initiated an ambitious plan. This plan is not limited to become the most important country for semiconductor foundries but also to create and develop a number of tier 1 semiconductor companies.
The first stage of the plan from 2015 to 2020 has an investment budget that rivals the R&D budgets of the 10 largest spenders in the semiconductor industry. Although the Chinese government has not specified what the investment pool is targeted at, it is reasonable to assume that it could be used for everything from fab investments to acquisitions.
Contrary to western companies, the Chinese government has a long-term perspective. They are not scared of the massive investments and delayed returns that are required to prolong the life of Moore's law.
The move from western governments to stop China from acquiring western semiconductor companies is not likely to stop the Chinese plan. Contrary to the western semiconductor industry the Chinese semiconductor industry is growing organically. Every year the grab market share from the established players.
If you can grow organically you don't need to buy them, you can beat them
The Chinese manufacturing industry is already is the largest consumer of semiconductor products, consuming 57% or 190B$ of the total global production in 2015. Both manufacturing and supply have had high organic growth rates over the last 5 years in an industry that is virtually flat.
The current semiconductor production allows China to support approximately 30% of their current manufacturing. The Chinese government has ambitions to raise the self-sufficiency rate to 40% in 2020 and then to 70% in 2025. Even if the Chinese consumption of semiconductors stayed flat at 2015 levels, the 2020 goal would require the Chinese semiconductor industry to capture revenue the size of Broadcom and Qualcomm's combined revenue. To get to the 2025 target, Intel and Samsung's revenue would be needed as well.
To survive, semiconductor companies need to grow organically again.
A false assumption of the western semiconductor companies is that the industry can survive by acquiring each other. The winning companies become larger but they are not growing organically. At the same time, they spend most of their energy and resources on integration, cost cutting and synergy exercises. The R&D budget is aimed at making chips cheaper and not better.
While western semiconductor companies are busy making the quarter, the Chinese are busy winning the decade.
Semiconductor companies used to be innovation factories. New products like the processor created the foundation for the PC industry. Networking chips revolutionised the telecom industry and created the foundation for the internet, but as time passed the industry forgot to innovate, Semiconductor companies became obsessed with getting orders. As a result, customers has commoditized the industry by using standard components and keeping semiconductor companies out of the buying process.
The way back to organic growth
To get back to organic growth, Semiconductor companies need to create more value for their customer. They need to be involved earlier in the customer's development process. Sales people need to be transformed into buying consultants that are experts in the customer's problems and their markets and they need to be visible so they get invited to the project early. The rise of the Social network gives semiconductor companies a chance to demonstrate to customers that they can benefit from inviting experts to their development process. Unfortunately most semiconductor companies only use Social Media to advertise and not to demonstrate that they have valuable experts available to customers.
Engagement Group is helping semiconductor and other hi-tech companies grow organically through social networks. We create new revenue from new customers.
- We make your organisation ready for the social network tsunami. You need to understand the social principles to survive and surf the wave.
- We transform your sales people from sellers to buying consultants attracting new customers and new orders.
- We transform your advertising to content valued by the customer, at a fraction of the advertising cost.
- Social Selling is so good it is irresponsible not to use it.
Visit Engagement Group for more information or contact Claus Aasholm
Principal Engineer | Device Design | FEOL | TCAD | FinFET
8 年Great article, Claus! Thanks! I’d like to comment about Moore’s law. The problem is quite fundamental 1) we need better materials (to switch transistors faster), better device arachitectures (FinFET, gate all around) thereby we are playing with strained silicon. But the performance gains are really marginal. Next step could be A3B5 semiconductors, to accomplish this we really have to alter process flows (and the wafer manufacturing equipment). 2) Transistors with tiny features – EUV is a hot topic and long overdue for industrial applications. From equipment point of you the industry just put all the eggs in one basket. The wiser approach we be pursuing alternatives like: X-ray, e-beam. 3) Demand side: Smartphone market – the key driver of leading edge nodes ICs is stagnant. The industry needs the next really BIG driver to justify cost of R&D and new fab construction.